ORIGINAL ARTICLE

Bond strength of orthodontic brackets with newself-adhesive resin cements

Mohammed Al-Saleha and Omar El-Mowafyb

Toronto, Ontario, Canada

Introduction: In this investigation, we determined the shear bond strength (SBS) of metallic and ceramic or-thodontic brackets with new self-adhesive cements. Methods: One hundred extracted premolars were used.They were sterilized and their roots embedded in stone bases, with the facial surfaces perpendicular to thebottom of the bases. The teeth were divided into 2 main groups, to receive metallic or ceramic brackets (Vic-tory series 3M Unitek, Monrovia, Calif). In each group, the specimens were further divided into 5 subgroupsaccording to the cement used: an etch-and-rinse control, Transbond-XT (3M Unitek); a resin cement withself-etch primer, Esthetic Cement system NC-100, (Kuraray, Okayama, Japan); and 3 self-adhesive resin ce-ments: Rely-X Unicem (3M ESPE, Seefeld, Germany), Biscem DC (Bisco, Schaumburg, Ill), and Breeze (Pen-tron, Wallingford, Conn). Ten brackets were cemented in each subgroup. The specimens were stored indistilled water at 37�C for 7 days and subjected to 3000 thermocycles between 5�C and 55�C. The bracketswere then debonded in shear with a testing machine. Results: Mean values for the metallic brackets cementedwith Transbond XT, Esthetic Cement system, Rely-X Unicem, Biscem DC, and Breeze were 18.6, 6.0, 6.0, 2.2,and 8.4 MPa, respectively. For the ceramic brackets, the values were 22.7, 17, 7.7, 1.6, and 9.5 MPa, respec-tively. Analysis of variance (ANOVA) showed significant differences among the subgroups (P \0.05) for bothbracket types. For the ceramic brackets, the Tukey test showed no statistical difference in mean SBS betweenTransbond XT and Esthetic Cement system. Conclusions: The SBS values of brackets cemented with etch-and-rinse cement were significantly higher than thosse of the 3 self-adhesive cements. However, when theself-etch adhesive, Esthetic Cement system, was used with ceramic brackets, no significant difference wasfound in the SBS compared with Transbond XT (P 5 0.052). (Am J Orthod Dentofacial Orthop2010;137:528-33)

Enamel bonding for orthodontic applications wasintroduced in 1965 and is considered a signifi-cant milestone in orthodontic treatment.1 As re-

ported by Owens and Miller,2 direct bonding oforthodontic brackets to enamel was made a reality byBuonocore, Bowen and Tavas. Adhesion occursprimarily through microscopic interlocking betweenthe adherent and the brackets with the adhesive.

Traditional adhesive systems used for bonding or-thodontic brackets require a clean enamel surface tobe etched with phosphoric acid for 30 to 60 seconds,rinsed thoroughly with water, dried, and coated withresin adhesive before bracket bonding with resin

From the Department of Clinical Sciences, Faculty of Dentistry, University of

Toronto, Toronto, Ontario, Canada.aPostgraduate student.bProfessor.

The authors report no commercial, proprietary, or financial interest in the prod-

ucts or companies described in this article.

Reprint requests to: Mohammed Al-Saleh, Department of Clinical Sciences,

Faculty of Dentistry, University of Toronto, 124 Edward St, Toronto, Ontario

M5G 1G6, Cananda; e-mail, m.alsaleh@utoronto.ca.

Submitted, March 2008; revised and accepted, April 2008.

0889-5406/$36.00

Copyright � 2010 by the American Association of Orthodontists.

doi:10.1016/j.ajodo.2008.04.027

528

cement. Although this is a relatively simple and effec-tive technique, the need to etch, rinse, and dry theenamel while maintaining an uncontaminated environ-ment can be challenging in some clinical situations, es-pecially when patient compliance is an issue.Contamination of the etched enamel before bracketbonding will negatively affect bond strength of thebracket, and this might compromise the stability ofthe appliance. Another disadvantage of traditional adhe-sive systems is that some of the etched enamel surfacemight not have adhesive protection, leading to deminer-alized enamel white spots.3 This will undermine the cli-nician’s primary concern to maintain a sound,unblemished enamel surface after removal of thebracket.

Studies showed that cements with high bondstrength to enamel can result in enamel fracture duringdebonding.4 Alternatively, considerable amounts of ad-hesive can remain on the tooth surfaces after debondingand require more chair time for removal. In addition, theprocess of removing the residual adhesive might resultin enamel loss.5

New self-adhesive cements have been introducedrecently to orthodontics. These could simplify the

American Journal of Orthodontics and Dentofacial Orthopedics Al-Saleh and El-Mowafy 529Volume 137, Number 4

bonding process by reducing the bonding steps andeliminating the need for etching and priming, thus less-ening the risk of contamination. Although these mate-rials are manufactured primarily for cementation ofcrowns, fixed partial dentures, and inlays or onlays, theirpotential use with orthodontic brackets has not yet beenfully explored. Since they are directly applied to toothsurfaces, they might reduce the risk of unnecessaryenamel etching with subsequent demineralization.Also during bracket debonding, the new self-etching ce-ments can reduce the risk of enamel fracture if theirbond strength to enamel is lower than that of conven-tional cement. Vicente et al6 found that Rely-X Unicem,a self-adhesive resin cement for bonding orthodonticbrackets, produced a shear bond strength (SBS) thatwas significantly lower than that of a conventionaletch-and-rinse cement after storing the specimens inwater for 24 hours at 37�C. However, they speculatedthat the lower SBS of Rely-X Unicem would be clini-cally sufficient for the normal time that the bracket isto be attached. To closely simulate the clinical situation,Bishara et al7 determined the SBS of metallic bracketsbonded with Rely-X Unicem 30 minutes after bonding.This is typically the time between bonding the bracketsto the teeth and ligation of the archwires. They con-cluded that the SBS of this self-adhesive resin cementwas insufficient for successfully bonding orthodonticbrackets.

Other new self-adhesive resin cements are alsoavailable. The purpose of this study was to determinethe SBS of 2 types of brackets, metallic and ceramic,with 3 self-adhesive resin cements. The null hypothesiswas that there is no significant difference in the SBSvalues of the 2 types of cements. The effects of waterstorage and thermocycling on SBS were also investi-gated. A conventional etch-and-rinse cement and a ce-ment with a self-etch primer were used as controls forcomparison.

MATERIAL AND METHODS

One hundred extracted premolars with intact buccalenamel surfaces were used. The teeth were cleaned andsterilized with gamma irradiation and stored in distilledwater. A split mold was used to mount the roots in dentalstone bases with the buccal surfaces aligned perpendic-ular to the bottom of the mold. The buccal surfaces werecleaned with coarse pumice and rubber prophylacticcups for 10 seconds and then rinsed and dried with anair-water syringe.

One hundred maxillary premolar brackets with 7� oftorque, no angulations, and 0.022-in archwire slots wereused; 50 were metallic, and the other 50 were ceramic

(Clarity, Victory series, 3M Unitek, Monrovia, Calif).The surface areas of the bracket bases were determinedwith a digital caliper (10.6 mm2 for the metallicbrackets and 11.9 mm2 for the ceramic brackets). Theteeth were divided into 2 equal groups according tothe type of brackets (metallic and ceramic). In eachgroup, the specimens were divided into 5 subgroups ac-cording to the resin cement used (n 5 10). These werean etch-and-rinse control, Transbond-XT LC adhesivesystem (3M Unitek) (TBXT), a resin cement withself-etch primer, Esthetic Cement system NC-100(Kuraray, Okayama, Japan) (ECS), and 3 self-adhesiveresin cements: Rely-X Unicem (3M ESPE, Seefeld,Germany) (RXU), Biscem DC (Bisco, Schaumburg,Ill) (BCM), and Breeze (Pentron, Wallingford, Conn)(BRZ). The brackets were then bonded to the mountedteeth according to the manufacturers’ instructions.

In the TBXT subgroups, the teeth of 1 subgroup ofthe metallic and 1 subgroup of the ceramic bracketswere etched with 35% phosphoric acid gel for 15 sec-onds. The etchant was applied at the center of the mid-dle third of the buccal surface. The teeth were thenthoroughly rinsed with water and dried, and a layer ofTBXT primer/sealant was applied to the etched area.TBXT adhesive paste was then applied to the bracketbase and placed on the tooth. The bracket was pressedfirmly for 10 seconds to ensure uniform adhesive thick-ness. Excess adhesive was removed with a microbrush,and the cement was cured with a halogen light (Spec-trum 800 Curing Unit, Dentsply, York, Pa) for 20 sec-onds (10 seconds from each proximal side).

In the ECS subgroups, the brackets of 2 subgroupssimilar to the above were bonded with ECS. Equalamounts of ECS bond liquids A and B were mixedand applied to tooth surfaces and left for 20 seconds.The tooth surfaces were then dried and light-cured for20 seconds. ECS was then applied to the bracket, placedon the tooth, and light cured as above.

In the RXU subgroups, the brackets of 2 subgroupssimilar to the above were bonded with RXU. An RXUcapsule was activated in a Maxicap activator (3MESPE) and mixed for 15 seconds in a high frequencymixing unit (Caulk, Dentsply). The capsule was placedin the Maxicap gun and applied to the bracket base. Thebracket was bonded and light cured as above.

In the BCM subgroups, the brackets of 2 subgroupssimilar to the above were bonded with BCM. Cementwas mixed and applied to the bracket base, which wasplaced on the tooth and cured as above.

In the BRZ subgroups, the brackets of 2 subgroupssimilar to the above were bonded with Breeze. Mixedcement was applied to the bracket and placed on thetooth and cured as above.

Table I. SBS of metallic bracket subgroups cementedwith 5 cements

n Mean SD Range

TBXT 10 18.6 4.9 12.3-26.3

ECS 10 6.0 A,B 1.7 3.2-9.3

RXU 10 6.0 A,B 1.5 3.9-8.8

BCM 6 2.2 B 1.0 0.9-3.5

BRZ 8 8.4 A 3.1 3.6-13.6

One-way ANOVA showed a significant difference in mean SBS

among the subgroups (P \0.001). The Tukey test showed that means

with the same letter were not significantly different.

Table II. SBS of ceramic bracket subgroups cementedwith 5 cements

n Mean SD Range

TBXT 10 22.7 A 5.0 16.0-32.1

ECS 10 17.0 A 5.3 8.5-26.1

RXU 9 7.7 B, C 2.5 4.4-12.7

BCM 4 1.6 C 0.4 1.0-1.8

BRZ 10 9.5 B 4.9 3.3-16.9

One-way ANOVA showed a significant difference in mean SBS

among the subgroups (P 5 0.032). The Tukey test showed that means

with the same letter were not significantly different.

530 Al-Saleh and El-Mowafy American Journal of Orthodontics and Dentofacial Orthopedics

April 2010

All bracket placements were carried out by the prin-cipal author (M.A.). The specimens were stored indistilled water at 37�C for 7 days. They were then sub-jected to 3000 thermocycles between 5�C and 55�C.After thermocycling, some brackets separated: 4metallic brackets cemented with BCM and 2 withBRZ, and 6 ceramic brackets cemented with BCMand 1 with RXU. The separated brackets were recorded,and the remaining specimens were subjected to a sheartest (Tables I and II).

The thermocycled brackets were debonded witha shear load applied in a universal testing machine (Ins-tron, Canton, Mass) with a metal chisel at a crossheadspeed of 1 mm per minute. The specimens weremounted on the machine so that the end of the chisel ap-plied a compressive load directly to the incisal aspect ofbracket-tooth interface parallel to the long axis of thebond interface. The maximum loads to debond thebrackets were recorded.

All tests were conducted by the principal author.Each tooth and its corresponding bracket were viewedwith a stereomicroscope (SMZ800, Nikon InstrumentsInc, Melville, NY) and scored according to the adhesiveremnant index (ARI). The values for the ARI are as fol-lows: 0, no adhesive left on the tooth; 1, less than half ofthe adhesive left on the tooth; 2, more than half of theadhesive left on the tooth; and 3, all adhesive left onthe tooth with an impression of the bracket mesh.

Statistical analysis

Data analyses were carried out with SPSS software(version 15.0, SPSS, Chicago, Ill). The data were statis-tically analyzed with 1-way analysis of variance(ANOVA) at P \0.05. Furthermore, the Tukey testwas used to show subgroups with significant differencesin SBS. Two-way ANOVA was used to determinewhether there was an interaction between the 2 vari-ables—bracket material and cement type.

The chi-square test was calculated for the cross-tabulation of cement type and ARI, which was initiallydichotomized into 2 categories (1 5 0 1 1; 2 5 2 1 3).P 5 0.05 was considered statistically significant.

RESULTS

Descriptive statistics (means, standard deviations,maximum and minimum values) for all subgroups areshown in Tables I and II. One-way ANOVA showed sig-nificant differences among the subgroups of eachbracket type. In the metallic bracket subgroups, thehighest SBS (18.6 MPa) was obtained when TBXTwas used, and the lowest value (2.2 MPa) was obtainedwith BCM. The Tukey test comparisons indicated that

mean SBS for TBXT was significantly higher than thoseof the other 4 cements (P\0.001). Mean SBS values forBRZ, RXU, and ECS were not significantly differentfrom one another (P .0.05); however, the mean SBSof BRZ only was significantly higher than that ofBCM (P 5 0.003).

In the ceramic subgroups, the highest SBS (22.7MPa) was also obtained when TBXT was used, andthe lowest value (1.6 MPa) was obtained with BCM.The Tukey test comparisons indicated that the meanvalues of TBXT and ECS with ceramic brackets werenot significantly different. However, they were signifi-cantly higher than those of the other 3 cements (P\0.05). The mean SBS values for RXU and BRZwere not significantly different (P 5 0.907).

Two-way ANOVA showed significant differences ofSBS in the ceramic and metallic subgroups for the var-ious types of cements. Also, a significant interactionbetween the variables was found (P \0.001).

After thermocycling, some brackets separated: 4metallic brackets cemented with BCM and 2 withBRZ, and 6 ceramic brackets cemented with BCMand 1 with RXU (Tables I and II).

The failure modes of the 2 types of brackets areshown in Table III. Chi-square comparisons of theARI values for metallic (13.3) and ceramic (17.6)brackets indicated that different adhesives had

Table III. Frequency distribution of ARI results of thegroups and findings of the statistical analyses

ARI scores

Brackets Group n 0 1 2 3

Metal TBXT 10 1 3 1 5

ECS 10 4 3 3 — c2 5 13.3

RXU 10 6 3 1 — P 5 0.010

BCM 6 6 — — —

BRZ 8 5 3 — —

Ceramic TBXT 10 1 2 2 5

ECS 10 1 1 4 4 c2 5 17.6

RXU 9 2 5 2 — (P \0.001)

BCM 4 4 — — —

BRZ 10 4 2 4 —

American Journal of Orthodontics and Dentofacial Orthopedics Al-Saleh and El-Mowafy 531Volume 137, Number 4

significantly different bracket failure modes (P 5 0.010for metallic and P \0.001 for ceramic). For TBXT,most of the adhesive remained on the tooth (scores 2and 3), indicating failure at the bracket-adhesive inter-face. For the self-adhesive cements, most of the adhe-sive remained on the bracket (scores 0 and 1),indicating failure at the enamel-adhesive interface.

DISCUSSION

These findings indicated that metallic bracketsbonded with RXU had a mean SBS of 6.0 MPa, whichis less than the value reported by Vicente et al6 (8.1MPa) for the same cement. However, in this study, thebrackets were debonded after 3000 thermocycles,whereas Vicente et al debonded them after storageonly in water for 24 hours at 37�C without thermocy-cling. In contrast, Bishara et al7 reported the meanSBS value for metallic brackets bonded with RXU tobe 3.7 6 2.1 MPa when the brackets were debondedwithin 30 minutes from the cement mixing and ata crosshead speed of 5.0 mm per minute. In our study,the specimens were stored for 7 days and then thermo-cycled, which took approximately 2 days, and then wereloaded at 1 mm per minute. This might explain the dif-ferent findings of our study compared with that of Bish-ara et al,7 because they might not have allowed for fullpolymerization of the cement.

Three thousand thermocycles between 5�C and55�C in this study equates to a number of years of in-traoral thermocycling, exceeding the average orthodon-tic treatment term. Thermocycling aims to thermallystress the adhesive joint interface.8 Versluis et al9 high-lighted the effect of a mismatch in thermal coefficient ofexpansion between resin restoration and tooth structure,which results in different volumetric changes duringtemperature fluctuations, causing fatigue of the adhe-sive joints with subsequent microleakage. The same

principle can be extrapolated to the tooth-bracket com-plex; 40% of the metallic brackets and 60% of the ce-ramic brackets bonded with BCM underwentseparation during thermocycling. The bond failuremodes indicated a weak bond between the tooth andthe BCM. In support of this finding, the lowest meanSBS values were obtained (1.6-2.2 MPa) when bothtypes of brackets were bonded with BCM. This is suffi-cient evidence to suggest that this cement should not beused for bracket bonding, because it does not meet theminimum requirement of SBS for this purpose.

It is difficult to determine the magnitude of bondstrength that is required to sustain active orthodontictreatment without bracket failure under oral conditionsfor the duration of the treatment.10 Some studies esti-mated a range of 6.5 to 10 MPa for SBS.11,12 Others sug-gested that a minimum SBS of 8.0 MPa is necessary formaintaining the bond of orthodontic brackets toteeth.13,14 In our study, the mean SBS values for ECS(6.0 MPa), RXU (6.0 MPa), and BCM (2.2 MPa) withthe metallic brackets were below this range, albeit notsignificantly for ECS and RXU.

In contrast, 3 cements used with ceramic bracketshad SBS values within or higher than the 6.5 to 10MPa range: RXU (7.7 MPa), BRZ (9.5 MPa), andECS (17.0 MPa). The high SBS value of ECS with ce-ramic brackets that was not statistically different fromthat of the control cement TBXT rendered it, withouta doubt, suitable for this clinical application, particu-larly since none of the 20 ECS specimens separated af-ter thermocycling. This equates to 1 specimen for RXU,2 for BRZ, and 10 for BCM. Nevertheless, the SBSvalues obtained for RXU and BRZ were within theabove range of minimum values; thus, these cementsmight be suitable in this respect.

Some standard deviation values in this study ex-ceeded 25% of the mean SBS value. This can be attrib-uted to variability in the anatomic form of the buccalsurfaces of the teeth that might have affected cementfilm thickness. Evans and Powers15 found that thickercement results in lower bond strength. Also, variationsin bracket alignment on the buccal surface and its rela-tionship with the chisel that applied the force could haveinfluenced the obtained stress. Although these variableswere difficult to control, careful selection of the teethand attempts to align the brackets with as much preci-sion as possible made the results sufficiently accurate.16

With the exception of BCM, the mean SBS wasalways higher with the ceramic brackets than with themetallic ones. This might be attributed to enhanced po-lymerization that took place with the ceramic bracketsduring light curing because their relative translucencypermitted some light transmission to the cement.

532 Al-Saleh and El-Mowafy American Journal of Orthodontics and Dentofacial Orthopedics

April 2010

El-Mowafy et al17 showed that, for ceramic inlay thick-ness of up to 3 mm, sufficient light curing reaches theunderlying dual-cured resin cement.

The site of failure also provides useful informationabout the bond. Ideally, an adequate bond that fails atthe enamel-cement interface is desirable because itmakes cleaning and polishing of the teeth much easier.18

Moreover, the cleaning procedures to remove adhesiveremnants can be accompanied with less enamel loss.5

When acid-etching techniques are used, few bond-ing failures were at the enamel-cement interface but,rather, at the cement-bracket interface. According toJou et al19 for light-cured cement, 70% of the failures oc-curred at the resin-bracket interface. This is most likelydue to incomplete polymerization of the resin below themetal base of the bracket because the curing light cannotreach the cement behind the bracket mesh.20 Even lon-ger curing times do not result in the same degree ofpolymerization obtained by direct light curing.21

In our study, TBXT had 6 failures in the metallicgroup and 7 in the ceramic group classified with ARIscores 2 and 3, indicating that all or more than half ofthe cement remained on tooth surfaces. This compareswith 3 for the metallic group and 8 for the ceramic groupwhen ECS was used; this complements the SBS findingsfor this cement. In contrast, only 10 BCM specimenswere available for testing, all were at 0 ARI failurescore, indicating a weak cement/enamel bond. WithBRZ, no metallic bracket specimens had failures withARI 2 or 3, whereas 4 ceramic bracket specimens hadfailures with a score of 2.

Considering the above, it can be concluded thatECS, the resin cement with the separate self-etchingprimer, with ceramic brackets would be expected togive results similar to those obtained with TBXT. How-ever, this does not apply to metallic brackets. Amongthe 3 self-adhesive cements, BRZ would be expectedto perform similarly to RXU with both types ofbrackets; however, BCM is not recommended at allfor cementation of orthodontic brackets.

CONCLUSIONS

1. The SBS of TBXT (control) to metallic and ceramicorthodontic brackets was significantly greater thanthose of all test cements except for ECS with ce-ramic brackets.

2. ECS resulted in SBS values with ceramic bracketsthat were significantly higher than those obtainedwith all self-adhesive cements. Thus, it has poten-tial for use with these brackets.

3. The bracket failure mode differed among theadhesives.

We thank 3M Unitek, Kuraray, 3M ESPE, Pentron,and Bisco for providing the materials for this study andMuneera Alshamrany of the Community Dentistry andPublic Dental Health, Faculty of Dentistry, Universityof Toronto, for assistance with the statistical analyses.

REFERENCES

1. Newman GV. Epoxy adhesives for orthodontic attachments: prog-

ress report. Am J Orthod 1965;51:901-12.

2. Owens SE Jr, Miller BH. A comparison of shear bond strengths of

three visible light-cured orthodontic adhesives. Angle Orthod

2000;70:352-6.

3. Ritter DE, Ritter AV, Bruggeman G, Locks A, Tulloch JF. Bond

strengths and adhesive remnant index of self-etching adhesives

used to bond brackets to instrumented and uninstrumented

enamel. Am J Dent 2006;19:47-50.

4. Bishara SE, Gordan VV, VonWald L, Olson ME. Effect of an

acidic primer on shear bond strength of orthodontic brackets.

Am J Orthod Dentofacial Orthop 1998;114:243-7.

5. Bishara SE, VonWald L, Laffoon JF, Jakobsen JR. Effect of alter-

ing the type of enamel conditioner on the shear bond strength of

a resin-reinforced glass ionomer adhesive. Am J Orthod Dentofa-

cial Orthop 2000;118:288-94.

6. Vicente A, Bravo LA, Romero M, Ortiz AJ, Canteras M. A

comparison of the shear bond strength of a resin cement and

two orthodontic resin adhesive systems. Angle Orthod 2005;

75:109-13.

7. Bishara SE, Ostby AW, Ajlouni R, Laffoon JF, Warren JJ. Early

shear bond strength of a one-step self-adhesive on orthodontic

brackets. Angle Orthod 2006;76:689-93.

8. El-Mowafy O, El-Badrawy W, Eltanty A, Abbasi K, Habib N.

Gingival microleakage of Class II resin composite restorations

with fiber inserts. Oper Dent 2007;32:298-305.

9. Versluis A, Douglas WH, Sakaguchi RL. Thermal expansion co-

efficient of dental composites measured with strain gauges. Dent

Mater 1996;12:290-4.

10. Kinami H, Sugimura M, Sakuda M, Okazaki M, Kimura H. New

type metal bracket for suppression of resin remaining in debond-

ing. Dent Mater J 1990;9:25-35.

11. Lopez JI. Retentive shear strengths of various bonding attachment

bases. Am J Orthod 1980;77:669-78.

12. Reynolds IR. A review of direct orthodontic bonding. Br J Orthod

1975;2:171-8.

13. Øgaard B, Bishara SE, Duschner H. Enamel effects during bond-

ing-debonding and treatment with fixed appliances. In:

Graber TM, Eliades T, Athanasiou AE, editors. Risk management

in orthodontics: experts’ guide to malpractice. Chicago: Quintes-

sence; 2004. p. 19-46.

14. Powers JM, Messersmith ML. Enamel etching and bond strength.

In: Brantley WA, Eliades T, editors. Orthodontic materials: scien-

tific and clinical aspects. Stuttgart, Germany: Stuttgart-Thieme;

2001. p. 105-22.

15. Evans LB, Powers JM. Factors affecting in vitro bond strength of

no-mix orthodontic cements. Am J Orthod 1985;87:508-12.

16. Fox NA, McCabe JF, Buckley JG. A critique of bond strength test-

ing in orthodontics. Br J Orthod 1994;21:33-43.

17. El-Mowafy OM, Rubo MH, el-Badrawy WA. Hardening of new

resin cements cured through a ceramic inlay. Oper Dent 1999;

24:38-44.

18. Toledano M, Osorio R, Osorio E, Romeo A, de la Higuera B,

Garcia-Godoy F. Bond strength of orthodontic brackets using

American Journal of Orthodontics and Dentofacial Orthopedics Al-Saleh and El-Mowafy 533Volume 137, Number 4

different light and self-curing cements. Angle Orthod 2003;73:

56-63.

19. Jou GL, Leung RL, White SN, Zernik JH. Bonding ceramic

brackets with light-cured glass ionomer cements. J Clin Orthod

1995;29:184-7.

20. Swartz ML, Phillips RW, Rhodes B. Visible light-activated

resins—depth of cure. J Am Dent Assoc 1983;106:634-7.

21. Cheung L, Ferguson J, Jones P, Wilson H. An investigation of the

polymerization of orthodontic adhesive by the transillumination

of tooth tissue. Br J Orthod 1989;16:183-8.