TECHNO BYTES

CAD/CAM/AM applications in the manufacture ofdental appliances

Noor Al Mortadi,a Dominic Eggbeer,b Jeffrey Lewis,c and Robert J. Williamsd

Cardiff, United Kingdom

FromaPostgbResecSeniodReadThe aucts oReprinropoliac.ukSubm0889-Copyrhttp:/

The purposes of this study were to apply the latest developments in additive manufacturing (AM) constructionand to evaluate the effectiveness of these computer-aided design and computer-aided manufacturing (CAD/CAM) techniques in the production of dental appliances. In addition, a new method of incorporating wire intoa single build was developed. A scanner was used to capture 3-dimensional images of Class II Division 1 dentalmodels that were translated onto a 2-dimensional computer screen. Andresen and sleep-apnea devices weredesigned in 3 dimensions by using FreeForm software (version 11; Geo Magics SensAble Group, Wilmington,Mass) and a phantom arm. The design was then exported and transferred to an AMmachine for building. (Am JOrthod Dentofacial Orthop 2012;142:727-33)

Computer-aided design/computer-aided manu-facturing/additive manufacturing (CAD/CAM/AM) scanning systems have been successfully in-

troduced in dentistry.1-3 Researchers have used thistechnology as a tool for orthodontic diagnosis andtreatment planning for determining the position ofimpacted maxillary canines4 and for the fabrication ofocclusal splints.5 However, CAD/CAM/AM innovationshave not been used successfully on a wide scale for re-movable orthodontic appliances.

Sassani and Roberts6 reported that it was possible touse CAD/CAM/AM techniques to build base plates of or-thodontic appliances, but they stated that it was notpossible to incorporate wires into the built parts. How-ever, in this study, we discuss a method to accomplishthis.

The technology mentioned above has been used inorthodontic treatments such as the invisible orthodonticappliance fabricated without metal wires or brackets(Align Technology, Santa Clara, Calif)7 and in producingcustomized lingual brackets.8 Recently, Wiechmannet al9 used this computer-based technology to make

Cardiff Metropolitan University, Cardiff, United Kingdom.raduate research student, School of Health Sciences.arch officer, National Centre for Product Design andDevelopment Research.r lecturer, School of Health Sciences.er, School of Health Sciences.uthors report no commercial, proprietary, or financial interest in the prod-r companies described in this article.t requests to: Robert J. Williams, School of Health Sciences, Cardiff Met-tan University, Cardiff CF5 2YB, United Kingdom; e-mail, rjwilliams@uwic..itted, January 2012; revised and accepted, April 2012.5406/$36.00ight � 2012 by the American Association of Orthodontists./dx.doi.org/10.1016/j.ajodo.2012.04.023

connectors for Herbst appliance hinges to custom lin-gual brackets, and the authors of a more recent studyused this technique for a digital titanium Herbst appli-ance.10

Because of the lack of application in general to re-movable appliances, the aims of this study were to applynew methods of producing dental appliances with wiresand hinges by using CAD/CAM/AM based techniquesand to evaluate these. Andresen and sleep-apnea appli-ances were chosen. Although an Andresen appliance israrely used, it provides a clear illustration for the inclu-sion of wire into a build that is straightforward. Asleep-apnea device shows where 4 coordinating hingescan be studied. Both devices were produced in a singlebuild.

MATERIAL AND METHODS

Opposing Class II case models suitable to be treatedwith Andresen and sleep-apnea removable deviceswere scanned by a hand-held laser sensor (HandyScan3D; Creaform, L�evis, Qu�ebec, Canada) (Fig 1). Laser lineswere projected from the scanner onto the surface of eachmodel to produce a 3-dimensional image from a “pointcloud.” The software automatically transformed thepoint cloud into a stereolithographic facet file, shownas it appears on a screen (Fig 2).

The scanner was moved around the casts by hand toproduce a clear and complete 3-dimensional surface inSTereoLithography (STL), a defacto file type used inAM (Fig 3), with a minimum of missing data.

The data were processed, and the STL files were gen-erated by using the proprietary scanner software (VXelements, version 2.1; Creaform). The appliances were

727

Fig 1. Hand-held scanner.

Fig 2. Point cloud after triangulation.

Fig 3. Three-dimensional image.

Fig 4. Phantom (haptic) arm, and appliance being built inFreeForm.

728 Al Mortadi et al

designed with specialist CAD software (FreeFormModel-ing Plus, version 11; Geo Magics SensAble Group, Wil-mington, Mass), in conjunction with a PhantomDesktop (Geo Magics SensAble Group) (haptic) arm(Fig 4). This apparatus allowed appliances to be “built”on scans of a patient’s cast in a virtual environment.

The software enables the design of complex, nongeo-metric forms and is ideal for designing anatomicallybased devices.11 It has various tools to accomplishspecific functions. Some tools are used for building,and others are used for trimming and shaping theconstructed parts. The Phantom device works as a 3-dimensional mouse because it mimics the work of a den-tal technician, enabling the design and construction ofthe appropriate appliances similarly to the conventionalmethod. The operator can feel force feedback sensationsduring the process.

November 2012 � Vol 142 � Issue 5 American

Before the virtual build, it was necessary to createa Class I molar relationship or achieve this as closely aspossible. This relationship was accomplished by postur-ing the mandible forward, until a Class I molar relation-ship was obtained and the vertical opening between theopposing premolars was 4 to 5 mm at the lingual cusps.

The labial inclination of the maxillary central incisorswas selected as a reference for the “path of insertion” ofthe appliance onto the upper model. Undercuts on bothvirtual models were detected and eliminated virtuallyautomatically by using the “extrude-to-plane” tool.The files were then saved as “buck” files, which meantthat the models were protected and could not bechanged.

The following is a brief description of the designprocess for each appliance.

Journal of Orthodontics and Dentofacial Orthopedics

Fig 5. Andresen monoblock in the FreeForm virtual envi-ronment, indicating where the upper and lower plateswere joined.

Fig 6. Intermaxillary edge.

Al Mortadi et al 729

For the Andresen, a 2-mm layer was applied to theupper and lower models with capping on the mandibularincisors. Palatal coverage was made by selecting the re-quired area with a cursor and offsetting a 2-mm thick-ness to form the plate. Upper and lower platesextended to the first molars. Capping was 1 mm thickon the mandibular anterior teeth to prevent proclinationand overeruption. The thickness of the appliance was setat 2 mm to withstand occlusal forces and patient han-dling. The virtual material so formed is termed “clay,”and it is added and shaped in a similar way to physicalwax. It was necessary to add clay in some areas of the ap-pliance to make the surfaces smooth and to form junc-tions of 2 surfaces continuous to each other. Borderswere smoothed.

Finally, the occlusal surfaces of the opposing offsetclay areas were connected to produce a 1-piece “mono-block” at the lingual surfaces of the teeth extending tothe middle of the occlusal surfaces (Fig 5).

These blockswere attached to the plates, and the junc-tions were smoothed with the “sculpting” and “smooth”tools. The middle distances of the blocks were stretchedbuccally by using the “smudge” tool up to the outlineof the buccal surfaces to form a sharp “intermaxillaryedge.” The blocks were cleared from themandibular teethto permit upward and forward eruption (Fig 6).

To allow wire to be inserted in the physical build,a square section tube (1.25 3 1.25 mm) on either sideof the appliance was created. Each tube consisted of 2parts. The first part was horizontal, 15 mm in lengthand 1 mm from the lingual walls of the intermaxillaryblocks, and was located in the maxillary part with itslower border at the level of the intermaxillary edge.The second part was vertical, projecting into the maxil-lary part, and 5 mm in length. The first part was at rightangles to the second part, and these were connected ata point between the second premolar and the first molar.These tubes were made so that the retentive part ofa 0.9-mm hard stainless steel wire labial bow could beinserted, as shown in Figure 7.

For thewire to be inserted in the tubes of the appliance,it was necessary to build “guiding jigs.” Fitting the wire inthe guiding jigs before the build ensured that the wiretagging would fit into the appliance during the build.

The guiding jigs (ie, small plastic components aroundwhich wire could be bent before the main build) werebuilt from rapid prototype material for the 3D SystemsProject 3000 Plus Machine (3D Systems, Rock Hill, SC),which was acrylate based. The jigs were produced by us-ing the “copy” tool to duplicate the upper virtual modelin conjunction with the outline of the tubes. Clays wereconstructed in the space between the model and thelevel of the tubes (Fig 8, A). Also, a small amount of

American Journal of Orthodontics and Dentofacial Orthoped

clay was added under the tubes to make sure that thewire would not be in contact with the model (Fig 8, B).This was accomplished by offsetting a 1-mm layer underthe tubes by using the “sculpting clay” tool.

The tubes formed spaces along which wire would bebent before insertion in the final device during the buildprocess. Figure 9 shows grooves 1.25 mm deep. Sharpangles were smoothed to facilitate positioning of thewire in the tubes.

Next, the guiding jigs were exported as an STL file toan AMmachine (3D Systems Project 3000 Plus Machine)to be built in a new colorless, acrylate-based material(Visijet EX200, 3D Systems).

The wax supporting material was removed bymeltingin an oven set at 72�C, and an oil bath was used to re-move residual wax. A citrus-based degreaser was then

ics November 2012 � Vol 142 � Issue 5

Fig 7. Virtual labial view of tubes built to accept wire.

Fig 8. Guiding jigs for the formation of wire tagging beforethe physical build.

Fig 9. Arrows indicate where the wire tagging will be bentvertically into the jig.

Fig 10. Labial bow fitted on the guiding jigs.

730 Al Mortadi et al

used to clean the jigs before they were fitted to the stonemodels.

The labial bow was bent in the conventional way, andthe tagging fitted to the jigs was positioned on the gyp-sum cast (Fig 10). The labial bow was constructed from0.9-mm hard stainless steel wire. The bow extendedfrom canine to canine, touching the labial surfaces ofthe maxillary incisors and the canines at the mesial thirdarea. From that point on each side, U-loops were formed1 to 2 mm below the gingivae on both sides. The wirewas then passed between the canines and the first pre-molars to enter the plastic between the occlusal surfaces.The wire tagging followed the tubes formed in the jigs.The horizontal parts of the wire tags were made straightand passively fitted on the jigs.

The next stage was to export the Andresen design toan AM machine as an STL file ready for fabrication inWaterShed XC 11122 (dsm somos; DSM Headquarters,Elgin, Ill).

A stereolithography machine (SLA 250-50; 3D Sys-tems) was chosen for fabrication, since it enables buildsto be paused and prefabricated pieces to be inserted and

November 2012 � Vol 142 � Issue 5 American

built around. This was necessary for embedding thewires.

Preparation software (Magics; Materialise, Leuven,Belgium) was used to orientate the Andresen designwith the recessed section to accommodate the wire sethorizontally. The same software was also used to generatesupport structures that are necessary to attach the part tothe build platform. Lightyear (version1.1; 3DSystems)wasused to prepare and slice the build in 0.1-mm layers beforetransferring the build to the SLA machine. The machinewas paused before it reached the top of the tubes to allowinsertionof thewire in the correct place and then restarted.

The “zephyr blade” recoating mechanism, which nor-mally sweeps unprocessed liquid from the build surfaceof the machine, was stopped from sweeping for the 3layers after insertion of the wire so that the wire wasnot disturbed during the process. The procedure wascontinued until building was completed (Fig 11).

Support structures were removed from the device,and it was cleaned in fresh isopropanol solvent (99%

Journal of Orthodontics and Dentofacial Orthopedics

Fig 11. The final prototype before cleaning.

Fig 12. Final Andresen in CAD.

Al Mortadi et al 731

minimum) for 20 minutes, and then dried by using com-pressed air and ventilated for a further 6 hours. Next, thedevice was postcured for 30 minutes in ultraviolet lightto ensure full polymerization.

The final design is shown in Figure 12, and the builtAndresen fitted to the master model in Figure 13.

For the sleep-apnea device, the design was built inthe virtual environment using FreeForm in combinationwith a Phantom arm and manufactured using the sameprocess and material used in the “guiding jigs” describedabove in relation to the Andresen device. In the virtualenvironment, a 2-mm layer was applied to the upperand lower models and extended along the occlusal andbuccal surfaces to approximately 1 mm short of thegingival margins facially to prevent irritation to thegingivae. One millimeter of capping was applied onthe 4 incisors to prevent labial inclination (Fig 14) andextended 2 to 3mmbelow the palatal and lingual gingivalmargins. The hinges on both sides connected the maxil-lary molars to the mandibular canines, and were parallelto allow free rotation. The role of these hinges is tomove the mandible forward into the corrected position.

Each bar connected 2 hinges located on the opposingarches. One bar was built on each side to connect themaxillary first molar hinge to the mandibular caninehinge. The 4 hinges were parallel.

Each hinge consisted of 2 parts (Fig 15). The first partwas a circular rod (2.4 mm in length, 3 mm in diameter)and a wide circular “stop” (6 mm in diameter) (Fig 15,A).The second part was a hollow cylinder with an internaldiameter of 3.5 mm and an external diameter of 6mm, as shown in blue in Figure 15, B. The 2 partswere separated in all areas to permit rotational freedomof the cylinder around the rod. There was a gap for clear-ance between sides of the cylinder and the first part. Theinner gap was 0.2 mm, and the outer gap was 0.1 mm.The gap between the inner surface of the cylinder andthe rod was 0.5 mm. These gaps were filled with waxat the build stage.

Finally, the design was exported as an STL file. A Pro-jet 3000 Plus machine (3D Systems) was chosen to man-ufacture the sleep-apnea device from a colorless material(Visijet EX200). Fundamentally, the process allows fea-tures such as hinges to be fabricated as 1 piece, withthe wax melted away to free the mechanism once pro-duced. The Projet process can also build extremely finedetail (656 3 656 3 1600 dpi in the x, y, and z axes).

The device was transferred into proprietary softwareand orientated in the same plane in which it would beworn (ie, the lowest z height) to obtain the optimumbuild speed. The build took 8 hours to complete. Thecompleted build was postprocessed by using an ovenset at 72�C to melt the wax support structure and an

American Journal of Orthodontics and Dentofacial Orthoped

oil bath to remove residual wax followed by degreasingto clean the device. This was then trial fitted to the stonemodels (Fig 16).

RESULTS

The appliances were fitted on the physical models,and their features were checked and compared withthe expectations of conventional appliances. Both appli-ances built by the AM method had smooth surfaces withno sharp edges that could not be easily dealt with bynormal finishing procedures.

With regard to the Andresen, the fit palatally and lin-gually was satisfactory. Because of the virtual prepara-tion of the models, no areas were positioned inundercuts. The labial bow fitted firmly in the tubes,was quite well connected to the appliance, and wasdeemed functional. The acrylic guiding channels couldbe trimmed according to conventional practices to

ics November 2012 � Vol 142 � Issue 5

Fig 13. Andresen prototype fitted to casts.

Fig 14. Virtual build of sleep-apnea device.

Fig 15. Ahinge showing stops (A) and rotating section (B).

Fig 16. The sleep-apnea prototype fitted to casts.

732 Al Mortadi et al

enhance the anterior movement of the mandibular teethand the posterior movement of the maxillary teeth, andto encourage expansion of the maxillary arch and up-ward eruption of the mandibular teeth into the correctedocclusion.

The sleep apnea device displayed an excellent fit onthe teeth, and on the palatal and gingival tissues.

November 2012 � Vol 142 � Issue 5 American

The hinges allowed free anterior and posteriormovements. Some sideways movement was also per-mitted by the hinges. Further experimentation couldreduce this, but there might be an argument in favorof maintaining such movement to allow the device tobe myodynamic.

DISCUSSION

The 2 appliances described above could be presentedto a clinic for fitting to a patient. These appliances high-light the technical possibilities of using CAD/CAM/AMtechnologies in dental device designs. Although thetechniques are currently slow and expensive, the benefitsof introducing digital technologies into other areas ofdentistry are well known. The ability to fabricate layerby layer enables features such as inserting wire, produc-ing hinges, and building threads to be made in 1 buildwithout assembly. In orthodontics, this has a potentialapplication in fabricating devices incorporating built ex-pansion screws and Twin-blocks; this is part of ongoingresearch. Other benefits of using CAD are accurate con-trol of the thickness of the material, producing an evenshape when required, and exact positioning of the pe-riphery of the design. Precise measurements can be easilytaken as the virtual build progresses or programmed in asthe software produces layers. A huge potential advan-tage is that the techniques could be linked to intraoralscanning; this would help to streamline the process ofproducing appliances by removing the impression anddental-cast production stages.

CONCLUSIONS

The article indicates that digital technologies can beapplied to the fabrication of appliance types not previ-ously considered to be possible.

Journal of Orthodontics and Dentofacial Orthopedics

Al Mortadi et al 733

REFERENCES

1. Stevens B, Yang Y, Mohandas A, Stucker B, Nguyen KT. A review ofmaterials, fabrication methods, and strategies used to enhancebone regeneration in engineered bone tissues. J Biomed MaterRes B Appl Biomater 2007;85:573-82.

2. Sun Y, L}u P, Wang Y. Study on CAD & RP for removable completedenture. Comput Methods Programs Biomed 2009;93:266-72.

3. Williams RJ, Bibb R, Eggbeer D. CAD/CAM-fabricated removablepartial-denture alloy frameworks. Pract Proced Aesthet Dent2008;20:349-51.

4. Faber J, Berto PM, Quaresma M. Rapid prototyping as a tool fordiagnosis and treatment planning for maxillary canine impaction.Am J Orthod Dentofacial Orthop 2006;129:583-9.

5. LaurenM,McIntyre F. A new computer-assisted method for designand fabrication of occlusal splints. Am J Orthod DentofacialOrthop 2008;133(Suppl):S130-5.

American Journal of Orthodontics and Dentofacial Orthoped

6. Sassani F, Roberts S. Computer-assisted fabrication of removableappliances. Comput Industry 1996;29:179-95.

7. Abe Y, Maki K. Possibilities and limitation of CAD/CAM basedorthodontic treatment. International Congress Series 2005;1281:1416.

8. Wiechmann D, Rummel V, Thalheim A, Simon JS, Wiechmann L.Customized brackets and archwires for lingual orthodontictreatment. Am J Orthod Dentofacial Orthop 2003;124:593-9.

9. Wiechmann D, Schwestka-Polly R, Hohoff A. Herbst appliance inlingual orthodontics. Am J Orthod Dentofacial Orthop 2008;134:429-46.

10. Farronato G, Santamaria G, Cressoni P, Falzone D,Colombo M. The digital-titanium Herbst. J Clin Orthod2011;45:263-7.

11. Bibb R, Eggbeer D, Williams R. Rapid manufacture of removablepartial denture frameworks. Rapid Prototyping J 2006;12:95-9.

ics November 2012 � Vol 142 � Issue 5