of fracture zones in porcelain-veneered-to-metal bond test spwimens ESCA analysis K. J. Anusavice, The Medical
Ph.D., D.M.D.,*
R. D. Ringle,**
College of Georgia, School of Dentistry,
and C. W. Fairhurst, Augusta,
by
Ph.D.*“*
Ga.
F
ractures of porcelain-veneered-to-metal restorations may be attributed to single or multiple deficiencies. Potential causes of failure include lack of ceramic-metal adherence, thermal contraction incompatibility, design and procedural errors, malocclusion, impact stresses exceeding normal service conditions, or combinations of these factors. Specific analysis of failure by visual and physical examination can be effectively made only after obtaining a thorough history of fabrication and clinical conditions. Visual classification of fractures has been considered useful to describe the macroscopic appearance and geometry of the fracture surface with respect to the original metal surface or unaffected porcelain structure. However, identification of the precise origin and composition of the fracture surface by visual examination alone is unreliable, since it is subjective in nature. O’Brien and co-workers’s 2 have suggested a system for classification of fracture surfaces on the basis of interfaces observed on fractured bond test specimens. This approach is potentially useful in classifying certain types of fractures. However, before generalizations regarding the nature of fracture surfaces are proposed, one should be concerned with the following factors: (1) correlation of fractures in bond test specimens with the types of fracture surfaces appearing on clinical porcelain-fused-to-metal crowns has not yet been established; (2) fracture paths can follow any of an infinite number of paths which may cross several interfaces in an unpredictable manner; (3) in diffuThis research HEW.
was supported
*Assistant **Research
Professor, Associate,
***Professor
and
0022-3913/79/100417
Dental Dental
by Grant Materials Materials
Coordinator,
+ 05$00.50/OQ
Dental
No.
DE
04252,
NIH-NIDR,
Division. Division. Materials
Division.
1979 The C. V. Mosby
Co
Fig. 1. Bond machine.
test
specimen
in
the
Instron
testing
sion profile studies, well-defined interfaces and the identity of thin structural regions such as metal oxides and bonding agent layers are frequently obscured by the significant diffusion and dissohttion interactions which have occurred; and (4) the overall appearance of a fracture surface may not be representative of the region of crack initiation. To completely describe the nature of fractures in ceramic-metal systems, previously established adherence theory and data must also be considered. This supplementary information is summarized in three principal subject areas: (1) theoretical development of adherence concepts,3-5 (2) evaluation of bond test
THE
JOURNAL
OF PROSTHETIC
DENTESTRY
417
ANUSAVICE,
RINGLE,
AND
FAIRHURST
types of fracture which occur in arch-shape specimens and (3) to correlate the. visual appearance of the fracture surface with spectroscopic data.
MATERIALS
-y-X-ray L X-ray monochromator
Source
-
Fig. 2. Simplified block diagram of HP-5950B ESCA spectrometer (Modified from Karasek, F. W.: Operating the new software-based ESCA. Research/Development Oct. 1975, p 33.)
Fig. 3. Enlarged view of a sample in the ESCA spectrometer (Modified from Karasek, F. W.: Operating the new software-based ESCA. Research/Development Oct. 1975, P 33.) data 6. ’ and (3) compositional analysis across ceramic-metal interfaces8-‘* The objectives of this study were (1) to develop a technique for identification of specific zones in which fracture was presumed to have occurred in ceramic-nonprecious metal test specimens subjected to flexural stress, (2) to chracterize the principal
418
AND METHODS
The test design selected for this study involved use of the previously developed semicircular casting (approximating the size of a human maxillary arch) to which porcelain was fired on the convex surface as shown in Fig. 1 .I3 The rationale for this test ‘configuration was that, as the arch was radially deformed in tension in an Instron* testing machine, the porcelain layers were under compressive stresses, whereas tensile stresses were confined within the metal. A nonprecious alloy (Ceramalloy),Jf designed for porcelain-metal restorations, was selected for study. This alloy has shown acceptable adherence when used with its recommended bonding agent. Twenty arch-shaped castings received firings of Ceramalloy bonding agent,? Paint-0-Pake,t and body porcelaint according to manufacturer’s directions. Five additional castings were similarly prepared with the exception that the bonding agent coating was not scrubbed prior to firing the porcelain, as recommended by the alloy supplier. This left a heavy, loose scale on the surface to be covered with porcelain. To analyze the composition of isolated fracture sites with minimal penetration to deeper structures nonrepresentative of the fracture, the ESCA technique of analysis was selected. ESCA, an acronym for electron spectroscopyfor chemical analysis, involved the use of x-rays to cause the emission from sample atoms of inner-shell electrons whose binding energies unambiguously define a specific atom (Fig. 2). Because the emitted electrons do not generally escape from depths greater than 5OA (about 10 atomic layers), ESCA is essentially a surface analysis technique. A Hewlett-Packard Model 5950B ESCAS unit was used for analysis of fracture zones ranging in area from 1 to 5 mm*. A relative accuracy in survey scans of 10% is attainable in quantitative analysis. Four standards were prepared for comparative analysis: a bare alloy, an oxidized alloy, a bonding agent fired on an alloy substrate, and an opaque porcelain fracture surface. The fractured ceramic-metal test specimens were grouped into three specific types of fracture patterns, and one surface representative of
*Instron Corp., Canton, Mass. tceramco, Inc., East Windsor, N. J. SHewlett-Packard, Palo Alto, Calif.
~CT~BERIW~
VOLUME
42
NUMBER
4
IDENTIFICATION
OF FRACTURE
ZONES
1 1000
500 Bindinq
Energy
(eV)
Fig. 4. ESCA survey scan of a bare metal standard.
OtAUGER) ‘Oil
I
1
1000
500 Bindinq
Energy
l
(ok’)
Fig. 5. ESCA survey scan of an oxidized each of the three fracture types was submitted for ESCA analysis. The deepest portion of each fracture zone was analyzed for compositional determination of the possible site of fracture initiation. To minimize the effects of surface contaminants, each specimen was argon ion-etched to remove a SOA layer before analysis. Areas for examination were isolated by confining the area from which electrons were accepted for analysis to the deepest area of the various fracture zones. The standard area of analysis was 5 mm2 (1 X 5 mm), although electronic suppression of counts in the surrounding area is possible down to an area of 1 mm2 (1 X 1 mm). The electron detector responds through a number of electron-multiplier elements to produce a pulse of lo7 electrons for each
m
JOURNAL
OF FROSTHETIC
DENTISTRY
0
alloy.
electron received. An enlarged view of the specimen is depicted in Fig. 3. Survey scans were m&e over the binding energy range 0 to 1000 eV. These scans revealed the principal elerknts pre~nt. in the fracture zone. Quantitative analyses were carried out by determining peak height to photoionization cross sections, which are proportional to specific atom concentrations. RESULTS Typical survey scans for the bare metal standard and metal oxide standard are shawn in Figs. 4 and 5. A survey scan, representative of fracture surface No. 41 and typical of failure within the bond&g agent, is shown in Fig. 6. The surface analysis of these seven specimens included significant concentrations df
419
ANUSAVICE,
RINCLE,
AND
FAIRHURST
Gr? (AUGER) Specimen: Fracture Surface t41 Alloy: Ceramalloy Bonding Agent: Cemmalloy Porcelain: Ceramco Bond Test: Single Cycle Arch Ftexure
5
d
4
yi
0 :
3 b-4
Sn
-u
O-
, 0
500
1000 Binding
Fig.
6.
Energy
(eV)
ESCA survey scan of fracture surface No.
41.
Table I. Surface composition analysis
Specimen
Fig. 7. Failed bond test specimen characteristic of metalmetal oxide-bonding agent complex fracture.
Metal standard Metal oxide standard Bonding agent standard Opaque porcelain standard Fracture No. 4 No. 24 No. 41
Fib. 8. Bond test specimens representing three types of fracture.
carbon, nitrogen, and adsorbed oxygen. After subtracting the effects of these contaminants, the resulting concentrations shown in Table I were obtained. DISCUSSION Fracture surface No. 4 was the fracture plane adjacent to the bonding agent layer visible as a gray area. The analysis of this fracture surface corre-
420
Composition
iatomicpercent)
Ni
Cr
Sn MO-
K
Si
Al
0
Na
67.5 5.4
19.9 58.9
- 3; - -
-
9.0 5.4
-
32.3
-
1.7
8.0 2.0
-
-
21.0
40.2
27.1
-
-
5.1
-
1.6
46.9
10.3
33.7 2.4
-
13.6 6.6
- 0.9 37.5 - 4.0 - 24.9 2.5 - 2.0 28.8
15.0
17.7 2.7
40.0 39.8 24.3
6.6
33.1
-
sponds well with the fractured opaque porcelain composition, although no kdium (Na) was detectable. Only one fracture of this type occurred. A secondspecimen also failed within the porcelain but at a point dloser to the external ceramic surface. In only two of the 20 specimens (Id%) did fracture occur in the porcelain structures alone. Fracture surface No, 24 appears to be a failure iti the bonding agent-metal oxide interfkial zone. It is noteworthy that this surface displayed apparent ba?e metal areas within the bonding agent regions as shown in Fig. 7. Ten of the twenty specimens(50%) failed in this manner. This fracture surface contains significant nickel (Ni), chromium (Cr), and silicon (Si) but matches r&the; metal, metal oxide, or the bonding agent-fGed-to-metal standards. This sur-
OCTOBER
1979
VOLUME
42
NUMBER
4
Fig. 9. Ceramalloy-porcelain bond test specimens representing fracture at the bonding agent-porcelain interface. Specimens were not scrubbed after bonding agent firing as recommended by the alloy supplier. face most likely represents a composition produced by the interaction of the bonding agent with the metal and metal oxide, since it contains molybdenum (MO) originating from the metal, a silicon concentration representative of the bonding agent, and nickel and chromium concentrations indicative of the inner layer of metal oxide. Fracture surface No. 41 shown in the center of Fig. 8 corresponds well in composition to the standard produced by fusing the bonding agent to an alloy substrate, except for a 2% potassium content probably contributed by interaction with the opaque porcelain. Failures of this type, i.e., extending into the bonding agent-porcelain interaction zone, were observed in eight of the 20 specimens (40%). It is interesting to note that all five specimens which were not scrubbed after bonding agent firing failed clearly along the bonding agent surface. as in Fig. 9.
CONCLUSIONS
JOURNAL
OF PROSTHETIC
REFERENCES 1.
DENTISTRY
O’Brien. Materials
W. J.: Dental Review. Ann
Michigan,
2. 3.
4.
pp
porcelains. .4rbor.
Mich.,
In Craig. 197’;.
R. ,J.: Dental University of
123-135.
O’Brien, W. J.: The cohesive plateau stress of ceramic-metal systems. J Dent Res 56:B177, 1977. (Abst No. 501) King, B. W., Tripp H. P., and Duckworth, W. H.: Nature adherence
of porcelain
42:504.
1959. M. P., and in the role
Borom, oxides”
enamels
to metals.
Pask, J. A.: of chemical
J Am
The role bonding
Ceram
of Sot
of “adherence at glass-metal
interfaces. J Am Ceram Sot 49:1, 1966. Vickery, R. C., and Badinelli, C. A.: Nature forces in porcelain-gold systems. J Dent Res
of attachment 47:683, 1968.
6.
Shell,
bond
7.
gold alloys and porcelain. McLean, J. W., and Seed,
J Dent Res I. R.: Bonding
to metal. II. The base-metal Ceram Sot 72:235, 1973.
alloy/porcelain
5.
8.
J. S., and
Nally,
J. N.,
9.
11.
Nielsen,
Monnier,
topographique porcelaine au Schweiz Monat
10.
The ESCA technique provides a useful approach to identifying probable sites of fracture initiation in ceramic-fused-to-metal systems in which the analyzed fracture area is 1 mm’ or larger. The primary disadvantages of this technique include high equipment cost, the need for meticulous removal of superficial oxide layers formed at room temperature, and susceptibility to surface contamination by carbon, nitrogen, and oxygen. From this study the following may be concluded: 1. Judgment of the composition of any fracture surface (except that which occurs exclusively in porcelain) by naked-eye examination or optical microscopic inspection may be erroneous. 2. Generalizations on the nature of adherence zone failures for nonprecious alloy-ceramic systems
THE
should not be made in the absence of fracture surface or adherence zone composition data determined by means of an acceptable analytical technique. 3. It appears unlikely that fractures originating in one of the regions of the adherence zone will have a composition analysis identicai to that of metal, metal oxide, or bonding agent standards, since considerable interdiffusion or chemical interaction between these regions will have altered the original chemical compositions.
J. P.: Study
D..
and
of the
bond.
Meyer.
de certains ClCments niveau de la liaison Zahn 78:868, 1968.
between
41:1425, 1962. of dental porcelain
J. hf.:
Trans
Br
Distribution
de l’aiiiage et de CCramco-rhetallique.
la
Lautenschlager, E. P., Greener, E. H., and Elkington, W. E.: Microprobe analysis of gold-porcelain bonding. j Dent Res 48: 1206, 1969. Baker, M. aluminum-Use
A.:
Study of adherence of the electron
microprobe. Proc Goeller, I., Meyer,
and
enamel to electron
Port Enam Inst 32:48, 1970. J. M., and Nally, J. N.: Comparative
study of three coating agents strength of porcelain-fused-to-gold
28:504,
of porcelain microscope
and
their allovs.
influence on bond J PROSTHE?. DENT
1972.
12.
Anusavice. K. J., Ringle, R. D., and Faxhurst, C. Bonding mechanism evidence in a ceramic-nonprecious alloy system. J Biomed Mater Res 11:701, 1977.
W.:
13.
Mackert. Fairhurst.
J. R., Jr., Anusavice, C W.: A flexure-shear
K. J.. test
and
to-metal
bonding.
55:B236,
J Dent
Res
Ringle, R. D., for porcelain-fusedi!?76.
(Abst
No.
698) Reprint requests to: DR.
K. J. ANUSAVICE
SCHOOL MEDICAL AUCUS~A.
OF DENTISTRY COLLEGE OF GEORGIA Gh.
30912
421