Abrasive machining of glass-infiltrated alumina with diamond burs

The abrasive machining characteristics of a glass-infiltrated alumina used for fabrication of all-ceramic dental crowns were investigated using a high-speed dental handpiece and diamond burs with different grit sizes. The material removal rate, surface roughness, and extent of edge chipping were measured as a function of grit size. The removal rate decreased substantially with decreasing bur grit size from supercoarse (180 μm) to fine (40 μm) and ultrafine (10 μm). The removal rate with the supercoarse burs was approximately twice that achieved with the fine burs and four times the removal rate with the ultrafine burs. Both surface roughness and edge chipping damage were sensitive to diamond grit size. Chipping damage was severe and the surface roughness substantial with the supercoarse burs, while negligible edge chipping and smooth surfaces were obtained with the ultrafine burs. The removal rate also decreased with continued machining for all grit sizes. The observed reduction in removal rate was found to be primarily due to wear of the diamond grit and accumulation of debris on the bur (i.e., bur loading). After prolonged use, a significant loss of diamond grit was observed that led to a substantial loss of cutting efficiency. It is concluded that, with respect to material removal rate and surface integrity, diamond machining is a feasible machining process for glass-infiltrated alumina in the final infiltrated state. However, caution should be exercised in the use of diamond grit larger than 40 μm. Such burs may result in excessively rough surfaces, chipped edges, and strength limiting surface and subsurface microcracks.

Keywords: Glass-infiltrated alumina; Dental ceramics; Abrasive machining; Diamond burs; Dental handpiece; Chipping damage; Surface integrity

Outstanding aesthetics and good oral biocompatibility of new dental ceramics and the development of improved fabrication techniques have lead to increased use of all-ceramic crowns. Pressed or slip-cast ceramic cores covered with a ceramic veneer, machinable glass ceramics, and feldspathic porcelains make up most of the all-ceramic crowns that are being used in dentistry [[1]]. Glass-infiltrated alumina, the subject of this investigation, has a higher strength, toughness and hardness than feldspathic porcelains and glass ceramics [2], [[5]], thereby serving as a strong all-ceramic core material for restorative procedures [[9]]. Since the high strength of the glass-infiltrated alumina renders the material more difficult to machine and shape than the lower strength machinable glass-ceramics and porcelains, this particular dental ceramic is usually machined to near net-shape in the preform state (i.e., prior to glass infiltration) [5]. The purpose of the present study is to evaluate the feasibility of abrasive machining of this material in the final infiltrated state to reduce the cost of restorations and obtain better shape control by eliminating the need for glass infiltration after green machining.

In dental practice, abrasive machining is also necessary for final shape adjustments and for repair of failed restorations. Clinical studies of glass-infiltrated alumina crowns have indicated that failures occur when the veneer layer fractures, with or without core exposure, or the core fractures with exposure of prepared tooth structure [12]. Replacement or repair of the failed crowns requires removal of the fractured material by the dentist often with a high-speed dental handpiece utilizing burs of appropriate type, size, and shape for the required procedure. For hard materials such as alumina, diamond burs are the only suitable tools. Due to a growing trend in recent years for the use of disposable or single-patient-use diamond burs to minimize the risk of cross-contamination of blood-borne pathogens [13], the present study is focused on evaluating such burs for abrasive machining of glass-infiltrated alumina.

Only a few studies have examined the machining characteristics of dental ceramics. Siegel and Von Fraunhofer [14] evaluated the cutting efficiency of diamond dental burs against a mica-containing machinable glass-ceramic under three different applied loads. They found that the cutting efficiency was dependent on the handpiece load and also appeared to be dependent on debris accumulation on the bur (i.e., bur loading) and wear of the diamond grit. Dong et al. [15] investigated the machining characteristics of a commercial dental glass-ceramic and three experimental glass-ceramics with different mica platelet sizes. The removal rate was found to increase as the size of the mica platelets was increased, but the extent of edge chipping was decreased. It was also found that coarse burs did not necessarily remove material faster than fine burs, but produced more extensive chipping damage as compared to the damage produced by fine burs.

The aim of the present study is to investigate the machining characteristics of a glass-infiltrated alumina utilizing a high-speed dental handpiece and diamond burs covering a range of grit sizes. Material removal rate, surface roughness, and edge chipping are evaluated as a function of diamond grit size in the dental burs. The cutting efficiency of diamond burs is evaluated based on the phenomena of bur loading and grit wear.

The material investigated was a commercial glass-infiltrated alumina used for dental restorations, commonly known as In-Ceram Celay Alumina (Vita Zahnfabrik, Bad Sackingen, Germany).[1] The test specimens were received in the form of fully processed infiltrated bars with a rectangular cross section (3 mm × 4 mm × 25 mm). The processing procedure was described as follows: First, dry-pressed alumina plates were prepared and partially sintered, resulting in preform blanks with 30% porosity (pore size ranging from 1 μm to 5 μm). The preform blanks were then machined into bars and infiltrated with lanthanum-aluminosilicate glass for 6 h at 1170 °C. After cooling, the excess glass was removed by sand blasting. The microstructure of the glass-infiltrated alumina, as determined by Jung et al. [16], consists of a fairly uniform distribution of roughly round alumina particles ranging in size from 1 μm to 3 μm. The glass infiltrate occupies the interconnected pores that existed in the preform prior to glass infiltration. The relevant mechanical properties of the glass-infiltrated alumina used in this study, as determined by Jung et al. [16], are as follows: Young's modulus = 254 ± 4 GPa, Vickers hardness = 11.8 ± 0.4 GPa, and indentation toughness = 2.5 ± 0.2 MPa m1/2.

The machining apparatus utilized in this investigation was designed by Dong et al. [15]. This test device incorporates a high-speed air-turbine-type handpiece equipped with an air-water mist delivery system (625C Super Torque, KaVo American Corp.). The dental handpiece is clamped to an arm that pivots freely on a ball bearing support. The load between the bur and workpiece is established by a pulley and suspended weight arrangement.

Disposable dental diamond burs manufactured by NTI Diamond (Axis Dental Corporation, Irving, TX) were used. Each bur comprises a steel blank of fixed diameter for mounting in the handpiece and a cutting portion that consists of a single layer of synthetic diamond grit, electrolytically codeposited with nickel. Burs of three different grit sizes designated by the manufacturer as supercoarse (862 012SC), fine (862 012F), and ultrafine (862 012UF) were used. The average diameter of the diamond particles were estimated from a series of micrographs obtained by scanning electron microscopy (SEM) on new burs to be 180 μm, 40 μm, and 10 μm, respectively. The burs had a pre-plated nominal diameter of 1.2 mm and a straight cylindrical geometry with a flame shaped tip at the end. In the cutting experiments reported here, the burs were positioned so that only the straight cylindrical portion was in contact with the alumina workpiece (Fig. 1). The actual bur diameter in the diamond section was larger than the nominal diameter due to the added thickness of the plated layer and the diamond grit [15]. Inspection of the new burs in the SEM also indicated that the coarser burs contained a smaller number of diamond particles and larger inter-particle spacing than the finer burs.

Graph: Figure 1. Schematic drawing of a dental diamond bur and the geometry of a machined groove: (1) top surface, (2) bottom surface, (3) inside groove surface, (4) side groove surface, and (5) front surface.

The orientation of the bur during machining with respect to the specimen is illustrated in Fig. 1, which also shows schematically the various features associated with a single groove cut in the specimen. The bur was oriented approximately parallel to the 3 mm × 24 mm surface of the specimen designated front surface (5) in Fig. 1. A series of such cuts was made along the length of each specimen.

Water coolant was delivered to the handpiece at a constant flow rate of 15 ± 1 ml/min during cutting. The handpiece was positioned such that the water spray nozzle was located at 180 degrees with respect to the bur-specimen contact. Before each test series was begun, a light hydrocarbon lubricant (KaVo lubricant spray, KaVo American Corporation) was applied to the handpiece for approximately 1 s. The handpiece was then run without load for 60 s as suggested by Siegel and Von Fraunhofer [14] to remove excess lubricant from the handpiece.

A constant applied load of 2 N was used for all tests. This load is typical of that used by clinicians in finishing operations [14], [[17]]. At a constant air pressure of 241 ± 7 kPa (35 ± 1 psi), the pressure used throughout out this investigation, the rotational speed of the bur was approximately 320,000 rpm, measured by means of a strobe type tachometer (Strobotac 1538-A, General Radio Company). On contacting the specimen under the 2 N load the speed dropped to 260,000 rpm where it remained until the handpiece was removed from the specimen.

Three burs of each diamond grit size were used in the study for each test condition to allow for the evaluation of the repeatability of the results. The machining sequence consisted of cutting a series of grooves (twenty-five with each supercoarse bur and fourteen with each fine and ultrafine bur). The cutting duration for each groove was approximately 10 s.

Volume removal rate, surface roughness, and edge chipping were the quantities used to evaluate the machining characteristics of the glass-infiltrated alumina. Each measurement was repeated three times and the average was used to calculate the mean and one standard deviation for the three repeat tests.

The volume removal rate for each groove was determined by dividing the groove volume by the time to cut the groove. The groove volume was determined by multiplying the average of the top and bottom groove areas by the specimen thickness. An image analysis system (Leica Microimage Video System with Image-Pro Plus) was used to measure the top and bottom groove areas. The actual cutting duration was determined from the cutting force data that were recorded as a function of time [15].

Surface roughness was measured by means of a stylus profilometer (Perthen, Mahr GmbH). The traces perpendicular to the machining direction were made across the bottom of the grooves (designated 3 in Fig. 1) using a trace length of 1.75 mm and a cut off length of 0.25 mm. Arithmetic mean roughness (Ra) and the maximum roughness depth within the measurement length (Rmax) were used to characterize surface roughness.

The quantity used to determine the degree of edge chipping at each groove was the projected chipped area per unit length measured along the groove edges on the top and bottom surfaces of the specimens (designated 1 and 2 in Fig. 1). The projected chipped area is the area of the chip projected on the plane of the top and bottom specimen surfaces. Measurements of the chipped area and the edge length of the grooves were made with the same image analysis system used in the determination of groove volume. Several machined samples were coated with gold and were examined in the SEM to assess the mechanisms of material removal during machining.

Following the cutting experiments, the burs were sputter coated with gold and examined by means of a SEM, without any prior cleaning, to determine the condition of the burs with respect to bur loading and grit wear.

A series of prolonged cutting tests was conducted to further evaluate bur loading and wear of diamond grit. In these tests, a new fine bur was used to cut seventy-two grooves and the bur was ultrasonically cleaned in water for 30 s (Ultrasonic Cleaner, Cole-Palmer) between each cut. This procedure was followed to minimize the influence of bur loading by removing the debris attached to the bur. The bur was periodically sputter coated with gold and examined in the SEM to check its condition with respect to loading and grit wear. Note that the gold coating is weakly attached to the bur surface and it is removed when cutting is resumed. Therefore, the gold coating does not have a significant influence on the machining results.

The removal rates obtained with burs of three different grit sizes, without bur cleaning between the cuts, are shown in Fig. 2. The results clearly show that the removal rate decreases with decreasing grit size. The removal rate for supercoarse burs (SC) is approximately twice the removal rate for fine burs (F) and at least four times higher than the removal rates for ultrafine burs (UF).

Graph: Figure 2. Removal rate obtained with supercoarse (SC), fine (F), and ultrafine (UF) diamond burs without cleaning between the cuts. Machining time for each groove was about 10 s. Each data point is the mean value for three grooves cut with three different burs; the uncertainty bars correspond to ± one standard deviation for each test.

A common feature exhibited in Fig. 2 is the marked decrease in removal rate with increasing cutting time, irrespective of the bur grit size. For the supercoarse burs after 120 s of use the removal rate has dropped to about 50% of the initial value, and after 240 s it is approximately 30% of the initial removal rate. The fine burs show a similar trend. The reduction in removal rate for the ultrafine burs, however, is much greater, falling to about 50% of the initial value after only 20 s and to 17% after 120 s.

The results of surface roughness measurements on the inside surfaces of the grooves are shown in Fig. 3 as the arithmetic average roughness (Ra) and maximum roughness (Rmax). The results clearly indicate that surface roughness scales with the diamond grit size. The roughness parameters are higher for the surfaces machined with the supercoarse (SC) burs than those with the fine (F) or ultrafine (UF) burs. The results in Fig. 3 suggest that the surface roughness is relatively unaffected by the increase in cutting time, i.e., bur wear has a relatively small influence on roughness.

Graph: Figure 3. Arithmetic average roughness (Ra) and maximum roughness (Rmax) for grooves cut with supercoarse (SC), fine (F), and ultrafine (UF) diamond burs. Each data point is the mean value for three grooves cut with three different burs; the uncertainty bars correspond to ± one standard deviation for each test.

The SEM micrographs in Fig. 4 show typical features inside the groove surfaces produced by machining with supercoarse, fine, and ultrafine burs. Consistent with the measured surface roughness values (Fig. 3), there is a trend toward less fracture and smoother appearance as the diamond grit size becomes smaller. The micrograph in Fig. 5 (a) shows, at a higher magnification, the surface cut with a fine bur. A series of parallel machining striations cut by individual diamond grits are seen on the surface. The striations have a smeared appearance suggesting plastic flow of the material or the adhered machining debris. The micrograph also shows fractured regions between adjacent striations. The fracture pattern seen at a higher magnification in Fig. 5 (b), suggests that fracture is primarily associated with an intergranular crack extension process along the interface between the alumina grains and the glass infiltrate. This is evident as the round shaped alumina grains are seen on the fracture surface.

Graph: Figure 4. SEM micrographs showing typical inside groove surfaces cut with (a) a supercoarse bur, (b) a fine bur, and (c) an ultrafine bur.

Graph: Figure 5. SEM micrographs showing microfracture on the machined groove surfaces cut with a fine bur: (a) and (b) show different locations on the machined surface at different magnifications.

The micrographs in Fig. 6 show the top surfaces of three grooves cut with the burs having three different grit sizes. A large chipped area is seen at the edge of the groove cut with a supercoarse bur, Fig. 6 (a), and a smaller chipped region at the edge of the groove cut with a fine bur, Fig. 6 (b). The groove cut with an ultrafine bur, Fig. 6 (c), shows no chipping damage along the edge. In order to quantify these observations, the average chipping width (i.e., the projected chipped areas per unit length) is plotted in Fig. 7 for the grooves cut with the three grit sizes. The figure shows that edge chipping is greatly diminished with the decrease of diamond grit size, such that no edge chipping could be measured on the grooves cut with the ultrafine burs. It is noted that edge chipping is not greatly influenced by cutting time.

Graph: Figure 6. Typical chipping damage on the top surface along the groove edges cut with: (a) a supercoarse bur, (b) a fine bur, and (c) an ultrafine bur (shown at a higher magnification than in (a)).

Graph: Figure 7. Average chipping width (or the projected chipped area per unit length) along the groove edges cut with supercoarse (SC), fine, (F), and ultrafine (UF) burs. Each data point is the mean value for three grooves cut with three different burs; the uncertainty bars correspond to ± one standard deviation for each test.

Following the cutting tests, several used burs were examined in SEM, without any prior cleaning. The supercoarse burs appeared clean with no evidence for adhered debris, as shown in Fig. 8 (a). However, in comparison with the unused burs, occasional dulling of diamond cutting edges was observed. Therefore, the reduction in the removal rate with the supercoarse burs (Fig. 2) is attributed to wear of the diamond grit rather than bur loading.

Graph: Figure 8. SEM micrographs showing diamond grit on the bur surfaces without prior cleaning for: (a) supercoarse bur after cutting for 250 s, (b) fine bur after cutting for 140 s, and (c) ultrafine bur after cutting for 160 s. Note that different magnifications are used to highlight bur loading or accumulation of adhered material on the bur surfaces in (b) and (c).

SEM examination of the fine and ultrafine burs did not reveal much evidence for wear of the diamond grit. But in contrast with the observations on the used supercoarse burs, the fine burs were partially covered with patches of adhered material, Fig. 8 (b). The amount of adhered material (i.e., bur loading) was much less on the ultrafine burs, Fig. 8 (c), as compared to the fine burs. The patches of adhered material apparently have developed in the vicinity of grits that were actively involved in cutting. Such loading not only diminishes the cutting efficiency, but is likely to be a source of variations in cutting rate as adhered material breaks away and re-accumulates on the bur surfaces.

In order to evaluate the effects of bur loading and grit wear on removal rate and to determine the wear mechanisms, a series of prolonged tests was conducted in which the bur was ultrasonically cleaned after each groove was cut. Figure 9 shows the removal rate as a function of cutting time for a fine bur. The removal rates obtained without intermediate cleaning (from Fig. 2) are also shown for comparison. The results indicate slightly higher removal rates with the use of ultrasonic cleaning. However, in both cases there is a gradual decrease in the removal rate with increasing the cutting time. Measurement of surface roughness inside the cut grooves and inspection of the groove edges showed no significant effect on surface roughness or edge chipping as the cutting time was increased.

Graph: Figure 9. Removal rate with a fine bur with ultrasonic cleaning between the cuts. Data without cleaning are from Fig. 2.

Periodic examination of the bur in the SEM confirmed the effectiveness of ultrasonic cleaning in removing the adhered material from the bur. Initially, there was very little visible evidence of bur wear; but eventually, grit fracture, grit loss, and noticeable dulling of grit cutting edges began to appear. The effects of wear are shown in Fig. 10. Evidence of grit fracture is shown in Fig. 10 (a), grit loss in Fig. 10 (b), edge rounding and formation of wear-flats in Fig. 10 (c), and damage to the surrounding nickel matrix in Fig. 10 (d).

Graph: Figure 10. SEM micrographs showing typical features observed on the surface of the fine bur: (a) micro-fracture of diamond grit after 570 s of cutting, (b) grit loss after 570 s of cutting, (c) edge dulling and wear flats after 720 s of cutting, and (d) damage to the metal matrix after 570 s of cutting.

Preparation of dental restorations by machining requires the ability to obtain a desired geometry with precise dimensional control and acceptable levels of surface integrity and material removal rate. While the material removal rates achieved for the glass-infiltrated alumina are lower by approximately a factor of twenty compared to Dicor MGC (a commercial machinable glass-ceramic developed for dental restorations prepared by machining), the resulting surface roughness and chipping characteristics are comparable [15]. When higher removal rates are desired during machining or finishing of restorations, it is a common practice to use burs with coarser diamond grit. The present investigation shows that although higher removal rates can be achieved with coarser diamond burs, the resulting chipping damage and poor surface quality may prove to be limiting factors that cannot be ignored.

The material removal rate and the resulting surface roughness and chipping damage are all influenced by the fundamental deformation and fracture processes that take place during machining. Removal of material during abrasive machining of brittle polycrystalline materials occurs through a microscale fracture process that involves generation and/or extension of microcracks along weak interfaces (i.e., grain boundaries) [20]. The glass-infiltrated alumina used in the present investigation contains approximately 30% glass occupying the interconnected pores that existed in the preform prior to infiltration. Since the hardness and toughness of the glass regions are much less than those of the alumina phase, fracture events occur either within the glass or at the glass/alumina interface. The SEM micrographs of the machined surfaces, Fig. 5, confirm that the material removal process in this material follows a pattern consistent with the microscale fracture process with cracks propagating along the alumina/glass interface.

Related to the microscale fracture process associated with abrasive machining of brittle materials is the formation of subsurface damage [20] that could be detrimental to the performance of ceramic dental restorations. Recent studies have shown that diamond grit size is an important parameter in controlling the extent of subsurface damage and fracture strength of machined ceramic surfaces [[21]]. Machining-induced subsurface microcracks can also be generated in tooth enamel as a result of diamond bur machining. It has been shown [23] that large grit sizes cause larger cracks than smaller grit sizes. Therefore, the use of burs containing large diamond particles in machining of dental restoration and tooth structures could be problematic due to extensive edge chipping and possible detrimental subsurface damage. If it is necessary to use coarse or supercoarse burs to remove a large amount of material rapidly, the damaged surface layer should be removed with finer burs to obtain a more acceptable surface finish and to restore the strength of restoration.

The observed relationship between machining characteristics (i.e., material removal rate, surface and subsurface damage, and edge chipping) and grit size can be explained based on the contact load between the diamond particles and the workpiece rather than a direct influence of grit size on the removal process. The load applied to the bur is shared by the diamond particles that contact the workpiece during machining. As described earlier, the number of diamond particles was smaller in the coarser burs as compared to the fine burs. Although the load applied to the bur was constant, the average normal load acting on the grit increases as the diamond grit size is increased from 10 μm to 180 μm (from ultrafine to fine and supercoarse burs). It has been shown that the material removal rate associated with a single abrasive particle during the microscale fracture process is proportional to the load applied to the grit [20]. Therefore, the larger grit loads associated with the coarser burs result in a higher rate of material removal. The larger grit loads associated with coarser burs also result in more extensive fracture because of larger grit penetration depths and higher contact stresses during machining.

An important issue in abrasive machining is the removal of the machining debris by the cutting fluid (i.e., water spray). If the debris is not efficiently removed from the cutting zone, it can be deposited onto the work surface and/or the bur, resulting in poor surface integrity and loss of cutting efficiency. The space available for the transport of the debris is determined largely by grit protrusion height above the supporting metal matrix and the distance between adjacent diamond particles. For the burs used in the present study, the usable space for debris transport is related to the grit size; thus, more efficient debris removal is expected for the supercoarse burs compared to the fine and the ultrafine burs. In fact, examination of the burs, Fig. 8, revealed no adherent debris or bur loading for the supercoarse burs; whereas the diamond particles in the fine burs were covered with adherent debris. However, the surfaces of the ultrafine burs were fairly clean with only a small amount of adhered debris, probably because of much smaller amount and possibly smaller particle size of debris generated during machining with the ultrafine burs.

The phenomenon of bur loading has been observed in machining of glass ceramics [14]. Although it is possible to remove the adhered debris from the bur to reverse the associated loss of cutting efficiency, the renewed cutting action lasts only for a short period of time. As soon as machining is commenced and debris is generated, the debris can accumulate on the bur. It should be noted that loading occurs despite the fact that a mist of water and air is directed at the bur during cutting. For a given bur and machining condition, bur loading might be minimized through the use of more efficient water spray and possibly utilizing surfactants that could reduce the tendency of the debris to adhere to the bur surface. It is important to distinguish bur loading and wear because the former could be removed to partially restore the cutting efficiency, however the latter can only shorten the tool life especially for single-layer diamond burs. Although the diamond burs used in the present investigation were designed for single-patient use to reduce the risk of cross-contamination, our study shows that these burs can be used to remove a substantial amount of material before losing cutting efficiency due to wear and grit loss.

The following conclusions are drawn from the results of the present investigation on the machining behavior of glass-infiltrated alumina using a high-speed dental handpiece and diamond burs:

• 1. Glass-infiltrated alumina can be machined in the infiltrated state with diamond burs. However, caution should be exercised in the use of diamond grit larger than 40 μm. Such burs may result in excessively rough surfaces, chipped edges, and strength limiting surface and subsurface microcracks.
• 2. The material removal rate increases with increasing bur grit size when comparing ultrafine (10 μm) to fine (40 μm) and supercoarse (180 μm) dental burs.
• 3. The material removal rate for all grit sizes gradually decreases with continued machining due to wear of the diamond grit.
• 4. Surface roughness and edge chipping damage are sensitive to diamond grit size, but insensitive to bur loading and grit wear.

The authors acknowledge the generous supply of materials from H. Horn-berger of Vita Zahnfabrik. They also wish to thank I. M. Peterson, B. R. Lawn, H. H. K. Xu, T-W. Hwang, J. Quinn, G. Quinn, and E. Whitenton, of the National Institute of Standards and Technology and X. Dong of Lucent for helpful discussions and assistance on various aspects of this work. Sincere appreciation goes to V. P. Thompson of the University of Medicine and Dentistry of New Jersey for overall guidance in this program. This project was supported by the National Institute of Dental and Craniofacial Research (NIDCR) Project Grant No. 1-P01 DE10976.