The Fit of Gold-alloy Full-crown Castings Made with Ceramic Casting Ring Liners R. EARNSHAW2 and E.F. MOREY13 Department of Prosthetic Dentistry and 'Department of Operative Dentistry, University of Sydney, 2 Chalmers Street, Surry Hills, NSW, Australia 2010 Measurements were made ofthe fit of gold-alloy full-crown castings produced with dry ceramic ring liners. When used with vacuum investing, these liners absorb relatively large amounts of water from the investment mix (thereby reducing its original W/P ratio) and then function as wet liners, thus increasing the investment's potential expansion and giving castings which are consistently larger than when air investing is used. With four ofthe five liners tested, investing in air produced many castings which were unacceptably undersized (inaccuracy worse than -0.2%). The fifth liner, an industrial material 2 mm thick, gave only one casting out of 12 which was outside this limit, although all castings were undersized to a lesser extent. Vacuum investing gave improved casting accuracy; with four of the five liners, the improvement was highly significant (p < 0.001), and with the fifth, probably significant (p < 0.05). Even with vacuum investing, however, with only two of the liners did all castings show inaccuracies within ± 0.2%. With the other three liners, some castings (ranging from 2/10 to 7/9) had inaccuracies worse than -0.2%. With both air and vacuum investing, changing from one liner to another caused changes in relative casting accuracy which were often significant (p < 0.01) or highly significant (p < 0.001). In casting techniques where a ceramic ring liner is used, the choice of specific lining material and the choice between investing in air or under vacuum are important factors which can have a major effect on the fit of castings. J Dent Res 71(12):1865-1870, December, 1992
Introduction. A study has already been described (Morey and Earnshaw, 1992) of the effects of pre-wetted asbestos and cellulose ring liners on the dimensional inaccuracy of gold-alloy full-crown castings. In the present investigation, measurements were made of casting inaccuracies obtained with five different ceramic liners, used dry. Under normal atmospheric conditions, ceramic liners have a low water absorption; however, it was found that they readily absorbed large amounts of water under reduced pressure. Because of this, during vacuum investing, water was abstracted from the investment mix, reducing its original W/P ratio, shortening its working time, and increasing both setting and thermal expansions to an uncontrolled extent; the then-wet liner increased the setting expansion still further. Relevant properties of the ceramic liners were therefore measured both on dry specimens and on specimens vacuum-impregnated with water. Similarly, casting inaccuracy tests were made with and without vacuum investing.
Materials and methods. The same batches of gold alloy (G5, Engelhard Industries Pty. Ltd., Thomastown,Australia) andinvestment (Cristobalite Inlay, Sybron/ Received for publication January 22, 1991 Accepted for publication August 7, 1992 2Present address: 1 Wall Avenue, Asquith, NSW, Australia 2077 3Present address: Government Employees Health Fund Dental Clinic, 8 Quay Street, Haymarket, NSW, Australia 2000
Kerr Products, Romulus, MI) were used, manipulated in the same way under the same ambient conditions. Except in tests in which the W/P ratio was varied deliberately, all mixes were made with a W/P ratio of 0.40. Liners tested.-These are listed in Table 1, with an identifying code. Of the five, three (DK, GC, and KF) are sold specifically as dental casting ring liners. The other two (K1 and K2) are industrial materials of the same type-nominally 1 and 2 mm thick, respectively-used principally for high-temperature thermal insulation. Experimental methods.-In general, testing procedures used with the ceramic liners followed those already described for tests on asbestos and cellulose materials (Morey and Earnshaw, 1992). Points of difference are described below. Water uptake.-The water uptake under reduced pressure was determinedby simulation ofthe vacuum-investing process. Casting rings were lined (liners 3 mm short of each end of the ring, 2 mm overlap at the joint) and weighed. They were then attached to the Vac-U-Spat investor, which contained water only, and the usual investing procedure was followed. The ring was then removed, drained of excess water, and re-weighed. Thickness and compressibility.-For measurements ofwet thickness and compressibility, specimens were impregnated with water under a gauge vacuum of 96 kPa maintained for 30 s under vibration. Effective WIP ratio in the ring.-The effect ofvacuum investing in a ring lined with a ceramic material was determined as described in the previous report, except that the liner was not pre-wetted. Since the effect here was to reduce the W/P ratio of the mix, the calibration curve of W/P ratio against working time, measured in unlined casting rings, covered the range from 0.40 down to 0.29 (the thickest mix which could satisfactorily be manipulated). This curve was then used to convert working times recorded in lined rings, filled in air or under vacuum, to effective W/P ratios. Investment expansion.-Laboratory measurements of setting and thermal expansion were made as described previously (Earnshaw, 1988). Here, measurements were first made on specimens mixed in the normal W/P ratio of 0.40, setting against dry ceramic liners; these conditions simulated the use of casting rings with dry ceramic liners and air investing. Then, measurements were repeated under conditions simulating the use of dry ceramic liners with vacuum investing. The investment was mixed in the effective W/P ratio already determined, and the trough was lined with the ceramic materials after they had been vacuum-impregnated with water. Control tests were made on specimens setting against a PTFE trough liner, at a W/P ratio of 0.40. In the tests simulating vacuum investing, it was found that all liners could be removed intact from the set specimens, which had not been possible with all pre-wetted asbestos and cellulose liners or dry ceramic liners. Therefore, in these tests, serial weighings ofthe liners allowed estimation of the amount of water absorbed by the specimens from the liners during setting. Casting inaccuracy.-Castings were made in rings with dry ceramic liners, with vacuum investing, and with vacuum mixing followed by air investing. All investment mixes were made in the standard W/P ratio of 0.40. When vacuum investing was used with K2, the large amount of water absorbed from the mix reduced its working time to such an extent that the investing process, even under vibration, was often not effective in filling the ring, and 1865
EARNSHAW & MOREY
1866
J Dent Res December 1992
TABLE 1
CERAMIC RING-LINING MATERIALS TESTED Name
Code
Supplier
Dentsply Kaoliner
DK
Dentsply/York Division, York, PA 17405-0872
G C Casting Liner
GC
G-C Dental Industrial Corp., Tokyo, Japan
Kerr Flask Liner
KF
Triton Kaowool 1 mm
K1
Triton Kaowool 2 mm
K2
Sybron/Kerr Products, Romulus, MI 48174 Morganite Ceramic Fibres, Alexandria, Australia 2015 Morganite Ceramic Fibres, Alexandria, Australia 2015
defective castings resulted. Of a total of 27 castings, only nine were measurable. Statistical analyses.-All individual sets of results were subjected to ANOVA (GLM procedure, Tukey's Studentized Range HSD test, at both the 0.01 and 0.05 confidence levels). In addition, t tests were used for assessment of the significance of changes in liner compressibilityon wetting; the significance of differences in potential setting, thermal and total expansion; and differences in casting accuracy, when vacuum rather than airinvestingwas used. For tests with air investing, linear regression (SAS PLOT and GLM procedures) was used to seek possible correlations between casting inaccuracy and (a) dry compressibility of the liners and (b) potential investment expansion. For tests with vacuum investing, correlations were sought between measured water uptake of the liners and (a) changes in thickness and compressibility and (b) potential investment expansion; and between casting inaccuracy and (a) effective W/P ratio in the ring, (b) wet and (c) dry compressibility ofthe liners, and (d) potential investment expansion. Multiple linear regression was applied to the relation between casting inaccuracy and combinations of relevant variables, for both air and vacuum investing. In the statistical analyses, the following levels of significance were applied: p < 0.001 = highly significant; 0.001 < p < 0.01 = significant; 0.01 < p < 0.05 = probably significant; p > 0.05 = not significant.
Results. Properties of the liners.-Reproducible results for thickness made multiple measurements unnecessary (Table 2); mean values are given for water uptake under vacuum, and for dry and wet compressibility. There was a significant correlation between thickness
and compressibility, both dry and wet. Unlike the asbestos and cellulose liners tested previously, which all increased in thickness and compressibility on wetting, the different ceramic liners showed either positive or negative changes in thickness and compressibility after impregnation with water; in general, the changes were not as great as with the absorbent liners. Changes in compressibility were highly significant except for K1, where the increase was probably significant. Predictably, there was no correlation between water uptake and change in either thickness or compressibility. Effective WIP ratio in the ring.-The Fig. shows the effect of varying the W/P ratio in the range 0.29 to 0.40 on the working time, measured on control specimens setting in dry, unlined casting rings. Table 3 lists the working times measured on specimens from standard mixes setting in rings lined with the various ceramic materials, initially dry, after being filled under vacuum; also listed are the effective W/P ratios of these latter mixes, estimated from the curve in the Fig. With K2, it was not possible for a value for working time-and hence for effective W/P ratio-to be obtained; the large amount of water it absorbed under vacuum (Table 2) increased the rate ofsetting ofthe mix to such an extent that it was already plastic at the end of the investing process (1 min 20 s). With the other four ceramic liners, simulated vacuum investing gave working times highly significantly less than those of the control specimens setting in unlined rings. There was a probable negative correlation between the water uptake of the liners (Table 2) and the effective W/P ratio in the ring. When rings fitted with dry ceramic liners were filled under atmospheric pressure, no significant changes occurred in working time, nor therefore in W/P ratio. Investment expansion.-Results of measurements of potential expansion are given in Table 4. Where tests simulated conditions of vacuum investing, results are also given for the estimated
TABLE 2 WATER UPTAKE UNDER VACUUM, DRY AND WET THICKNESS, AND DRY AND WET COMPRESSIBILITY OF THE MATERIALS TESTED Liner Water Uptake Under Vacuumt Thickness (mm) Compressibility(mm)t (mL) Dry Wet Dry Wet DK 3.56 ± 0.05 (5) 1.34 1.27 a 0.40 + 0.02 (12) 0.38 ± 0.02 (16) GC 2.29 ± 0.02 (6) 0.90 0.87 0.23 + 0.03 (20) a 0.27 ± 0.03 (21) KF 2.15 + 0.12 (5) 0.83 0.88 0.34 ±0.02 (11) a 0.29 ± 0.02 (13) K1 3.32 + 0.17 (5) 1.05 1.19 a 0.41 ± 0.03 (20) 0.44 + 0.04 (10) K2
5.71 ± 0.28 (5)
2.64
2.54
0.75 ± 0.07 (19)
0.66 ± 0.04 (10)
*See Table 1. tValues for water uptake and compressibility include standard deviations and (in parentheses) numbers of tests. Means marked "a" are not significantly different (p > 0.05).
FIT OF CASTINGS MADE WITH CERAMIC LINERS
Vol. 71 No. 12
amounts of water absorbed by the setting specimen. Tests simulating air investing gave results which showed only small differences compared with those of control tests. Except for DK, increases in setting expansion, though small, were significant (GC) or highly significant. Similar small increases in setting expansion with dry ceramic liners were observed in an earlier investigation (Earnshaw, 1988). No changes in thermal expansion were significant. Only with K2 (8%) and KF (3%) was there a significant increase in total expansion. Under conditions simulating vacuum investing, much more dramatic changes occurred. Tests could not be performed on K2, but with the other four liners increases in setting expansion were highly significant. Concomitant decreases in thermal expansion were significant with all liners except KF. Net increases in total expansion were highly significant. There was a highly significant positive correlation between the amount ofadditional water absorbed by the setting specimen and setting expansion-a relationship recognized by Asgar et al. (1955), who used it as the basis for the "water-added" technique. The amount of absorbed water also showed a probably significant negative correlation with thermal expansion and a significant positive correlation with total potential expansion. Casting inaccuracy.-With air investing, all liners gave undersized castings (Table 5). There was no correlation between casting inaccuracy and potential investment expansion, or between inaccuracy and dry compressibility of the liners. However, there was a II
II
101-
1867
TABLE 3 THE EFFECTIVE W/P RATIO OF THE UNSET INVESTMENT MIX IN LINED CASTING RINGS AFTER VACUUM INVESTING, ESTIMATED FROM ITS WORKING TIME (Control tests were made in unlined rings.)
Liner* Control
Working Time t (min) 10.07 ± 0.12 (3)
Effective W/P Ratio t 0.40
DK
1.60 ± 0.02 (3)
0.29
GC
a 4.41 ± 0.45 (6)
0.31
KF
6.68 ± 0.97 (8)
0.34
K1
a 4.52 ± 0.55 (6)
0.31
< 0.29 K2 0.05).
probably significant correlation between inaccuracy and the combination of both these variables. With vacuum investing, all liners gave improved casting accuracy; the improvement was highly significant with all materials except GC, where it was probably significant. Apart from the exceptions marked in Table 5, all liners, whether used with air or vacuum investing, gave inaccuracies which were significantly or highly significantly different from each other. There were significant correlations between casting inaccuracy and water uptake by the liners under vacuum, wet compressibility, and dry compressibility, and a probably significant correlation between casting inaccuracy and the combination of these three variables. For K2, results could not be obtained for effective W/P ratio or potential investment expansion. However, with the other four liners, there was a probably significant negative correlation between casting inaccuracy and effective W/P ratio and a probably significant positive correlation between inaccuracy and potential investment expansion.
510
Discussion.
2.-c
0 cc
01--
I
I
I
0.30
I 0.40
W/P RATIO Fig.-The relationship between W/P ratio of the investment mix and its rate of setting in unlined rings, shown by the working time.
Previously, it was found (Morey and Earnshaw, 1992) that, wetted in the usual way, asbestos and cellulose liners showed mean water absorptions in the range 1.25 to 2.12 mL. The amounts absorbed by ceramic liners when immersed under vacuum were very much higher (Table 2). Under practical conditions of vacuum investing, this water uptake would occur at the expense ofthe investment mix, greatly reducing its effective W/P ratio (Table 3); this, coupled with the presence of a now-wet liner, caused the large increase in total potential expansion shown in Table 4. The various results for expansion, particularly setting expansion, emphasize the importance of practical conditions being reproduced as closely as possible in laboratory measurements of this property. If ± 0.2% is accepted as an allowable tolerance in casting accuracy (Morey and Earnshaw, 1992), with air investing only K2 gave a mean inaccuracy within this range (Table 5); even with this
J Dent Res December 1992
EARNSHAW & MOREY
1868
TABLE 4 POTENTIAL EXPANSION OF CRISTOBALITE INLAY INVESTMENT, MEASURED UNDER LABORATORY CONDITIONS (During setting, specimens were surrounded by the various lining materials tested. They were then heated to 650'C. Test conditions simulated investing in air, and investing under vacuum. With vacuum investing, estimates are given of the amount of water absorbed from the liners by the setting specimens.) Water Absorbed W/P No. by Setting Thermal Setting Total of Ratio Investing Specimen Expansiont Expansion Expansiont Tests Used (% Liner Method (mL) (%) (%) 6 a 0.37 + 0.02 A a 1.17 ± 0.05 PTFE 0.40 a 1.54 ± 0.04 (control) A a 0.35±0.01 a 1.16±0.01 a 1.51±0.02 DK 3 0.40
GC KF
K1 K2
V
0.29
8
A
0.40
6
V
0.31
4
A
0.40
6
V
0.34
6
A
0.40
6
V
0.31
6
A
0.40
6
V
1.90 ±0.10
1.36 ±0.12 1.06 ± 0.09
1.50 ±0.10
2.58 ±0.04
0.97 ± 0.03
3.56 + 0.05
0.41±0.01
a 1.11±0.03
a 1.52 ±0.02
1.85 ± 0.03
1.03 ± 0.03
2.88 ± 0.04
0.46±0.04
a 1.13±0.04
1.59±0.01
1.13 ± 0.05
a 1.19 ± 0.03
2.32 ± 0.07
0.45 ± 0.03
a 1.11 ± 0.06
a 1.56 ± 0.03
2.08 ± 0.04
1.04 + 0.04
3.11 ± 0.08
0.53 ± 0.03
a 1.14 + 0.06
1.67 ± 0.03
Tests could not be performed, as the effective W/P ratio was not known.
*See Table 1. tA = air investing. V = vacuum investing. tMean and standard deviation. In each column, means marked "a" are not significantly different (p > 0.05).
material, one of the individual castings had a discrepancy worse than -0.2%, and would probably have been unacceptably undersized. With the other four materials, all or most of the castings had a discrepancy worse than -0.2%. With the sprueing method used here, relative casting accuracy depends largely on expansion occurring diametrally in the core of the mold. This, in turn, depends on the potential investment expansion, modified by the presence of the casting ring, its liner, and the wax pattern. When the pattern is invested in air, the setting investment is not exposed to additional water; therefore, the only important liner property is its dry compressibility, which can affect both setting and thermal expansion. There was a probably significant correlation between casting inaccuracy and the combination of potential investment expansion and dry liner compressibility. With vacuum investing, a consideration ofthe effect of the liner on casting inaccuracy is more complex. Liner properties which directly modify investment expansion diametrally in the ring (and therefore casting inaccuracy) are water uptake (which affects setting expansion), wet compressibility (also affecting setting expansion), and dry compressibility (affecting thermal expansion). There was a probably significant correlation between casting inaccuracy and the combination of these relevant liner properties. As a result of the improved accuracy produced by vacuum investing, four of the liners gave mean values within the tolerance of + 0.2%; the exception, KF, wasjust outside it. However, only with K1 and K2 did all individual castings show inaccuracies within ± 0.2%. All other liners produced some castings which, by this
criterion, would have been unacceptably undersize-for DK, 2/10, for GC, 2/9, and for KF, 7/9. The fact that many ofthe differences in mean casting inaccuracy given by the different liners were significant or highly significant, both with air and with vacuum investing (Table 5), emphasizes the importance of liner choice in any casting technique that uses a ceramic liner. The increase in measured setting expansion found with dry liners and vacuum investing was caused by a reduction in WfP ratio, accompanied by the presence of a wet ring liner which induces hygroscopic setting expansion. With the investment used here, the hygroscopic setting expansion can be greatly reduced by mechanical restraint (Earnshaw, 1969); nevertheless, since vacuum investing produced, in most instances, a small decrease in thermal expansion (Table 4), the improved accuracy of these castings must have been the result of effective hygroscopic setting expansion occurring in the investment core. As with pre-wetted liners (Morey and Earnshaw, 1992), it is possible for an indication to be obtained of the effective total expansion shown by the investment core with each of the liners, by summing the values for expected alloy casting shrinkage and the mean casting inaccuracy. These values, for both air and vacuum investing, are shown in Table 6, column 3. Similarly, by then subtracting the values for potential thermal expansion given in Table 4, an estimate can be made of the effective setting expansion occurring in the core (Table 6, column 4). These values are all considerably less than the potential setting expansion ofthe investment with the same liners (Table 4). The percentage reduction in setting expansion found in the investment core is shown in Table 6,
Vol. 71 No. 12
FIT OF CASTINGS MADE WITH CERAMIC LINERS
1869
reduction in setting expansion and dry compressibility of the liners. For DK, GC, KF, and K1, vacuum investing caused increases in effective setting expansion when compared with air investing, reflected in the improvements in casting accuracy shown in Table 5. As with air investing, the various reductions in setting expansion shown in Table 6, column 5, would have been the result ofvariations in liner compressibility. However, there was no correlation between Dimensional Inaccuracy of Castings (%)t percentage reduction in setting expansion and wet compressibility. Liner Air Investing Vacuum Investing A comparison of relative casting accuracy obtained with prewetted asbestos and cellulose liners (Morey and Earnshaw, 1992), Control: - 0.61 ± 0.05 (11) with dry ceramic liners used with air investing, and with vacuum DK a -0.31±0.07(10) a - 0.16 ± 0.05 (10) investing, shows that with only four liners did every casting have inaccuracies within the range + 0.2%. These were pre-wetted KA b - 0.23 ± 0.04 (10) ab - 0.18 +0.05 (9) GC and BE, and K1 and K2 used dry but with vacuum investing. In all other instances, at least some castings were produced which were KF - 0.49 ± 0.08 (10) b - 0.21 ± 0.05 (9) undersized by more than 0.2%. With pre-wetted SH and WM, and with dry DK and KF used with air investing, all castings had Ki a,b - 0.29 ± 0.09 (10) - 0.08 ± 0.06 (9) inaccuracies worse than -0.2%. Considered together, the results of these two studies show the K2 - 0.16±0.05 (12) - 0.05 ±0.08 (9) importance of the choice of ring liner in dental casting techniques. *See Table 1. The simplest example is represented by the use of a dry ceramic tMean, standard deviation, and (in parentheses) number of liner and air investing, since only two variables are involvedtests. potential investment expansion and dry compressibility of the TMorey and Earnshaw, 1993. liner-and both are easily controlled. There was a probably signifiIn each column, means marked with the same letter were not cant correlation between the combination ofthese two variables and significantly different (p > 0.05). castinginaccuracy. Underthese conditions, though, the investment used here did not have a potential expansion sufficiently high to provide an acceptable level of accuracy. This probably also applies column 5. For air investing, it ranged from about 25% (GC) to about 90% to many other currently available inlay casting investments. The use of a wet liner, either pre-wetted or dry (in conjunction (KF). The effect of the wax pattern in reducing setting expansion would have been the same, so that the differing reductions in setting with vacuum investing), increases the mold expansion but introexpansion would have resulted from differences in liner compress- duces more variable factors, and in the latter case has the disadvanibility. However, there was no correlation between percentage tage of reducing the working time of the investment. Even so, only
TABLE 5 THE DIMENSIONAL INACCURACY OF GOLD-ALLOY FULL-CROWN CASTINGS MADE WITH NO RING LINER (CONTROL) AND EACH OF THE CERAMIC LINERS TESTED
TABLE 6 CALCULATED VALUES FOR EFFECTIVE TOTAL EXPANSION OF THE INVESTMENT CORE, ITS EFFECTIVE SETTING EXPANSION, AND THE PERCENTAGE REDUCTION IN SETTING EXPANSION OCCURRING IN THE CORE UNDER PRACTICAL CONDITIONS
Investing Method Air
Vacuum
Effective Total Expansion t
Effective Setting Expansion t
Percentage Reduction in
Setting Expansion §
Liner DK
(%) 1.34
(%) 0.18
49
GC
1.42
0.31
24
KF
1.16
0.03
93
1K1 1K2
1.36
0.25
44
1.49
0.35
34
DK
1.49
0.52
80
GC
1.47
0.44
76
KF
1.44
0.25
78
1K1
1.57
0.53
75
K2 Calculations not possible, since values for potential expansion were not available
*See Table 1. tSum of alloy shrinkage and casting inaccuracy. tDifference between effective total expansion and potential thermal expansion. §Percentage difference between potential and effective setting expansion.
1870
EARNSHAW & MOREY
four of the 11 wet liners studied here consistently produced fullcrown castings with an acceptable dimensional accuracy. A logical approach to improvement in the dimensional accuracy of gold-alloy full-crown castings would be to use a casting ring with a ceramic liner in conjunction with air investing, since this involves only two easily controlled variables. It would be necessary to develop an investment with an increased potential expansion when setting under dry conditions. The variation in casting inaccuracy found with the different ceramic liners studied here, used with air investing (Table 5), suggests that investment expansion should be adjusted to the specific ring liner being used.
Acknowledgments. We acknowledge with gratitude the assistance given by Mr. Ken Tyler, who constructed the apparatus; Dr. Peter Brockhurst and Dr.
J Dent Res December 1992
Andrew Ruys, who determined the thermal expansion and the solidus temperature, respectively, of the gold alloy; and Mrs. Anne Lear and Mr. Chris Johnson, who made many of the statistical analyses. REFERENCES Asgar K, Mahler DB, Peyton FA (1955). Hygroscopic technique for inlay casting using controlled water additions. J Prosthet Dent 5:711-724. Earnshaw R (1969). The effect ofrestrictive stress on the hygroscopic setting expansion of gypsum-bonded investments. Aust Dent J 14:22-29. Earnshaw R (1988). The effect of casting ring liners on the potential expansion of a gypsum-bonded investment. J Dent Res 67:1366-1370. MoreyEF, Earnshaw R (1992). The fit ofgold-alloy full-crown castings made with pre-wetted casting ring liners. J Dent Res 71: 1858-1864.
Introduction. A study has already been described (Morey and Earnshaw, 1992) of the effects of pre-wetted asbestos and cellulose ring liners on the dimensional inaccuracy of gold-alloy full-crown castings. In the present investigation, measurements were made of casting inaccuracies obtained with five different ceramic liners, used dry. Under normal atmospheric conditions, ceramic liners have a low water absorption; however, it was found that they readily absorbed large amounts of water under reduced pressure. Because of this, during vacuum investing, water was abstracted from the investment mix, reducing its original W/P ratio, shortening its working time, and increasing both setting and thermal expansions to an uncontrolled extent; the then-wet liner increased the setting expansion still further. Relevant properties of the ceramic liners were therefore measured both on dry specimens and on specimens vacuum-impregnated with water. Similarly, casting inaccuracy tests were made with and without vacuum investing.
Materials and methods. The same batches of gold alloy (G5, Engelhard Industries Pty. Ltd., Thomastown,Australia) andinvestment (Cristobalite Inlay, Sybron/ Received for publication January 22, 1991 Accepted for publication August 7, 1992 2Present address: 1 Wall Avenue, Asquith, NSW, Australia 2077 3Present address: Government Employees Health Fund Dental Clinic, 8 Quay Street, Haymarket, NSW, Australia 2000
Kerr Products, Romulus, MI) were used, manipulated in the same way under the same ambient conditions. Except in tests in which the W/P ratio was varied deliberately, all mixes were made with a W/P ratio of 0.40. Liners tested.-These are listed in Table 1, with an identifying code. Of the five, three (DK, GC, and KF) are sold specifically as dental casting ring liners. The other two (K1 and K2) are industrial materials of the same type-nominally 1 and 2 mm thick, respectively-used principally for high-temperature thermal insulation. Experimental methods.-In general, testing procedures used with the ceramic liners followed those already described for tests on asbestos and cellulose materials (Morey and Earnshaw, 1992). Points of difference are described below. Water uptake.-The water uptake under reduced pressure was determinedby simulation ofthe vacuum-investing process. Casting rings were lined (liners 3 mm short of each end of the ring, 2 mm overlap at the joint) and weighed. They were then attached to the Vac-U-Spat investor, which contained water only, and the usual investing procedure was followed. The ring was then removed, drained of excess water, and re-weighed. Thickness and compressibility.-For measurements ofwet thickness and compressibility, specimens were impregnated with water under a gauge vacuum of 96 kPa maintained for 30 s under vibration. Effective WIP ratio in the ring.-The effect ofvacuum investing in a ring lined with a ceramic material was determined as described in the previous report, except that the liner was not pre-wetted. Since the effect here was to reduce the W/P ratio of the mix, the calibration curve of W/P ratio against working time, measured in unlined casting rings, covered the range from 0.40 down to 0.29 (the thickest mix which could satisfactorily be manipulated). This curve was then used to convert working times recorded in lined rings, filled in air or under vacuum, to effective W/P ratios. Investment expansion.-Laboratory measurements of setting and thermal expansion were made as described previously (Earnshaw, 1988). Here, measurements were first made on specimens mixed in the normal W/P ratio of 0.40, setting against dry ceramic liners; these conditions simulated the use of casting rings with dry ceramic liners and air investing. Then, measurements were repeated under conditions simulating the use of dry ceramic liners with vacuum investing. The investment was mixed in the effective W/P ratio already determined, and the trough was lined with the ceramic materials after they had been vacuum-impregnated with water. Control tests were made on specimens setting against a PTFE trough liner, at a W/P ratio of 0.40. In the tests simulating vacuum investing, it was found that all liners could be removed intact from the set specimens, which had not been possible with all pre-wetted asbestos and cellulose liners or dry ceramic liners. Therefore, in these tests, serial weighings ofthe liners allowed estimation of the amount of water absorbed by the specimens from the liners during setting. Casting inaccuracy.-Castings were made in rings with dry ceramic liners, with vacuum investing, and with vacuum mixing followed by air investing. All investment mixes were made in the standard W/P ratio of 0.40. When vacuum investing was used with K2, the large amount of water absorbed from the mix reduced its working time to such an extent that the investing process, even under vibration, was often not effective in filling the ring, and 1865
EARNSHAW & MOREY
1866
J Dent Res December 1992
TABLE 1
CERAMIC RING-LINING MATERIALS TESTED Name
Code
Supplier
Dentsply Kaoliner
DK
Dentsply/York Division, York, PA 17405-0872
G C Casting Liner
GC
G-C Dental Industrial Corp., Tokyo, Japan
Kerr Flask Liner
KF
Triton Kaowool 1 mm
K1
Triton Kaowool 2 mm
K2
Sybron/Kerr Products, Romulus, MI 48174 Morganite Ceramic Fibres, Alexandria, Australia 2015 Morganite Ceramic Fibres, Alexandria, Australia 2015
defective castings resulted. Of a total of 27 castings, only nine were measurable. Statistical analyses.-All individual sets of results were subjected to ANOVA (GLM procedure, Tukey's Studentized Range HSD test, at both the 0.01 and 0.05 confidence levels). In addition, t tests were used for assessment of the significance of changes in liner compressibilityon wetting; the significance of differences in potential setting, thermal and total expansion; and differences in casting accuracy, when vacuum rather than airinvestingwas used. For tests with air investing, linear regression (SAS PLOT and GLM procedures) was used to seek possible correlations between casting inaccuracy and (a) dry compressibility of the liners and (b) potential investment expansion. For tests with vacuum investing, correlations were sought between measured water uptake of the liners and (a) changes in thickness and compressibility and (b) potential investment expansion; and between casting inaccuracy and (a) effective W/P ratio in the ring, (b) wet and (c) dry compressibility ofthe liners, and (d) potential investment expansion. Multiple linear regression was applied to the relation between casting inaccuracy and combinations of relevant variables, for both air and vacuum investing. In the statistical analyses, the following levels of significance were applied: p < 0.001 = highly significant; 0.001 < p < 0.01 = significant; 0.01 < p < 0.05 = probably significant; p > 0.05 = not significant.
Results. Properties of the liners.-Reproducible results for thickness made multiple measurements unnecessary (Table 2); mean values are given for water uptake under vacuum, and for dry and wet compressibility. There was a significant correlation between thickness
and compressibility, both dry and wet. Unlike the asbestos and cellulose liners tested previously, which all increased in thickness and compressibility on wetting, the different ceramic liners showed either positive or negative changes in thickness and compressibility after impregnation with water; in general, the changes were not as great as with the absorbent liners. Changes in compressibility were highly significant except for K1, where the increase was probably significant. Predictably, there was no correlation between water uptake and change in either thickness or compressibility. Effective WIP ratio in the ring.-The Fig. shows the effect of varying the W/P ratio in the range 0.29 to 0.40 on the working time, measured on control specimens setting in dry, unlined casting rings. Table 3 lists the working times measured on specimens from standard mixes setting in rings lined with the various ceramic materials, initially dry, after being filled under vacuum; also listed are the effective W/P ratios of these latter mixes, estimated from the curve in the Fig. With K2, it was not possible for a value for working time-and hence for effective W/P ratio-to be obtained; the large amount of water it absorbed under vacuum (Table 2) increased the rate ofsetting ofthe mix to such an extent that it was already plastic at the end of the investing process (1 min 20 s). With the other four ceramic liners, simulated vacuum investing gave working times highly significantly less than those of the control specimens setting in unlined rings. There was a probable negative correlation between the water uptake of the liners (Table 2) and the effective W/P ratio in the ring. When rings fitted with dry ceramic liners were filled under atmospheric pressure, no significant changes occurred in working time, nor therefore in W/P ratio. Investment expansion.-Results of measurements of potential expansion are given in Table 4. Where tests simulated conditions of vacuum investing, results are also given for the estimated
TABLE 2 WATER UPTAKE UNDER VACUUM, DRY AND WET THICKNESS, AND DRY AND WET COMPRESSIBILITY OF THE MATERIALS TESTED Liner Water Uptake Under Vacuumt Thickness (mm) Compressibility(mm)t (mL) Dry Wet Dry Wet DK 3.56 ± 0.05 (5) 1.34 1.27 a 0.40 + 0.02 (12) 0.38 ± 0.02 (16) GC 2.29 ± 0.02 (6) 0.90 0.87 0.23 + 0.03 (20) a 0.27 ± 0.03 (21) KF 2.15 + 0.12 (5) 0.83 0.88 0.34 ±0.02 (11) a 0.29 ± 0.02 (13) K1 3.32 + 0.17 (5) 1.05 1.19 a 0.41 ± 0.03 (20) 0.44 + 0.04 (10) K2
5.71 ± 0.28 (5)
2.64
2.54
0.75 ± 0.07 (19)
0.66 ± 0.04 (10)
*See Table 1. tValues for water uptake and compressibility include standard deviations and (in parentheses) numbers of tests. Means marked "a" are not significantly different (p > 0.05).
FIT OF CASTINGS MADE WITH CERAMIC LINERS
Vol. 71 No. 12
amounts of water absorbed by the setting specimen. Tests simulating air investing gave results which showed only small differences compared with those of control tests. Except for DK, increases in setting expansion, though small, were significant (GC) or highly significant. Similar small increases in setting expansion with dry ceramic liners were observed in an earlier investigation (Earnshaw, 1988). No changes in thermal expansion were significant. Only with K2 (8%) and KF (3%) was there a significant increase in total expansion. Under conditions simulating vacuum investing, much more dramatic changes occurred. Tests could not be performed on K2, but with the other four liners increases in setting expansion were highly significant. Concomitant decreases in thermal expansion were significant with all liners except KF. Net increases in total expansion were highly significant. There was a highly significant positive correlation between the amount ofadditional water absorbed by the setting specimen and setting expansion-a relationship recognized by Asgar et al. (1955), who used it as the basis for the "water-added" technique. The amount of absorbed water also showed a probably significant negative correlation with thermal expansion and a significant positive correlation with total potential expansion. Casting inaccuracy.-With air investing, all liners gave undersized castings (Table 5). There was no correlation between casting inaccuracy and potential investment expansion, or between inaccuracy and dry compressibility of the liners. However, there was a II
II
101-
1867
TABLE 3 THE EFFECTIVE W/P RATIO OF THE UNSET INVESTMENT MIX IN LINED CASTING RINGS AFTER VACUUM INVESTING, ESTIMATED FROM ITS WORKING TIME (Control tests were made in unlined rings.)
Liner* Control
Working Time t (min) 10.07 ± 0.12 (3)
Effective W/P Ratio t 0.40
DK
1.60 ± 0.02 (3)
0.29
GC
a 4.41 ± 0.45 (6)
0.31
KF
6.68 ± 0.97 (8)
0.34
K1
a 4.52 ± 0.55 (6)
0.31
< 0.29 K2 0.05).
probably significant correlation between inaccuracy and the combination of both these variables. With vacuum investing, all liners gave improved casting accuracy; the improvement was highly significant with all materials except GC, where it was probably significant. Apart from the exceptions marked in Table 5, all liners, whether used with air or vacuum investing, gave inaccuracies which were significantly or highly significantly different from each other. There were significant correlations between casting inaccuracy and water uptake by the liners under vacuum, wet compressibility, and dry compressibility, and a probably significant correlation between casting inaccuracy and the combination of these three variables. For K2, results could not be obtained for effective W/P ratio or potential investment expansion. However, with the other four liners, there was a probably significant negative correlation between casting inaccuracy and effective W/P ratio and a probably significant positive correlation between inaccuracy and potential investment expansion.
510
Discussion.
2.-c
0 cc
01--
I
I
I
0.30
I 0.40
W/P RATIO Fig.-The relationship between W/P ratio of the investment mix and its rate of setting in unlined rings, shown by the working time.
Previously, it was found (Morey and Earnshaw, 1992) that, wetted in the usual way, asbestos and cellulose liners showed mean water absorptions in the range 1.25 to 2.12 mL. The amounts absorbed by ceramic liners when immersed under vacuum were very much higher (Table 2). Under practical conditions of vacuum investing, this water uptake would occur at the expense ofthe investment mix, greatly reducing its effective W/P ratio (Table 3); this, coupled with the presence of a now-wet liner, caused the large increase in total potential expansion shown in Table 4. The various results for expansion, particularly setting expansion, emphasize the importance of practical conditions being reproduced as closely as possible in laboratory measurements of this property. If ± 0.2% is accepted as an allowable tolerance in casting accuracy (Morey and Earnshaw, 1992), with air investing only K2 gave a mean inaccuracy within this range (Table 5); even with this
J Dent Res December 1992
EARNSHAW & MOREY
1868
TABLE 4 POTENTIAL EXPANSION OF CRISTOBALITE INLAY INVESTMENT, MEASURED UNDER LABORATORY CONDITIONS (During setting, specimens were surrounded by the various lining materials tested. They were then heated to 650'C. Test conditions simulated investing in air, and investing under vacuum. With vacuum investing, estimates are given of the amount of water absorbed from the liners by the setting specimens.) Water Absorbed W/P No. by Setting Thermal Setting Total of Ratio Investing Specimen Expansiont Expansion Expansiont Tests Used (% Liner Method (mL) (%) (%) 6 a 0.37 + 0.02 A a 1.17 ± 0.05 PTFE 0.40 a 1.54 ± 0.04 (control) A a 0.35±0.01 a 1.16±0.01 a 1.51±0.02 DK 3 0.40
GC KF
K1 K2
V
0.29
8
A
0.40
6
V
0.31
4
A
0.40
6
V
0.34
6
A
0.40
6
V
0.31
6
A
0.40
6
V
1.90 ±0.10
1.36 ±0.12 1.06 ± 0.09
1.50 ±0.10
2.58 ±0.04
0.97 ± 0.03
3.56 + 0.05
0.41±0.01
a 1.11±0.03
a 1.52 ±0.02
1.85 ± 0.03
1.03 ± 0.03
2.88 ± 0.04
0.46±0.04
a 1.13±0.04
1.59±0.01
1.13 ± 0.05
a 1.19 ± 0.03
2.32 ± 0.07
0.45 ± 0.03
a 1.11 ± 0.06
a 1.56 ± 0.03
2.08 ± 0.04
1.04 + 0.04
3.11 ± 0.08
0.53 ± 0.03
a 1.14 + 0.06
1.67 ± 0.03
Tests could not be performed, as the effective W/P ratio was not known.
*See Table 1. tA = air investing. V = vacuum investing. tMean and standard deviation. In each column, means marked "a" are not significantly different (p > 0.05).
material, one of the individual castings had a discrepancy worse than -0.2%, and would probably have been unacceptably undersized. With the other four materials, all or most of the castings had a discrepancy worse than -0.2%. With the sprueing method used here, relative casting accuracy depends largely on expansion occurring diametrally in the core of the mold. This, in turn, depends on the potential investment expansion, modified by the presence of the casting ring, its liner, and the wax pattern. When the pattern is invested in air, the setting investment is not exposed to additional water; therefore, the only important liner property is its dry compressibility, which can affect both setting and thermal expansion. There was a probably significant correlation between casting inaccuracy and the combination of potential investment expansion and dry liner compressibility. With vacuum investing, a consideration ofthe effect of the liner on casting inaccuracy is more complex. Liner properties which directly modify investment expansion diametrally in the ring (and therefore casting inaccuracy) are water uptake (which affects setting expansion), wet compressibility (also affecting setting expansion), and dry compressibility (affecting thermal expansion). There was a probably significant correlation between casting inaccuracy and the combination of these relevant liner properties. As a result of the improved accuracy produced by vacuum investing, four of the liners gave mean values within the tolerance of + 0.2%; the exception, KF, wasjust outside it. However, only with K1 and K2 did all individual castings show inaccuracies within ± 0.2%. All other liners produced some castings which, by this
criterion, would have been unacceptably undersize-for DK, 2/10, for GC, 2/9, and for KF, 7/9. The fact that many ofthe differences in mean casting inaccuracy given by the different liners were significant or highly significant, both with air and with vacuum investing (Table 5), emphasizes the importance of liner choice in any casting technique that uses a ceramic liner. The increase in measured setting expansion found with dry liners and vacuum investing was caused by a reduction in WfP ratio, accompanied by the presence of a wet ring liner which induces hygroscopic setting expansion. With the investment used here, the hygroscopic setting expansion can be greatly reduced by mechanical restraint (Earnshaw, 1969); nevertheless, since vacuum investing produced, in most instances, a small decrease in thermal expansion (Table 4), the improved accuracy of these castings must have been the result of effective hygroscopic setting expansion occurring in the investment core. As with pre-wetted liners (Morey and Earnshaw, 1992), it is possible for an indication to be obtained of the effective total expansion shown by the investment core with each of the liners, by summing the values for expected alloy casting shrinkage and the mean casting inaccuracy. These values, for both air and vacuum investing, are shown in Table 6, column 3. Similarly, by then subtracting the values for potential thermal expansion given in Table 4, an estimate can be made of the effective setting expansion occurring in the core (Table 6, column 4). These values are all considerably less than the potential setting expansion ofthe investment with the same liners (Table 4). The percentage reduction in setting expansion found in the investment core is shown in Table 6,
Vol. 71 No. 12
FIT OF CASTINGS MADE WITH CERAMIC LINERS
1869
reduction in setting expansion and dry compressibility of the liners. For DK, GC, KF, and K1, vacuum investing caused increases in effective setting expansion when compared with air investing, reflected in the improvements in casting accuracy shown in Table 5. As with air investing, the various reductions in setting expansion shown in Table 6, column 5, would have been the result ofvariations in liner compressibility. However, there was no correlation between Dimensional Inaccuracy of Castings (%)t percentage reduction in setting expansion and wet compressibility. Liner Air Investing Vacuum Investing A comparison of relative casting accuracy obtained with prewetted asbestos and cellulose liners (Morey and Earnshaw, 1992), Control: - 0.61 ± 0.05 (11) with dry ceramic liners used with air investing, and with vacuum DK a -0.31±0.07(10) a - 0.16 ± 0.05 (10) investing, shows that with only four liners did every casting have inaccuracies within the range + 0.2%. These were pre-wetted KA b - 0.23 ± 0.04 (10) ab - 0.18 +0.05 (9) GC and BE, and K1 and K2 used dry but with vacuum investing. In all other instances, at least some castings were produced which were KF - 0.49 ± 0.08 (10) b - 0.21 ± 0.05 (9) undersized by more than 0.2%. With pre-wetted SH and WM, and with dry DK and KF used with air investing, all castings had Ki a,b - 0.29 ± 0.09 (10) - 0.08 ± 0.06 (9) inaccuracies worse than -0.2%. Considered together, the results of these two studies show the K2 - 0.16±0.05 (12) - 0.05 ±0.08 (9) importance of the choice of ring liner in dental casting techniques. *See Table 1. The simplest example is represented by the use of a dry ceramic tMean, standard deviation, and (in parentheses) number of liner and air investing, since only two variables are involvedtests. potential investment expansion and dry compressibility of the TMorey and Earnshaw, 1993. liner-and both are easily controlled. There was a probably signifiIn each column, means marked with the same letter were not cant correlation between the combination ofthese two variables and significantly different (p > 0.05). castinginaccuracy. Underthese conditions, though, the investment used here did not have a potential expansion sufficiently high to provide an acceptable level of accuracy. This probably also applies column 5. For air investing, it ranged from about 25% (GC) to about 90% to many other currently available inlay casting investments. The use of a wet liner, either pre-wetted or dry (in conjunction (KF). The effect of the wax pattern in reducing setting expansion would have been the same, so that the differing reductions in setting with vacuum investing), increases the mold expansion but introexpansion would have resulted from differences in liner compress- duces more variable factors, and in the latter case has the disadvanibility. However, there was no correlation between percentage tage of reducing the working time of the investment. Even so, only
TABLE 5 THE DIMENSIONAL INACCURACY OF GOLD-ALLOY FULL-CROWN CASTINGS MADE WITH NO RING LINER (CONTROL) AND EACH OF THE CERAMIC LINERS TESTED
TABLE 6 CALCULATED VALUES FOR EFFECTIVE TOTAL EXPANSION OF THE INVESTMENT CORE, ITS EFFECTIVE SETTING EXPANSION, AND THE PERCENTAGE REDUCTION IN SETTING EXPANSION OCCURRING IN THE CORE UNDER PRACTICAL CONDITIONS
Investing Method Air
Vacuum
Effective Total Expansion t
Effective Setting Expansion t
Percentage Reduction in
Setting Expansion §
Liner DK
(%) 1.34
(%) 0.18
49
GC
1.42
0.31
24
KF
1.16
0.03
93
1K1 1K2
1.36
0.25
44
1.49
0.35
34
DK
1.49
0.52
80
GC
1.47
0.44
76
KF
1.44
0.25
78
1K1
1.57
0.53
75
K2 Calculations not possible, since values for potential expansion were not available
*See Table 1. tSum of alloy shrinkage and casting inaccuracy. tDifference between effective total expansion and potential thermal expansion. §Percentage difference between potential and effective setting expansion.
1870
EARNSHAW & MOREY
four of the 11 wet liners studied here consistently produced fullcrown castings with an acceptable dimensional accuracy. A logical approach to improvement in the dimensional accuracy of gold-alloy full-crown castings would be to use a casting ring with a ceramic liner in conjunction with air investing, since this involves only two easily controlled variables. It would be necessary to develop an investment with an increased potential expansion when setting under dry conditions. The variation in casting inaccuracy found with the different ceramic liners studied here, used with air investing (Table 5), suggests that investment expansion should be adjusted to the specific ring liner being used.
Acknowledgments. We acknowledge with gratitude the assistance given by Mr. Ken Tyler, who constructed the apparatus; Dr. Peter Brockhurst and Dr.
J Dent Res December 1992
Andrew Ruys, who determined the thermal expansion and the solidus temperature, respectively, of the gold alloy; and Mrs. Anne Lear and Mr. Chris Johnson, who made many of the statistical analyses. REFERENCES Asgar K, Mahler DB, Peyton FA (1955). Hygroscopic technique for inlay casting using controlled water additions. J Prosthet Dent 5:711-724. Earnshaw R (1969). The effect ofrestrictive stress on the hygroscopic setting expansion of gypsum-bonded investments. Aust Dent J 14:22-29. Earnshaw R (1988). The effect of casting ring liners on the potential expansion of a gypsum-bonded investment. J Dent Res 67:1366-1370. MoreyEF, Earnshaw R (1992). The fit ofgold-alloy full-crown castings made with pre-wetted casting ring liners. J Dent Res 71: 1858-1864.
