Longevity of NR structural bearings.
The first bridge in the United Kingdom to be supported by laminated rubber steel bearings is the Pelham Bridge in Lincoln, England. The original bearings were manufactured in 1957 and have performed without any problems since installation. The maintenance required for the bridge has involved the concrete supports. The present project entailed the removal of eight of the original bearings from one portal support and the installation of new replacement bearings. Mechanical and analytical tests have been performed on the extracted bearings. Results from the Japanese part of the study are given elsewhere (refs. 3 and 4). Preliminary results of the U.K. tests have already been reported (ref. 1); the full results are presented here. In addition to stiffness tests on the whole bearing to determine any changes since their manufacture, mechanical and analytical tests were performed on specimens cut from one of the bearings to assess the depth and significance of any surface oxidative aging.
Design and original tests
The design of the Pelham Bridge bearings is shown in figure 1. The bearings are completely encased in rubber with side and top and bottom cover layers all approximately 7 mm thick. When installed under the bridge, they were fixed to outer plates by dowel pins that locate in the upper and lower, thicker internal bearing plates. Thus, the outer rubber layers are not active when the bearing is sheared, though they contribute to the vertical compliance. The other internal plates contain an array of 28 holes of 10-20 mm diameter, possibly to provide mechanical keying between the rubber and steel layers.
[Figure 1 ILLUSTRATION OMITTED]
The rubber for the bearings had a hardness of about 70 IRHD, according to reference 5, and 65-70 Shore, according to reference 6. The formulation of the bearing compound is not available. A compound of similar hardness recommended in the early 1960s for use in bridge bearings is detailed in table 1; it is likely to be similar to the one actually used. The steel internal plates were said to be bonded to the rubber using the brass-plating process (ref. 6). No evidence of a copper peak, however, has been found from x-ray analysis of the plates in the SEM.
Table 1 - rubber formulation Material MRPRA bridge Amount from bearing analysis of recommendations Pelham bearing Natural rubber 100 100 Carbon black 60 (lampblack) 50(medium thermal grade) Zinc oxide 30 24 Oil 2 (Dutrex R) Dioctyl paraphenyl- 4 >3 amine diamine (DOPD) Sulfur 2.50 2.80 Accelerator 0.7 (CBS) ? Stearic acid 1 ? Other antidegradants 1 (PBN) >0.2
(*) (parts oer hundred of rubber [by weight])
The original stiffness specifications for the bearings is given in table 2. The shear stiffness is presumed to refer to the situation in which the top and bottom rubber cover layers are not active. Indeed, raking the shear modulus figure of 205 psi assumed in reference 5 for a 70 IRHD rubber, and the total thickness of rubber less the cover layers results in the design stiffness value of 10 tons/in. (3.95 kN/mm). Shear and compression tests were carded out (ref. 5) to verify the properties of the original bearings, but unfortunately, full details are not known. The observed values reported are given in table 2. Relying on photographs taken during the testing (ref. 5) and the few details provided, it seems that the compression stiffness was measured under 200 tons load, and the shear stiffness at a deflection of 25-32 mm, with the outer rubber layers active (i.e., the outer plates are not in place). It is likely a vertical load was applied during the shear test. From the lack of bulging of the rubber layers (according to the photographic records) any vertical load applied was substantially less than 200 tons, but, as the bearings under test appear to have been held only by friction, some vertical load must have been imposed. Provided it is adequate to prevent slip, the shear stiffness should not depend significantly on the magnitude of the vertical load.
Table 2 - Pelham Bridge bearings stiffness Original Design observations Design Ton/in. kN/mm kN/mm Compression 1,000 400 470 Shear 10 4.0 3.7
The eight bearings from Pier 14 were removed. The pier runs approximately E-W. Because of an overhanging footway, even the outer bearings are shielded from direct sunlight except when the sun is low in the sky. The mean temperature on the site is estimated at 10 [degrees] C. According to the bridge maintenance records, some bearings were cleaned by grit-blasting and painted in about 1980.
The two bearings examined in the U.K. were numbers 1 and 4 (bearings numbered E-W). Being nearest to the edge of the bridge deck, bearing no. 1 had a very dirty appearance. In contrast, no.4, which came from under the center of the deck, looked clean.
Superficial examination showed both bearings nos. 1 and 4 had been painted. In the case of the former, the coating was peeling off in places and the rest was easily removed. It had a thickness of about 0.25 mm. Analysis of the coating by NMR, x-ray analysis in an SEM and thermogravimetric analysis (TGA) suggests it is perhaps chlorosulfonated polyethylene. It is rubbery in nature, being flexible and slightly stretchable; it is thus assumed to have been permeable. The coating still adheres well in the case of bearing no. 4. Before any surface measurements were made, the coating was removed.
Mechanical tests: Experimental details
Compression and shear stiffness
The most important tests performed were measurement of the compression and shear stiffnesses of the bearings for comparison with the original test data on a similar bearing. Initially, the following series of tests was performed on the bearings in the order given:
* Chord compression stiffness (25-50t);
* shear stiffness under vertical load 58t, outer layers active;
* chord compression stiffness (50-150t);
* tests (i) and (ii) were repeated after increasing intervals.
The shear stiffness was checked at 25 and 32 mm; one of these, probably the former, is thought to be the shear displacement for which the original stiffness value is quoted. The shear test was carried out both before and after the large load compression measurement because the order of the original tests is not known. Strain history effects may influence the data. Further shear tests under the vertical load of 58t were performed on bearing 4 over a period of two years. At the beginning of the present series of tests the bearings were 38 years old.
Observation of the edge of a bearing during shear testing indicated movement between it and the load platen. Shear stiffness results initially reported (ref. 1) contained a correction for this apparent slip. In the later tests it was realized that the movement was not permanent, but reversed during unloading. Since any slip could only have been partial, it was felt that no correction should be applied. The shear stiffness results reported here for bearings tested with the outer layer active contain no slip correction.
Because of doubt about the influence of any partial slip on the observed shear stiffnesses, a series of tests on bearing 4 was performed with end-plates attached to it by dowels located in the holes in the upper and lower internal plates (figure 1). The original end-plates and dowels were used; the former were machined to remove corrosion and give a flat surface. In this test, the top and bottom rubber cover layers are in principle inactive; it was noted, however, that the dowels were slightly undersize (about 2 mm on the diameter). The bearing was 40 years old at the time of these tests.
Tests on sections cut from bearing 1
Figure 2 shows the sections into which bearing 1 was cut for further testing.
[Figure 2 ILLUSTRATION OMITTED]
The sections B, E, H and G were tested in shear to 25 mm displacement under a vertical load of 60 kN with the intention of seeing whether the stiffness varied between blocks at the corner and the center of the complete bearing. Double shear test pieces were prepared from: (a) center of comer block; (b) close to center of bearing; and pulled to failure to enable the shear modulus and ultimate strain to be determined. The rubber was approximately 80 mm long and 20 mm wide. The test pieces comprised the central rubber layer and an adjacent one.
The bond strength was assessed by a series of eight 90 [degrees] peel tests on samples (width approximately 20 mm) cut from the bearing. The test pieces 1 to 6 were cut from section F, as shown in figure 3. Numbers 7 and 8 were from near the centre of the whole bearing; their peel direction was in the opposite sense (relative to the original bearing) to those of numbers 4 to 6. The interface peeled in all cases was that between the second (internal) rubber layer and the third reinforcing plate.
[Figure 3 ILLUSTRATION OMITTED]
A rubber strip (15 mm thick x 95 mm wide; designated sample 2) was cut with a sharp blade lubricated with talc from the central rubber layer of section F. At certain distances from the outer surface, 2 mm thick slices were cut from the strip. The surface of the slices was ground using fine silicone carbide paper with talc to remove debris; this ensured the surfaces were free of flaws. Dumbbell test pieces (Type 2 BS903: Part A2) were die-stamped from the slices. These were tested (according to the above British Standard) and the modulus at 100% elongation, the tensile strength and the elongation at break noted. Close to the outer surface the slices were adjacent to each other to ensure that any variation of the properties in this region would be detected.
Using similar techniques, additional strips of rubber were cut from the central layer of section B (sample designated 5) and section E (sample designated 4). There was a slight lateral offset between the location of these two samples, but they will be treated as a continuous strip extending in from the center of the short side of the bearing. Slices were cut and ground to give 2 mm thick sheets whose plane was perpendicular to the outer surface of section B. Dumbbell test pieces (type 2) were cut from the slices, two at each of the following distances: 50, 150, 250 and 350 mm; where this is measured to the center of the length of the dumbbell.
Finally, a set of three 2 mm thick dumbbell (type 2) test pieces was prepared from a slice cut parallel to and including the outer surface of section B. The dumbbells were oriented vertically with respect to the bearing and centered on the middle rubber layer.
Several series of hardness measurements were made with one particular aim being to see whether there was any variation of hardness with depth into the bearing. The results were found to be sensitive to the roughness of the surface obtained by sectioning and the presence of any soft debris. The test surfaces were produced by the cutting and polishing technique described in the previous section. Care was taken to avoid raising the surface temperature of the rubber, and to remove any sticky debris with the talc.
The tests were all performed on rubber in the central bearing layer, except where indicated. In several cases, the test planes prepared were parallel to the adjacent outer surface of the bearing; this avoided sample edge effects for measurements close to the bearing surface. Where the test plane was perpendicular to the adjacent outer surface, that surface was pressed against another block of rubber to avoid the edge affecting the readings.
Six sets of measurements (designated a-f) were obtained as follows:
* (a) Microhardness (IRHD) on the slices cut from section F for the tensile test pieces (sample 2). Eight readings were made at each depth and the average taken.
* (b) Microhardness (IRHD) on rubber strip (sample 1) obtained from the central rubber layer of section D. The test surfaces, planes cut parallel to the outside surface of D, were at every 2 mm between 0 to 8 mm and 120 to 128 mm from the bearing surface.
* (c) Shore A on planes obtained from section B (sample 5) and section E (sample 4) and prepared perpendicular to the outside surface of B. There was a slight offset between the location of the surfaces from the two sections.
* (d) Shore A on the surfaces detailed in (c). The measurements were concentrated near the surface and at a depth of 300 mm to compare the hardness in these two zones.
* (e) Shore A on four test planes prepared from section C; the planes were at distances of 40, 80, 120 and 160 mm from the corner of section C and were cut parallel to the short side of the bearing. Measurements were made on all five rubber layers over distances up to 16 mm from the bearing outer surface. The intention was to look for any variation of hardness with depth in the surface region.
* (f) Microhardness (IRHD) on the dumbbells cut from sections B and E for tensile tests. Measurements (the average of three per test piece) were made on the three dumbbells prepared at the surface and the two at each of the distances 50 and 150 mm.
Mechanical test results
The stiffness results obtained from bearings 1 and 4 with the outer layer active are summarized in tables 3 and 4. The compression force-deflection plot to 200 metric tons for bearing 4 is shown in figure 4. The shear force-deflection plots (figure 5) indicate a slight softening with increasing displacement; the stiffness at 32 mm is about 3% less than the value for 25 mm given in the tables. The stiffnesses measured after the application of 200t vertical load are 3 to 5% less than the initial values, and show no sign of recovery in tests several months later.
Table 3 - stiffness of bearing 1 Time, hours - 0 0.5 2.5 19.5 Compression 440 - 420 420 430 stiffness, kN/mm (25-50t) Compression - 550 - - - stiffness, kN/mm (50-200t) Shear stiffness, 3.7 - 3.5 3.5 3.5 kN/mm (25 mm) Table 4 - stiffness of bearing 4 Time, hours - 0 2 70 2 mon. Compression 520 - 490 500 - stiffness, kN/mm (25-50t) Compression - 590 - - - stiffness, kN/mm (50-200t) Shear stiffness, 4.2 - 4.1 4.0 4.0 kN/mm (25mm)
[Figure 4 and 5 ILLUSTRATION OMITTED]
The compression stiffness (50-200t) of 570 kN/mm observed as the average of results from the two bearings compares with the original observation (table 2) of 470 kN/mm. The precise load limits for the latter are not known and because of the pronounced lead-in to the force-deflection curve (figure 4) the value will be sensitive to the lower limit. The comparison cannot, therefore, be taken as a good indicator of any stiffening over service life. For this, the shear stiffness should be much more reliable. In estimating any increase it is conservative to consider the shear stiffness measured at 25 mm and before the softening produced by the 200 metric ton test. The average of the stiffnesses observed for bearings 1 and 4 -3.95 kN/mm - is only 7% above the original prototype value of 3.7 kN/mm; the difference is of the same order as the variability between the two test bearings (1 and 4).
The results of the tests on bearing 4 with end-plates attached by dowels are given in table 5. The 200 metric ton test appears this time to have had no significant effect on the shear stiffness observed. The values at 32 mm were on average only 1% less than those for 25 mm, so the non-linearity was negligible. For comparison with the earlier shear stiffness data, the value needs to be adjusted by the ratio (total thickness of internal rubber layers/total thickness of rubber =) 0.86 to give a stiffness of 3.8 kN/mm. The somewhat low figure compared with that (4.0 kN/mm) in table 4 is consistent with the slightly undersize dowels.
Table 5 - stiffness of bearing 4 with endplates attached Time, hours - 0 0.5 1.3 19 Compression 500 - 470 470 470 stiffness, kN/mm (25-50t) Compression - 610 - - - stiffness, kN/mm (50-200t) Shear stiffness, 4.4 - 4.4 4.4 4.3 kN/mm (25mm)
Tests on samples from bearing 1
The results of the shear tests on the cut sections from bearing 1 are listed in table 6. Each section was restrained laterally only by the friction between it and the platens; no correction for slip is made. There is no significant difference between the stiffness of the central (E) and corner (G) sections, suggesting that the surface aging effects during service have not influenced the results significantly.
Table 6 - shear stiffness of bearing 1 sections Section B E G H Shear stiffness kN/mm (25 mm) 0.35 0.34 0.35 0.32
The results from the double-shear test pieces are given in table 7. The rubber modulus (given at a rubber shear strain [24%] equivalent to 25 mm shear displacement of the bearings) is the same for two locations. The shear failure strain is modest, but acceptable for a bridge bearing.
Table 7 - shear tests on small samples Location Center Corner G, MPa 1.51 1.54 Shear failure strain, % 330
Table 8 gives the results of the peel tests, including an indication of the locus of failure. The peel strength for the sample (4) near the surface is particularly low. It is interesting that the results on test pieces 1 to 3 are consistently higher than those oriented perpendicularly. The latter do not meet current requirements (7 N/mm). The low value from sample 4 possibly results from in service aging near the surface of the bearing, but the results for other test pieces in that orientation are similar whether they are near the surface (sample 5) or the center of the bearing (samples 7 and 8).
Table 8 - rubber-metal bond peel energies Test piece 1 2 3 4 Peel energy, N/m 7.1 10.2 8.8 2.7 Type of failure RC+ R(*)(90%) RC RC Test piece 5 6 7 8 Peel energy, N/m 4.2 5.6 5.4 6.4 Type of failure RC RC RC RC
(+) rubber-cement failure; (*) failure within rubber
Plots of the tensile strength, elongation at break and 100% modulus are given in figures 6, 7 and 8 as a function of depth beneath the bearing outer surface. The results from samples 2 and 4/5 are combined on the same plot; there is no sign of any significant difference between data from the two locations. (There is not a one-to-one correspondence between the test data plotted in the three figures because the followers used to monitor the elongation-at-break failed in some cases). The tensile strength values are 16-18 MPa at depths over 20 mm; the results from sample 2 show rather more scatter on the low side, possibly due to, a slightly poorer surface finish. Close to the surface (depths [is less than] 10 mm) there is a significant drop in strength. The elongation-at-break is about 450% at depths over 20 mm; there is a fall near the surface as seen for the tensile strength. The 100% modulus data indicate a rise of up to 10-15% above the bulk value in the first 30 mm of rubber.
[Figures 6-8 ILLUSTRATION OMITTED]
The results of the microhardness measurements on sample 2 (test a: Section on hardness tests) (figure 9) suggest a three unit rise between a depth of 20 mm and the surface. The measurements on sample 1 (test b), however, show no significant difference between the readings at depths of 0 to 8 mm and those at 120 to 128 mm; an average hardness of 66 IRHD was found in each case. The hardness data from samples 4 and 5 (test c) (figure 10) give a constant value at about 67 Shore A, except for an indication of a one or two point rise at depths less than 30 mm. Microhardness data on the dumbbell test pieces prepared in this location (test f) support the presence of a slight rise at the surface. Further measurements on sample 5 (test d) concentrated at distances of 0 to 40 mm did not, however, show a systematic trend with depth. The mean value is about one point higher than that from tests concentrated at distances of 280-320 mm. The scatter is somewhat greater in the surface region. The hardness results (test e) obtained from the surface region of all five rubber layers in section C are shown in figure 11. Considering these data as a whole, it appears that there is little evidence of a consistent, systematic decrease with increasing depth.
[Figures 9-11 ILLUSTRATION OMITTED]
Analytical tests and results
A strip of rubber from the central layer of sections B and E was cut out and samples taken at the following distances from the external surface of section B:
Sample: s w x y z Distance, mm: 0 20 50 200 250
Other samples were cut from strips taken from the central rubber layer extending in from the surface of section D. A1 and A2 were at depths of 5 and 50 mm, respectively; samples B1 to B8 were from points, respectively, 2, 4, 6, 8, 12, 16, 20 and 50 mm beneath the surface. Samples A3 and B9 were taken from approximately the center of the bearing.
Rubber and filler
Analysis of trace chemicals indicates the polymer is natural rubber. Thermogravimetric analysis of samples S and Z both gave the carbon black and inorganic filler levels shown in table 1. SEM x-ray spectra suggest the latter is zinc oxide; the diameter of the black particles was 260 nm, a figure suggesting MT type black rather than lampblack (diameter 120 nm) as in the MRPRA recommended formulation.
Table 1 gives the level of combined sulfur determined from the analysis. Trace quantities of the secondary accelerator ZDBC were detected, but otherwise no other components of the vulcanizing system were found, consistent with the use of a sulfenamide accelerator such as CBS.
Rubber from samples A1 to A3 was swollen in deuterochlorofonn and subjected to tests using [sup.1]H NMR to assess crosslink density from the broadening of the olefinic peak (ref. 7). Because this is influenced by rubber-black interaction, as well as crosslinking, it is difficult to produce an absolute figure for the crosslink density. The degree of crosslinking found was very similar for the three samples, and quite high, consistent with the compound's high shear modulus of 1.5 MPa and the presence of 50 pphr of non-reinforcing black.
The proportion of polysulfidic ([is greater than or equal to] 3) links was determined by repeating the NMR tests on rubber from samples A1 to A3 treated with a chemical probe (a solution of propane-2-thiol and piperidine in n-heptane); the treatment reduces the polysulfidic to mono- and di-sulfidic links (ref. 8). The degree of crosslinking was reduced to about one-third in all cases. The two-thirds proportion of polylsuffidic links implied suggests that the vulcanizing system was a conventional one. The proportion of such links to be expected from a conventional cure depends upon the cure temperature. The present two-thirds figure is consistent with a cure at about 140 [degrees] C - somewhat high for a large bearing. A lower temperature gives a higher proportion, implying some maturation of the links, leading to stiffening of the compound, has occurred during service. The bearing tests only support the presence of slight stiffening.
The acetone extract from rubber from samples s, w, x, y and z was analyzed to identify the antidegradants used. The results from the latter three samples show no significant variation with depth and the average wt % levels are: DOPD - 1.3; PBN - 0.026; diphenylamine-acetone condensate - 0.099.
The last two are present in extremely low quantities. An unlikely low level of addition to the original compound or very high rate of reaction during service are implied. Allowing for about 25 % consumption of the DOPD during vulcanization of the bearing, the analysis gives a figure of [is greater than] 3 pphr in the formulation; some material will have been consumed during service.
Analysis did not reveal the presence of any antiozonant wax. The DOPD would have functioned as an antiozonant.
Changes near surface
The amount of acetone extractable material (which includes soluble material in the natural rubber and added antidegradants) showed a spread within the data from 5.0 to 6.2 wt %; given that spread, no trend with depth in the amount of total extract could be identified. According to HPLC analysis of the extract, however, the level of DOPD decreased markedly from the bulk value between a depth of 20/50 mm and the surface. Near the surface, the fall must represent consumption in oxidative reactions. Away from the surface, the fall may only be due to diffusion towards the surface, and thus the depth of the depleted zone is not necessarily equivalent to the depth of oxidative reactions. The missing material was obviously not fully replenished by diffusion of DOPD from the bulk. The trace levels of PBN and diphenyl-amine acetone condensate did not decrease significantly near the surface.
The figure for the percentage of nitrogen in the samples (which is contained in proteinaceous material and the antidegradants) varied widely between 0.41 and 0.56 wt %. Within that spread, no significant dependence on depth was noted. After acetone extraction, the nitrogen content at depths greater than 20 mm appeared constant at an average of 0.36 wt %. The figure for samples near the surface (depth [is less than] 6 mm) averaged a significantly higher value - 0.46 wt %. The difference is presumably due to bound reacted anti-degradant; the increase of 0.1 wt % is equivalent to about 2 pphr of DOPD having reacted in the surface region.
Two additional analytical tests were performed to assess the magnitude of any oxidation of the natural rubber chain. The results for the glass transition temperature ([T.sub.g]) gave an average value of -69.0 [degrees] C. The figure is that expected for vulcanized natural rubber; there was no sign of a significant rise due to oxidative modification of the chain, even in the surface region. The other test measured the amount of oxygen in the vulcanizate. The technique, using a Perkin Elmer Model 240 Analyzer operating in "oxygen mode," measures the amount of oxygen evolving during pyrolysis of the sample; the inorganic oxygen content (e.g., that in the zinc oxide) is therefore not included. The results indicate a level in the bulk of the bearing comparable to that expected from natural rubber non-hydrocarbon material. Most of the rise as the surface is approached occurs at depths less than 6 mm; the extra 1 wt % of oxygen close to the surface indicates a moderate level of oxidation there.
The analytical tests taken together suggest that significant oxidative activity has been confined to depths less than 10-20 mm. This is consistent with the observations that the drop in tensile strength and elongation at break is confined to a narrow region of less than 10 mm.
The average of the shear stiffnesses of the two Pelham Bridge bearings tested after 38 years service is only 7% greater than the stiffness observed for an original prototype; the figure is less than the variation in stiffness between the two removed bearings. Mechanical tests on rubber specimens prepared from a sectioned bearing show the tensile properties still to be acceptable in the bulk of the bearing; a drop of no more than 30-40% in strength and elongation at break is seen at depths [is less than] 10 mm from the surface. Tensile modulus data suggests a slight increase relative to the bulk begins at depths of about 30 mm as the surface is approached and reaches a maximum of 10-15%; such stiffening, however, was not always detected from the hardness measurements near the surface. Analytical tests show that the presence of significant rubber oxidation and reactions involving the antidegradants are confined within a surface zone extending to depths of only 10-20 mm. The mechanical and analytical tests on rubber samples confirm that aerobic changes were restricted to relatively close to the surface of the bearings.
(1.) Fuller, K.N.G., Ahmadi, H.R. and Pond, T.J. (1996) High Damping Natural Rubber Seismic Isolators: Longevity and Modeling. Proceedings PVP-ICPVT Pressure Vessel and Piping Conference, Montreal, ASME.
(2.) Davies, B. (1989) The Longest Serving Polymer. Mechanical Incorporated Engineer, 1, 91.
(3.) Kato, M., Watanabe, Y., Yoneda, G. et al. (1996) Investigation of Aging Effects for Laminated Rubber Bearings of Pelham Bridge. Proc. 11th World Conference on Earthquake Engineering, Acapulco, Mexico.
(4.) Kato, M., Watanabe, Y., Kato, A. et al. (1997) Aging Effects on Laminated Rubber Bearings of Pelham Bridge. Proc. of 14th International Conference on Structural Mechanics in Reactor Technology, Session K.
(5.) Gent, A.N. (1959) Rubber Bearings for Bridge & Design Consideration. Rubber J. International Plastics, 137, 426.
(6.) Anon (1957) The Use of Rubber in Bridge Bearings. Rubber Developments, 10, 34-37.
(7.) Brown, P.S. and Tinker, A.J. (1995) The Use of FT-NMR in the Analysis of Rubber Blends: Crosslink distribution in carbon black filled blend of NR and cis BR. Kautschuk Gummi Kunstoffe, 48, 606.
(8.) Saville, B. and Watson, A.A. (1967) Structural characterization of sulfur-vulcanized rubber networks. Rubber Chem. Technol., 40, 100.
K.N.G. Fuller and A.D. Roberts, Tun Abdul Razak Research Centre
|Printer friendly Cite/link Email Feedback|
|Article Type:||Statistical Data Included|
|Date:||Dec 1, 1999|
|Previous Article:||PNR for paper roller application.|
|Next Article:||Thermoplastic rubber as a shoe soling.|