___.The main qualities of a direct-bonding bracket reside in the characteristics of the slot and the retention of its base. The reason clinicians prefer more expensive brackets is because these have mesh bases. Less expensive, one-piece or monoblock ones can offer everything but the last feature. In recent years, manufacturers have tried their best to come up with inexpensive injection mold-made brackets, which were claimed to provide bond strengths similar to the ones endowed with mesh pads. Among the latter, those having the very fine 100-mesh pad (one hundred openings per inch) have become standard.
___.Being able to pinpoint specific inexpensive brackets capable of acceptable bonding strengths translates in saving about half the amount one spends for attachments. Unfortunately, the few independent reports published once in a while cannot keep up with the flood of new attachments launched almost each year on the market, and guesswork doesn’t help. Take, for example, the bases of the one-piece brackets: for technical reasons, it is not feasible to have their locking grooves both deep and numerous. As a result (see bases shown on the back page), the manufacturer makes choices that translate in a multitude of designs.
___.The best available evaluation of the attachments’ bond strength is performed by specialized labs and involves trained technicians, acrylic-mounted bovine teeth and adequate testing equipment. The latter, a universal tensile strength tester, is a large machine such as the one shown in Fig. 1. Sold by companies such as Tinius Olsen, Chatillon or Instron, such instruments are priced around $30,000.

___.A standard test providing statistically reliable data, or even a comparison with a well-known brand one, may run into the thousands of dollars. Aside from the cost, the test itself is basically inaccurate. Indeed, teeth are far from exhibiting standard features, and the bracket pad seldom matches their contour. In addition, being too small, teeth have to be mounted in acrylic blocks. While comparisons give significant results, the method used is cumbersome and requires sophisticated and expensive equipment.
___.To simplify the test and transform it in an easy, do-it-yourself method, instead of measuring the debonding forces, in what follows we used a variation of the Adhesive Remnant Index (ARI)⁵ as criteria. Indeed, if both adhesive and substrate are the same, and the last has been properly selected, the examination of the debonded attachment and especially what is left on the substrate can lead to an evaluation of the bonding ability of a particular base.
Materials and method
___.An attempt to find an already made, easy to acquire substrate, which could provide always a standard surface, failed. Stainless steel mesh-laminated sheets were tried using the Elcometer tester F106/1 (Paul N. Gardner Co., Gardner Bldg., 316 NE First St., Pompano Beach, FL 3306), Fig. 3. The supplied aluminum dollies, designed for testing coating adhesion, are cemented to the surface of the coated substrate with a cyanoacrylate or epoxy adhesive. After the adhesive is cured, the claw of the tester is placed under the dolly head and slowly pulled by turning the spanner nut at the top of the instrument. At some point, the bond gives up and the test poundage is read on the scale indicator.
___.As is, the tester couldn’t be used for testing attachments’ adhesion: the existing calibrated spring didn’t match the force needed to debond these and the dollies were inadequate for holding the brackets. As a result, the dollies were replaced with an adjustable bracket holding clamp, and the spring with another one fitting the range of the debonding forces encountered.

___.The system is presented in Fig. 4 and the adjustable clamp in Fig. 5. A portion of the laminated mesh with bonded brackets is shown in Fig. 6. Despite the reliability of the system and the claims of the seller, for practical purposes we preferred to compare brackets with one another, instead of providing force values.
___.While the 100 mesh laminated to stainless steel sheets may actual offer a good substrate for measuring debonding forces, it was not appropriate for a do-it-yourself test. In almost all cases, the adhesive remained on the substrate.
___.The second substrate tried were tiles made of marble, a form of calcium carbonate that has a relatively similar composition throughout the world. To increase roughness, their finished side were subjected to a short treatment with diluted hydrochloric -muriatic (HCl) acid. Various brackets were bonded on its surface, as shown in Fig. 8. This time, the results were the reverse of our previous attempt: the acid attack, instead of producing micro-pits and fissures, generated irregular but smooth surfaces which didn’t hold the bonded bracket.
___.A similar bonding test, using as substrate etched glass, commonly used in photography to avoid glare, also failed as all the debonded brackets retained chunks of the substrate.
___.In a fourth and final attempt, we tried ceramic tiles used for floors. These exhibit both a glazed, smooth surface, as well as a rough one which has built in indentations to retain the adhesive. While perhaps adequate as a mechanical interlocking surface, the rough side does not lend itself to the test due to its uneven profile. The glazed side, however, if properly treated with hydrofluoric acid, leads to a surface rich in micro-pits and micro-fissures that were thought to offer a surface similar to the etched enamel.
___.After testing several tiles, we selected the 8x8” (20x20 cm) Americer Ceramic Floor black tiles 8310 H I-RH (Home Depot). These are not only the least expensive ones, but also quite common.
___.After washing two tiles with water and detergent, and then rinsing and drying them, their glazed surface was degreased with amyl acetate and dried again. On each tile, 2 ml hydrofluoric acid 48% (Matheson, Coleman & Bell, Norwood, OH) was spread, and then the wet surfaces were covered with a plastic film to reduce evaporation. A tile has been subjected to the etching for an hour, while the other for 24. After washing and drying, the exposed surfaces became somewhat whitish and rough. After being rubbed with some mineral oil, these became again black, enhancing the traces left after bracket debonding.
___.For each brand tested, ten new brackets having flat bases and approximately the same size were selected and degreased with amyl acetate. Their bases were covered with an excess of the initiator-containing part (“Base”) of a two-paste adhesive ( Maxibond, Ortho-Source, Hollywood, CA) and then pressed over a particular zone of each tile. The whole was heated for half an hour in a microwave oven, achieving thus a homogeneous and complete polymerization. After cooling, the brackets were subjected to debonding.
___.To duplicate simple means available to any office, instead of using for debonding the tensile strength provided by the Elcometer, we chose first to use the adjustable clamp, Fig. 9, and later a a common pincer. After seizing the bracket, the pincer was gently tilted on its side until debonding took place. After debonding, both the bracket pad and the trace left on the tile were examined to quantify the adhesive left behind: the latter proved to be enough. Instead of the Adhesive Remnant Index (ARI) as defined by Artun⁵, which starts from 0 (no adhesive left on the enamel), for our purpose we chose the Modified Remnant Index (MARI) where for the same situation the value of 1 is assigned⁶:
__1. All the adhesive is removed from the enamel
__2. Less than half of it has remained on the enamel
__3. More than half of it has remained on the enamel
__4. All the adhesive has remained on the enamel
___.A typical view of the tile surface after the debonding of the brackets is shown in Fig. 10 which is the area where ten Ortho-Organizer’s Edgeway brackets once stood (site 4, Fig. 14). According to MARI, the sites where the glaze was torn without leaving any adhesive were marked with (1). Sites where no less than half of the adhesive was left on the surface were marked with (2); and sites where practically all the adhesive remained on the tile, with (4). These cases were given values corresponding to the MARI values shown below:


Results
___.Ten brackets each from twenty-three brands were tested twice. A first test was performed on the tile that was etched only an hour. A second one was performed under the same conditions on the tile exposed to acid attack for 24 hours. In the first case, the penetration of the adhesive in the tile was too superficial, Fig. 12, only seldom showing the tearing of glaze chunks commonly found on the second tile, Fig. 14.
___.Marking the sites found on the second tile according to MARI, the values shown in Table 1 were obtained. The only difference encountered was the unexpected fact that in a few instances, the glaze was not affected and there was no adhesive left on the tile. While according to MARI, these sites (7 & 8) could have been marked with 1, we preferred not to count them, calculating instead the average for the other sites that qualified. The reason behind this measure was the possibility that, in comparison with the other 228 sites, in these two cases, the glaze may have been stronger or the acid attack weaker.
___.In examining Table 1, it is obvious that the pads with mesh and MicrolocRbases (GAC) performed better than the ones relying on dents, protrusions and grooves. As variations may have occurred between tests, it is fair to consider that the bases summing 12 or less (shown as ratings) showed an acceptable retention. It was surprising to see that Unitek’s Twin Torque brackets rated low: the reason may well reside in the tack welding points used to join the mesh to the bases mesh, a fact consistent with research done by Maijer and Smith⁷. Another observation is that the bases exhibiting protrusions adhered better than those having grooves or dents: GAC Microloc^R bases may be an exception. Micro-etched and multiple mesh bases showed also better bonding, confirming previous research⁸.
___.Surprisingly, a larger mesh size base (American Orthodontics, #21) gave as good results as the fine mesh preferred today, confirming previous studies^9-14. Bases with smaller surfaces (mini) presented a performance similar to those having a medium size, confirming previous research stating that bond strength is independent of nominal area and mesh size of the bases¹⁴.
Discussions
___.Disclosing the bond strength of the many various bases sold today is economically important, as the price of some attachments can be as high as five times that of others which may have identical properties. Unfortunately, the measurement of bond strength is costly, difficult and inaccurate.
___.As the economic part has already been presented, we will focus on the other problems. Indeed, while attempts were made in the past to replace the traditional acrylic embedded bovine teeth with other substrates⁴, a literature search shows that the method was not adopted by subsequent studies. The concept, however, is sound as long as the adhesion is based upon mechanical interlocking and not on chemical affinity.
___.Aside from the fact that the bovine teeth enamel is weaker than the human one^1,2, it is unlikely that any bracket base will evenly match the profile of the tooth to which it should be attached. As a result of the variance in adhesive thickness, the related bond strength is affected, as shown both by Matasa¹² and Evans¹³. In contrast, in the bracket base-tile case, the distance between the base and substrate is even throughout the experiment, as is also the adhesive system used .
Conclusions
___.In the present research no attempts were made to determine debonding forces. For the practitioner’s daily use, however, it may be enough to know that a certain new brand he feels tempted to use, exhibits bonding strengths equal to that of a brand he is familiar with, and which is more expensive or doesn’t have other desirable features.
___.No matter if for debonding one uses pincers, Elcometers or Universal Testing Machines, the method proposed relies upon a more reliable substrate than teeth. It also steers away from mixing chemical affinity and mechanical locking concepts, as it considers only the latter as governing (hydrophobic) resin bonding.
___.By using heat-activated polymerization, it allows a departure from the individual (and often uneven) bonding. Indeed, chemical and photochemical polymerization of the adhesive varies from teeth to teeth due to slight differencesin the part ratios (concentration), timing, exposure, etc.
___.Instead of testing bonding bases, as explored in this study, the method can be used to test adhesives, of course keeping as a constant the brackets used.

References
1. Oesterle LJ, Shellhart WC, The use of bovine enamel in bonding studies Belanger GK, Am J Orthod Dentofacial Orthop 1998: 114; 514-9
2. Barkmeier WW, Erickson RL. Shear bond strength of composite to enamel and dentin using the Scotchbond multi-purpose adhesive, Am J Dent. 1994; 7: 175-9.
3. Matasa CG, The poor man’s tensile strength, Phoenix without Ashes, April 1993, 6(1); 6-7
4. Chung CH, Brendlinger EJ, Brendlinger DL, Bernal V, Mante FK Shear bond strengths of two resin-modified glass ionomer cements to porcelain, Am J Orthod Dentofacial Orthop 1999;115:533-5
5.Artun J, Bergland S, Clinical trials with crystal growth conditioning as an alternative to acid-etch enamel pretreatment, Am J Orthod Dentofacial Orthop 1984; 85: 333-340
6. Mimura H, Deguchi T, Obata A, Yamagishi T, Ito, M, Comparison of different bonding materials for debonding, Am J Orthod Dentofacial Orthop 1995; 108: 267-73
7. Maijer R, Smith, DC, Variables influencing the bond strength of metal orthodontic bracket bases, Am J Orthod Dentofacial Orthop. 79: 20-34, 1981
8. Smith DC, Maijer R, Improvements in bracket base design, Am J Orthod Dentofacial Orthop . 1983 Apr, 277-281
9.Reynolds IR, von Fraunhofer, JA, Direct bonding of orthodontic attachments to the teeth: The relation of adhesive bond strength to gauze mesh size, Br. J. Orthod. 1976; 3: 91-95
10. Reynolds IR, von Fraunhofer JA, Direct bonding in orthodontics: A comparison of attachments, Br. J. Orthod. 1977; 4: 65-69
11. Buzzitta VAJ, Hallgren SE, Powers JM, Direct-bonding cement-bracket systems of orthodontic direct-bonding cement-bracket systems as studied in vitro Am J Orthod Dentofacial Orthop.1982 Feb, 87-92
12. Matasa CG, Adhesion and its ten commandments, Am J Orthod Dentofacial Orthop.1989; 95: 355-6
13. Evans LB: Factors affecting bond strength of no-mix orthodontic adhesive systems, Master's thesis, Ann Arbor, 1984, University of Michigan School of Dentistry
14. Dickinson, PT, Powers, JM, Evaluation of fourteen direct-bonding orthodontic bases, Am J Orthod Dentofacial Orthop 1980; 78: 630-639

Do brand new brackets bond better than those recycled
by Ortho-Cycle Co.?

___.Having a simple method allowing to differentiate bond strengths, we were anxious to see how the brackets recycled by us compare with their new counterparts. We bonded to an etched tile, as described in the previous pages, “A” -Co Standard Edgewise laterals. A hundred were brand new (first 5 line on the tile), and a hundred recycled by us. We photographed them before and after debonding: with the exception of few zones (marked), the surface of the tile behaved quite evenly, chunks of glaze being removed in most instances in both cases. The sites were recorded using the MARI score:


___.In contrast with the previous tests, in the present one, a new category (noted with a purple circle) has appeared: indeed, three new brackets detached from their bases (these were not “fire-tested” like the recycled ones!). The unexpected, highly desirable low MARI average (1.15 vs. 1.4) exhibited by the recycled brackets may well be due to their extreme cleanliness and the increased mesh roughness caused by wear and treatments.