Tubular Tire Adhesion Performance - Part 7 Examination of Tubular Tire Adhesion to Carbon Fiber Rims Colin S. Howat Kurata Thermodynamics Laboratory Department of Chemical & Petroleum Engineering 1530 W 15th , Room 4132 Learned Hall Lawrence, Kansas 66045-7609 www.engr.ku.edu/~ktl ____________________________________________________________________________________

Abstract Tubular tires are still the choice of many cyclists. The lighter tire/rim combination provides superior acceleration performance. The tires do not pinch flat. Many tubular tire cross sections are uniform, providing predictable cornering. Recent trends supplant aluminum rims with carbon fiber ones. The latter provide superior stiffness for the same mass. Braking heat dissipation is problematic. Tubular tire adhesives were developed for aluminum rims. Evidence suggests adhesion to carbon fiber is inferior to adhesion to aluminum. The purpose of this study was to quantify adhesion to carbon fiber with readily available glues. Previous studies demonstrated the superior performance of Continental glue. This glue, along with Vittoria Mastik’One, Panarace Panacement, Clement Gutta and Tufo tape, were compared. The rims used are Bontrager carbon fiber and Wolber Profil 19’s. The tires used are Continental LA’s and Competition 20’s. Mastik’One is the best glue to use for carbon fiber rims. Continental was the worst performing of the conventional glues. This is a laboratory study performed under laboratory conditions. Mechanics are urged to use their experience to temper the results presented herein. The mating geometry was typical but not ideal. Gluing to the edges is critical to performance when there is a protruding seam that may interfere with good contact. The safety of the rider is paramount and these results should be combined with experience to maximize the rider safety.

Introduction Tubular tires provide acceleration and cornering advantages over conventional clincher tires. Further, they are not prone to sudden deflation due to pinch. They are still the tires of choice of many competitive cyclists. KTL has published six papers covering adhesive, tire, rim, application and temperature impact on adhesive performance. For all around use with aluminum rims, Continental glue was superior. It has one of the shortest times to come to full strength. It has less degradation at higher temperatures. It was one of the best for tires with and without latex coated base tapes. It was best or one of the best with non anodized and anodized aluminum rims. It did not separate during storage. Tubular tires have two adhesive joints. The first, controlled by the tire manufacturer, holds the base tape to the casing. The second, controlled by the mechanic, holds the base tape to the rim. Solvents in adhesives that are not designed for tubular tire installation can attack the base tape - casing joint. This attack increases the likelihood 1 C. S. Howat 12/04

that the tire will roll during cornering or loss of traction situations. Readily available glues at hardware and auto supply stores may not be appropriate for tubular tire applications. Further, ones that may appear to work for some tires will not for others because of the differing properties of the base tape - casing adhesive. In general, mechanics should restrict options to those designed for tubular tire installations. Carbon fiber with its superior stiffness to weight ratio is supplanting aluminum as a rim material. The surface properties of carbon fiber resin are substantially different from those of aluminum. Mechanics and riders should anticipate differences in braking and tubular tire adhesion performance. The purpose of this study is to examine adhesion to carbon fiber rims.

Materials Six glues, two rims and two tires were used in this study. Three of the glues - Continental, Vittoria Mastik’One and Panaracer Pana Cement - were ‘clear’, synthetic. One, Clement Gutta, was ‘red’, natural based. One was a tape, Tufo. The last was a polyurethane based glue sold in hardware stores, Probond. The rims were blank (no spoke holes) Profil 19’s and ‘paired-spoke’ Bontrager Race Lite XXX. These have nearly the same cross section. The tires were Continental LA’s and Competition 20’s. These have a noticeable seam. The base tape is uncoated. The protocol used requires a torque wrench, electrical conduit, C-clamp and tire insert. Glue application, when applicable, was done with an ‘acid’ brush.

Procedure The protocol used for these studies is that developed and documented in the previous six papers. Briefly, the rims receive three coats of glue with 24 hour curing between the first and second coats, and 12 hour curing between the second and third. The base tape receives one coat of glue at the time the first coat of glue is put on the rim. The rims were cut into 80 mm sections ensuring that the section did not contain a spoke hole. After cutting, the rims were washed with soap and water, rinsed and then rinsed with acetone. They were placed in the oven to dry. The tires were cut and tube removed. The tires were then cut into 80 mm lengths. Flexible electrical conduit, flattened to accommodate the seam, was inserted into the tire to simulate air pressure.

2 C. S. Howat 12/04

After applying the third coat of the glue, the tire section was positioned on the rim and held in place during curing with zip ties. Figure 1 shows a mounted tire/rim section. The mounted tire/rim combination was allowed to cure at least 24 hours.

Figure 1: Mounted Tire/Rim Section

The tools used to determine the roll off torque are: a tire insert, a C-clamp, a slotted socket and a torque wrench. Figure 2 shows three of the four tools.

1: Mounted Tire/Rim SectionSocket FigureFigure 2: C-clamp, Tire Insert and Slotted Once cured, the tire insert was placed in the tire. This insert provides resistance to the C-clamp. The insert provides ‘form’ for the tire to simulate air pressure. The C3 C. S. Howat 12/04

clamp is mounted next. The clamp is tightened against the insert. The tightening torque must be sufficient to withstand movement when torque is applied to the C-clamp. A slotted socket is placed on the end of the C-clamp. The width of the tire at the C-clamp is measured. The zip ties are cut and the assembly is placed in the vice. A torque wrench is used to simulate lateral forces on the tire. Figure 3 shows the assembly prior to application of the roll off torque.

Figure 3: Assembly Ready for Roll Off

A steadily increasing force is applied to the torque wrench. The torque at failure is the maximum torque that the adhesive joint can withstand. During this study, once roll-off initiates, two types of situations occurred. The first is that observed in previous studies. The torque falls slightly and then remains steady as roll off progresses. The tire appears to ‘peel’ from the rim. The second is a snap failure. Once maximum torque is reached and roll off initiates, the torque required to continue falls to near zero. These two observations imply that duration of force application is a factor in adhesive performance. Consider the typical roll off situation. A rider enters a corner, loses traction and regains it abruptly. In the first situation, roll off might initiate but because there is a ‘peeling’ effect, the tire will not dislodge. In the latter, the sudden force would dislodge the tire. The ‘snap’ failure had not been observed in previous studies with aluminum rims. Snap failure occurs with some of the glues, notably Continental, on carbon fiber rims. Even though the maximum torque to initiate roll off is the same, the performance could be different on the road. The protocol developed does not address force duration. Therefore, some of the measurements indicating good performance may be misleading in actual on-the-road applications.

4 C. S. Howat 12/04

Glue Solids Content The percentage nonvolatiles in each of the tubular adhesives was measured by discharging tubes into weighed sample dishes. These were allowed to outgas in either an oven or a fume hood. The results are given in Table 1. Table 1: Percentage Solids in Tubular Tire Glues Continental Vittoria Mastik’One Panaracer Pana Cement Clement Gutta

30% 35% 41% 76%

The procedure briefly described above and described in detail elsewhere has the purpose to build a bed of contact between the tire and rim. A high percentage solids may give a better bed of contact. But, it may be more difficult to apply. As the solvent evaporates during drying, the glue volume will decrease. This may result in poor contact area. This is problematic when the tire has a protruding seam.

Baseline Studies The baseline studies provide a comparison between rim types and glue types. In this set, aluminum and carbon fiber rims with Continental LA and Competition 20 tires were glued using Continental glue from a 350g can. Six samples of each combination contribute to the mean performance presented in Figure 4. The performance is essentially the same among the four combinations when only the torque is compared. The performance during roll off, however, is different. The roll off from the aluminum rim is a ‘peel’. The failure from the carbon fiber rim is a ‘snap’. Under short duration spikes in lateral force, the likelihood of roll off failure from the carbon fiber rim is greater than that from the aluminum rim even though the apparent adhesion is the same. Further, the adhesion failure is substantially different. The goal in adhesion is to have failure in the adhesive. That is, the cohesive forces in the glue are less than the adhesive forces between the glue and substrate (base tape or rim). In these samples, the adhesion to the rim was poor for both rim types. The failure was about 80% of the surface area for the aluminum rims. That is, 80% of the gluing surface of the rim was free of glue after roll off. If the adhesive had performed optimally, there would be a thin coat of adhesive on the rim and on the base tape. The adhesion failure to carbon fiber was worse, in excess of 90% failure. Comparing the two, this means that the adhesion to the carbon fiber is less than 50% of the strength of that to aluminum. The absolute torques given in Figure 4 can be translated to a ‘relative strength’. The relative strength is defined as the torque for a particular combination divided by the performance for Continental glue - Continental LA tire - aluminum rim. Figure 5 shows the performance on a relative basis. The relative performance will be used for the remainder of this paper.

5 C. S. Howat 12/04

Figure 4: Absolute Torque 60

Roll-off Torque, in-lb

50

40 LA/AL

C20/AL

30

LA/CF

C20/CF

20

10

0 Performance with Continental Glue (Can)

Figure 5: Base Line Performance on a Relative Basis 2.0 1.8

Roll-off Torque, in-lb

1.6 1.4 1.2

LA/AL 1.00

C20/AL 1.06

1.0

LA/CF 0.98

C20/CF 0.93

0.8 0.6 0.4 0.2 0.0 Performance with Continental Glue (Can)

6 C. S. Howat 12/04

Sanding Carbon Fiber Rims Mechanics in the past have roughened aluminum rims to try to enhance adhesion. This did not improve adhesion. Nevertheless, the practice has carried forward to sanding the resin of carbon fiber rims to try to improve adhesion. Samples were prepared. The rims were sanded with 50 grit paper. The scarring of the resin was noticeable. The rims were washed again with soap and water and rinsed with acetone. After drying, the three step procedure was repeated. Figure 6 gives the relative performance. Figure 6: Effect of Sanding Carbon Fiber Rims with Continental Glue 2.0 1.8

Roll-off Torque, in-lb

1.6 1.4 1.2

LA/AL 1.00

LA/CF 0.98

1.0

LA/CF/Sanded 0.94

0.8 0.6 0.4 0.2 0.0 Performance with Continental Glue (Can)

The performance is essentially the same. There is no benefit to sanding Bontrager rims. Further, Figures 7 and 8 show that the adhesion performance is the same. There is little residual adhesive on the rims. The sanding marks are readily apparent. This further demonstrates that sanding provides no benefit and is a waste of the mechanics’ time.

Tubular Adhesive Performance The purpose of these studies was to compare the four liquid adhesives designed for tubular tire applications. Six samples for each adhesive were prepared. Bontrager carbon fiber rims and Continental LA tires were used in all cases. All glue samples came from small tubes. Figure 9 presents the statistical results. The results clearly show that

7 C. S. Howat 12/04

Figure 7: Four Samples of Sanded Rims after Roll Off

Figure 8: Two Samples of Sanded Rims after Roll Off

Vittoria Mastik’One is superior to the other adhesives under these conditions. Figures 10-12 present rim photographs after roll off. The figures show the residual adhesive on five of the rim samples. The pattern is the same as that of Continental. The edges have a residual adhesive. There is more adhesive left on the gluing surface, however, than in the Continental case. The base tape adhesive failed in one of the cases, i.e. when the tire rolled, the base tape adhered to the 8 C. S. Howat 12/04

rim as the casing rolled off. Mastik’One is a thicker adhesive than Continental. This may have contributed to the increased performance with a thicker bed of contact and more contact area. This hypothesis is supported by the Clement case. Figure 9: Comparison of Glues Using Carbon Fiber Rims 2.0 1.8 Mastik'One 1.43

Roll-off Torque, in-lb

1.6 1.4 1.2

Clement 1.36 Panacement 1.04

LA/AL Cont(C) Continental 1.01 1.00

1.0 0.8 0.6 0.4 0.2 0.0 Performance with Carbon Fiber Rims and LA Tires

Figure 10: Four Rim Samples with Mastik’One Residue

Clement Gutta performed nearly as well as Mastik’One. This is an old technology developed long before carbon fiber rims. Previous studies showed that 9 C. S. Howat 12/04

Clement performance with aluminum rims is inferior to nearly all other adhesives. Further, those studies showed that Clement perfo rmance degraded with increasing temperature such as that which is encountered on hot days or under intense braking. Therefore, it was a surprise when Clement performed nearly as well as Mastik’One.

Figure 11: Single Rim Sample with Mastik’One Residue

Figure 12: Mastik’One Sample with Base Tape Adhesion Failure

Figure 13 gives insight. Clement is very thick. It is difficult with which to work. Consequently, it was difficult to spread thinly on either the tire or the rim. The adhesive thickness was greater providing a better bed of contact. Nevertheless, the photograph 10 C. S. Howat 12/04

shows that Clement adhered to the tire and rim. It adhered to the rim to a far greater extent than the other glues did.

Figure 13: Rim and Tire Samples with Clement Glue

Tufo Gluing Tape Many avow that Tufo Tape is a good substitute for traditional glues. Tufo was tested with aluminum and carbon fiber rims. The tape is the Activated Tub ular Tire Gluing Tape. The instructions state that this tape was developed specifically for Tufo tires and that performance with other tires may differ. Continental Competition 20’s were used for this study. The tape is two sided. The side that adheres to the rim looks like conventional cellophane tape. The tape that adheres to the tire has a white strip running the length of the tape. This white strip adheres to the tire. The tape has two protective backings. The backing covering the cellophane surface is removed. The tape is positioned on the rim. A small section of the backing covering the tire adhesive is peeled. The semi- inflated tire is positioned and centered on the rim. The protective strip is then pulled from between the tire and the tape. One to four hours is the cure time. The white strip is marginally wider than the seam on the casing. Unlike conventional glues for which the primary bonding is the edge of the rim to the tire, the primary bonding is at the seam. The recommended procedure was followed to the extent that the cellophane backing was removed and the tape was centered on the rim. The protective backing was removed and the tire positioned on the white strip. Zip ties were used as before. Curing time was twenty- four hours. Roll-off measurements were done as before. Figure 14 gives the results. Clearly, the performance of Tufo Tape in this application, i.e. with Continental Tires, is inferior to that of conventional glues.

11 C. S. Howat 12/04

12 C. S. Howat 12/04

Figure 14: Tufo Tape Adhesion Relative to Continental (Can) Glue 2.0 1.8

Roll-off Torque, in-lb

1.6 1.4 1.2

Cont Glue LA/AL 1.00

1.0 0.8 0.6

C20/AL 0.52

C20/CF 0.48

0.4 0.2 0.0 Performance with Tufo Tape

Probond Performance An adhesive readily available in hardware stores is Elmer’s Probond. The claim is that it bonds anything. To test that claim, it was used with aluminum and carbon fiber rims with LA tires. This is a polyurethane adhesive activated by water. The rim and tire were both wetted, the glue applied and tire mounted. Zip ties were used to hold the tire in place during curing. The instructions note that the adhesive expands during curing. This is an understatement. If this adhesive were to ever be used for this application, the mechanic needs to be frugal in the glue application. Figure 15 presents the results. The adhesive does adhere to aluminum giving strikingly better performance than Continental rim cement. But, despite the claim, the adhesive does not bond to carbon fiber resin. It shouldn’t be used in this application.

Observations The mating geometry between the aluminum and tires and the carbon and tires was not optimal. The seam has a marked protrusion. This type of protrusion is also found on Vittoria tires. The tire-rim contact is three-point, i.e. the two edges and the seam line. Good practice always requires glue at the edges. But, in this situation with poor tire contact, it is particularly important. Clement’s performance was unexpectedly

13 C. S. Howat 12/04

Figure 15: Probond Adhesion Relative to Continental (Can) Glue 2.0

LA/AL, 1.83

1.8

Roll-off Torque, in-lb

1.6 1.4 1.2

Cont Glue LA/AL 1.00

1.0 LA/CF 0.69

0.8 0.6 0.4 0.2 0.0

Performance with Probond Adhesive

good as it has shown poor performance in previous studies. The performance was enhanced here because the thick glue made thin coats difficult to apply. The bed of contact was clearly complete as shown in the photographs. Rims, such as Wolber Profil 20, which ha ve a center relief to accommodate the seam should be considered when using tires with a protruding seam. Alternatively, tires without the prominent seam should be considered.

Conclusions Adherence to aluminum and to carbon fiber is the same for Continental rim cement under this testing protocol. But, the failure is more susceptible to impact failure. Sanding carbon fiber rims prior to coating with glue provides no benefit. Mastik’One is the best of the tested glues for carbon fiber rims. Tufo Tape is inferior to conventional glues in both aluminum and carbon fiber applications. Elmer’s Probond should not be used in either application. Gluing at the edge is of paramount importance in these applications.

14 C. S. Howat 12/04

Recommendations A new testing protocol needs to be developed to test the importance of impact failure v. peel failure. The tests should be redone with tires that better match the rim profile. Alternatively, a greater relief should be ground into the inserts to allow the seam to sink into the tire away from the rim.

Acknowledgments Occidental Petroleum is instrumental in supporting the myriad of activities at the Kurata Thermodynamics Laboratory. In particular, Dr. D. R. Laurance, President of Occidental Petroleum, is gratefully acknowledged. Park Tool Company through Mr. C. Calvin Jones is acknowledged for tools and supplies. Trek Bicycle Company supplied the carbon fiber rims. Continental supplied their tires and glue.

15 C. S. Howat 12/04