Sintering of dental porcelain: effect of time and temperature on appearance

and porosity
K.C. Cheung
a
, B.W. Darvell
b,
*
a
Dental Technology, The Prince Philip Dental Hospital, Hong Kong
b
Dental Materials Science, Faculty of Dentistry, The University of Hong Kong, Hong Kong
Received 27 July 2000; accepted 20 February 2001
Abstract
Objective: Mechanical condensation of powder-method dental porcelains can only achieve a limited effectthe majority of consolidation
and porosity elimination is achieved by sintering. However, there is a surprising lack of information on the process in the literature, and the
effects of the two basic conditions of sintering, time and temperature, are very poorly described despite theoretical expectations. The present
study was to investigate the effects of these conditions on porosity and to consider its relationship to the recognizable ring stages of low and
high biscuit, and other aspects of appearance, with a view to schedule recommendations.
Methods: The variation of appearance, pore form and percentage porosity with sintering time and temperature was studied for ve dental
dentine porcelains, two aluminous: Alpha (Vita Zahnfabrik, Bad Sackingen, Germany) and Vitadur-N (Vita); and three feldspathic: Omega
(Vita), VMK68 (Vita) and Carmen (Esprident, Ispringen, Germany). Disc specimens were sintered for 01000 min over 75011008C in a
systematic search pattern to establish limits of acceptable appearance (high biscuit to not quite slumped). Porosity was measured using an
image analyser on specimens red for sintering times of 24 and 30 s; 1, 3, 6 and 30 min; 1, 5, 10, 15 and 20 h; with sintering temperatures
from 750 to 9508C for Carmen and 80010508C for the others, all with 508C increments. Measurements were made on a ground and polished
surface (1 mm diamond), ve elds on each of ve specimens per condition. The percentage porosity, pore count, median pore area and pore
size distribution were analyzed.
Results: The boundaries of the acceptable appearance areas in maps of sintering temperature vs. sintering time were clearly delineated;
analysis showed that they may be related to the activation energy of the diffusive processes occurring during sintering. Minimum porosity
was obtained at high temperature and short time, close to but not consistently coincident with the manufacturers' recommendations. There is
also a conductivity-related minimum sintering time at high temperatures. The theoretical reduction on prolonging sintering at high tempera-
ture was not a general result, and porosity increased markedly in the feldspathic porcelains, particularly VMK-68.
Signicance: The reduction of porosity of dentine porcelain is much more sensitive to temperature than to time. It is possible system-
atically to identify optimum conditions using the present approach. Detailed study of the effects of formulation on the sintering process may
be made by reference to the activation energy. q 2002 Academy of Dental Materials. Published by Elsevier Science Ltd. All rights reserved.
Keywords: Dental porcelain; Sinter; Porosity; Activation energy
1. Introduction
The porosity of dental porcelain needs to be minimized to
attain the best optical appearance and strength since pores
scatter light, decreasing translucency, and can act as crack
initiators with high stress concentration, lowering strength
in tension and shear. Control of porosity would therefore
appear to be a fundamental consideration in the design and
processing of a such a porcelain.
Mechanical vibration and blotting are commonly used
in the rst instance to reduce the volume fraction of
porosity in a powder-method dental porcelain compact,
the number and size of voids remaining depending on
the particle size distribution [1,2], although the effect of
vibration on the reduction of porosity is limited [3,4]. To
enable a better simulation of tooth tissue, dental porcelain
restorations are commonly built in layers consisting of
different frit formulations. However, vibration tends to
cause the porcelain compact to slump (shear thinning of
a plastic dilatant system). Thus, surface detail is lost and
the porcelain layers may alter in position, spoiling the
work. Hence, even with an ideal particle size distribution,
the use of vibration must be carefully controlled, and
Dental Materials 18 (2002) 163173
dental
materials
0109-5641/02/$22.00 + 0.00 q 2002 Academy of Dental Materials. Published by Elsevier Science Ltd. All rights reserved.
PII: S0109-5641(01)00038-0
www.elsevier.com/locate/dental
* Corresponding author: Dental Materials Science, Prince Philip Dental
Hospital, 34 Hospital Road, Hong Kong. Tel: 1852-2859-0303; fax: 1852-
2548-9464.
E-mail address: hrdubwd@hkucc.hku.hk (B.W. Darvell).
substantial porosity remains, which is the cause of the
ring shrinkage.
The bulk of the porosity reduction occurs therefore in the
sintering process, the purpose of which is to create as far as
possible a non-porous solid. Although sintering is often
referred to as a solid-state process, here it is driven by
surface tension as a result of partial melting of crystalline
phases or sufcient softening of glassy material [57], the
pores eventually becoming spherical and contracting in
volume. This reduction of pore volume depends on sintering
time [68], temperature [7,8] and atmosphere [912], and
the viscosity of the melt [5,7,13]. The viscosity of silicate
melts is strongly dependent upon temperature [7] and its
chemical composition [14]. Although the reduction of
porosity is therefore expected to be promoted by longer
sintering time or increasing sintering temperature [15],
or both, Hodson [16] and Jones and Wilson [10] reported
porosity increase with increase in sintering tempera-
ture. Minimum porosities of 0.85% [16], 0.47% [10]
and 1.6% [15] have been reported. Without any measure-
ment, Olorunfemi [17], observing three specimens per
condition under a microscope, estimated that minimum
porosity was obtained by sintering for 1 min longer than
the recommendation.
McLean and Hughes [18] found that a reasonable corre-
lation was obtained between the results of the methods of
apparent specic gravity, heavy liquid gradient column and
microscopic point counting. The rst two methods depend
upon the assumptions that the pores are closed and that no
chemical changes, and therefore no changes in the density of
solid phases, have occurred during ring. However, there is
variation in the density of dental porcelains depending upon
shade and ring temperature [19].
Clearly, both time and temperature are important factors
in the ring of such ceramic materials. However, there
seems to have been no systematic analysis of the effects
of these factors on the porosity of dental porcelains, and
what mention there is of porosity is incomplete or contra-
dictory. In addition, some of the methods that have been
used do not permit the characteristics of the porosity, such
as shape, number of pores, pore size distribution or other
statistics to be analyzed.
Porosity is dened as the ratio of pore volume, V
p
, to the
total (bulk) volume V
s
of pore and solid together [20]:
e
V
p
V
p
1V
s
: 1
The volume of the closed pores of a sintered dental porce-
lain cannot be measured directly. However, direct measure-
ment of porosity can be done by a method of lower
dimensionality, that is, by working on a plane section of
an object, the section area of the porosity A
p
behaves propor-
tionately with respect to the solid area, A
s
[20]:
e
A
p
A
p
1A
s
: 2
This approach is unaffected by density, the connectedness
or otherwise of the porosity, nor does it require calibration.
The present study was thus to investigate the effect of
sintering time and temperature on the porosity of powder-
method dental porcelains using area counting. The relation-
ship of porosity to the recognizable ring stages known as
low and high biscuit or bisque, and other aspects of appear-
ance, could then be considered with a view to schedule
recommendations.
2. Materials and methods
Two aluminous porcelains and three feldspathic porce-
lains were studied (Table 1), using the corresponding liquid
binders: Modelling Liquid MV (Esprident) for Carmen and
Modelling Liquid (Vita) for the others. No instructions for
ring pressure were given.
2.1. Specimen preparation
Circular disc specimens, 8 mm diameter and 2 mm thick,
were prepared using the vibration and continuous blotting
technique for consolidation in a sheet acrylic mould bolted
to a baseplate. Moulds were initially overlled, specimens
being levelled by scraping before being ejected with a at-
ended plunger. Powder compacts showing any signs of
distortion, cracking or crumbling were rejected. After
drying for one hour at room temperature (to avoid disruption
of these thicker than normal compacts), ring was done in a
K.C. Cheung, B.W. Darvell / Dental Materials 18 (2002) 163173 164
Table 1
Firing schedules for the porcelains studied, and manufacturer's recommended sintering temperatures and times
Manufacturer
a
Type
b
Preheat time/min Initial temp./8C Heating rate/8C min
21
Recommended
sintering temp./8C
Recommended
sintering time/min
Alpha Vita A 6 600 55 960 1
Vitadur-N Vita A 6 600 55 960 1
Omega Vita F 6 600 55 930 1
VMK68 Vita F 6 600 55 930 1
Carmen Espr. F 6 400 53 860 1
a
Vita: Vita Zahnfabrik, Bad Sackingen, Germany; Espr.: Esprident, Ispringen, Germany.
b
A: aluminous; F: feldspathic.
programmable vacuum porcelain furnace (Multimat Mach
2, Dentsply DeTrey, Dreieich, Germany) on a mat (Fibrous
Pad Firing Support, Vita). The general form of the ring
schedule is shown in Fig. 1. The pre-heating times, initial
temperatures and heating rates followed the manufacturers'
recommendations (Table 1). For all rings, the pressure of
the ring chamber was reduced to 50 hPa before the furnace
temperature was increased from the initial temperature. The
atmospheric pressure was restored when the sintering
temperature was reached. The accuracy of the furnace
temperature (claimed to be ^28C) was veried by the silver
wire fusion test and measurement of D.C. voltage output
produced by the furnace thermocouple. Errors of 11 to
128C were found in the range 80012008C. These were
considered to be insignicant for the present purposes.
Two series of specimens were red: the rst to determine
the ranges of `acceptable ring' based on visual inspection,
the second for use in image analysis of the porosity.
2.2. Acceptable ring
Since in practice the assessment of whether a porcelain
has been appropriately red is based on a visual inspection
(and a suitable quantitative approach is unknown), the
sintering time and temperature ranges yielding acceptable
results were rst determined. The criteria for an acceptable
result were set as sharp edges with an egg-shell sheen [16]
or glossy appearance [21]. A chalky appearance and
rounded edges or slumping are obtained for under- and
over-ring respectively. Using a temperature step size of
108C and roughly logarithmically-spaced sintering times
from 0 to 20 h, the upper and lower bounds of the `accept-
able' ring conditions region were traced staircase fashion
using single specimens.
2.3. Porosity
With reference to the results of the `acceptable ring'
series, for each porcelain a group of ve specimens was
red for sintering times of 24 and 30 s; 1, 3, 6 and
30 min; 1, 5, 10, 15 and 20 h at sintering temperatures
ranging from 750 to 9508C for Carmen and 80010508C
for the others, all with 508C increments.
Specimens were then embedded as a ring schedule
group in acrylic thermoplastic resin (Resin 3, Struers,
Copenhagen, Denmark) in an embedding press (Duplopress-
1, Struers). To prevent cracking of specimens under pres-
sure in the embedding process, about 0.2 mm was rst
removed by wet-grinding by hand to leave a at surface.
To assist the retention of at and sharp-edged specimens
during subsequent grinding and polishing, the resin mounts
were edge-hardened by surrounding specimens to their
thickness with a mixture of acrylic resin and porcelain
powders in a ratio of 60:40% by bulk volume. The pro-
cessed blocks were wet-ground to a at and smooth surface
on 600 and 1200 grit silicon carbide abrasive papers
(Carbimet, Buehler, Lake Bluff, IL, USA) on a polishing
machine (Vibromet, Buehler), then polished using 3 and
1 mm diamond polishing pastes (Metadi II, Buehler) in
sequence on an ultra-short napped cloth (Texmet, Buehler).
Blocks were ultrasonically cleaned in deionized water
between stages. The grinding and polishing work was
directly checked regularly with an incident light microscope
(Universal M, Carl Zeiss, Oberkochen, Germany) at a
magnication of 50 . Individual grinding or polishing
steps were repeated until the required result was achieved.
The quality of the result was judged from sharp pore edges
and freedom from visible scratches. At this stage, the shapes
of the pores were classied and recorded as continuous,
irregular, or circular.
Porosity data were collected for ve elds without over-
lap on each of the ve specimens of a group using computer-
based image analysis. The image was captured by a video
camera (TK-1080E, JVC, Tokyo, Japan) through an inci-
dent light microscope (Orthoplan-pol, Leitz, Wetzlar,
Germany) at a nominal magnication of 140 . Sufcient
grey-scale contrast was present for automatic measurement
without further preparation. Pore size and count determina-
tions were done in software (NIH Image 1.60, Rasband W,
National Institute of Health, USA) on a personal computer
(Macintosh IIci, Apple Computer, Cupertino CA, USA).
The smallest recordable pore area was 7.2 mm
2
, and the
area of a eld 2.84 mm
2
. The measurement elds were no
closer than 0.5 mm from the specimen edge; no visually
detectable variation in porosity occurred over the whole
area, all elds being individually representative. In extreme
cases, a measurement might fail because of too long a peri-
meter or too large a pore area, in which case pore area was
determined as the complement after image inversion. The
volume fraction porosity was calculated from the pore size
distribution data according to Eq. (2) for each of the 25
K.C. Cheung, B.W. Darvell / Dental Materials 18 (2002) 163173 165
Fig. 1. Schematic diagram of the form of the ring schedule used.
elds as the ratio of pore area to eld area and a variety of
statistics computed. These data were converted to contour
plots in sintering time and temperature (SigmaPlot v6,
SPSS, Chicago IL, USA).
3. Results
3.1. Acceptable ring
There were two patterns of acceptable ranges of sintering
times and temperatures, corresponding to whether the
porcelain was aluminous or feldspathic, viz. the acceptable
temperature range was narrow at a given sintering time in
the case of the aluminous porcelains, wide for the felds-
pathic (Fig. 2). Except at very short times for the feldspathic
porcelains, two nearly linear (and nearly parallel) plots were
obtained for both the highest and the lowest acceptable
sintering temperature against logarithm of sintering time
over a substantial range of times. For Omega in particular,
the lower bound appears to have a different slope for time
below 1 min. A clear minimum possible sintering tempera-
ture was present in each case, but while the upper bound for
the aluminous porcelains seems to intersect this (at around
1000 min), for the feldspathic porcelain it levelled off some
70908C higher than this minimum, leaving the acceptable
region apparently open on the right.
3.2. Porosity topology
The maps of sintering times and temperatures used for the
porosity specimens are given in Fig. 3, with the acceptable
appearance bounds superimposed for reference. At the
lowest time and temperature combinations (black spots) a
continuous network of pores was present, the specimens
being fragile. This corresponded to the upper limit of
what is termed the low-biscuit stage of ring, and the pore
area determination required image inversion. The lower
bound of the `acceptable appearance' corresponded very
closely to the upper limit of the presence of irregular
pores (open diamonds), i.e. an acceptable ring on the
basis of appearance corresponded to the presence of spheri-
cal porosity. This boundary may be identied as represent-
ing the high-biscuit condition. At high enough temperatures,
and long enough times, specimens melted sufciently to
cause them to ball up into a shape incapable of permitting
the standardized porosity measurement elds (open circles)
and no data was collected. However, it is clear from the 24 s
data that there is a minimum time required for porosity to
become spherical, irrespective of the external appearance.
3.3. Porosity
The volume percentage porosity contour plots are shown
in Fig. 4, again with the acceptable appearance bounds
superimposed for reference. Generally, porosity decreased
with increase in sintering temperature at a given time, with
rapid changes occurring between the low- and high-biscuit
conditions. However, this was not monotonic. At short
sintering times there is evidence of a minimum porosity
being attained and then increasing again, although this
mostly was only a slight effect it is seen very clearly in
the case of VMK-68. The turning point appears to occur
at a different temperature for each porcelain. Otherwise,
within the acceptable appearance bounds, increasing the
sintering temperature had only a slight effect on porosity.
However, it is clear that for VMK-68 and Carmen there is a
distinct tendency to increase porosity with increasing sinter-
ing time at a given temperature. There is no evidence of this
effect for the aluminous porcelains. In addition, there is
evident a dramatic and rapid decrease in porosity on going
from 24 to 30 s sintering corresponding to the change in
pore topology noted above.
3.4. Other statistics
Similar contour plots were inspected for statistics such as
median pore size and pore count. However, since all such
statistics were highly correlated no further useful informa-
tion was extractablethe pattern of contours was very
similar to that of the volumetric measure in each case.
Pore size distributions, cumulative porosity and cumulative
count plots were also inspected, but no notable features were
identied.
4. Discussion
4.1. Method
The microscopic method of porosity determination was
found to be a high sensitivity, fast, accurate and repro-
ducible method. It has the advantage of being a direct
measurement of porosity, and permits a detailed picture of
the characteristics of the porosity because all (visible) indi-
vidual pores are measured, unconfounded by the dimensions
and densities of solid phases. The accuracy is, however,
contingent on a time-consuming specimen preparation if
errors due to curvature or break-out are to be avoided. In
particular, preservation of sharp pore edges required careful
selection of grinding and polishing media and schedules.
Since the porosity of the green compact is expected to be
below 50% [18] values over that for a coherent specimen,
i.e. one in the low biscuit stage [22] or initial sintering stage
[23], were taken to be indicative of the loss of some particles
during grinding, polishing and ultrasonic cleaning, since the
particles were only weakly joined at the neck formation
stage. If it was worthwhile pursuing more detailed data
in this region, inltration would be necessary, with, for
example, liquid epoxy resin, which is to be cured before
grinding.
The smallest pore area that could be recorded was
7.2 mm
2
, although, course, smaller pores must have been
present. Although higher magnication could be used, this
K.C. Cheung, B.W. Darvell / Dental Materials 18 (2002) 163173 166
K.C. Cheung, B.W. Darvell / Dental Materials 18 (2002) 163173 167
Fig. 2. Results of acceptable ring conditions tests. Points on left-most boundary represent zero sintering time. (Time on a logarithmic scale.)
K.C. Cheung, B.W. Darvell / Dental Materials 18 (2002) 163173 168
Fig. 3. Porosity topology vs. ring conditions. Each point represents conditions for which specimens were examined internally for pore shape. Superimposed is
the envelope of acceptable ring conditions for the corresponding product (from Fig. 2). (Time on a logarithmic scale.)
K.C. Cheung, B.W. Darvell / Dental Materials 18 (2002) 163173 169
Fig. 4. Fitted porosity contours (%) for data obtained under ring conditions mapped in Fig. 3. Superimposed is the envelope of acceptable ring conditions for
the corresponding product (from Fig. 2). (Time on a logarithmic scale.)
would compromise the measurement as more large pores
would be intersected at the boundary of a eld, and many
more elds would be required for a representative sample to
be obtained. Informal estimates suggested that the present
measurements are in error by missing no more than about 2
or 3% of the value of the recorded porosity.
4.2. Acceptable ring
The high biscuit [22] or nal sintering stage [23] can be
clearly identied with the rst occurrence of an acceptable
appearance, corresponding closely to the point at which
spherical pores are formed; the correlation apparent from
Fig. 3 is very good. Although there is only a limited effect of
temperature on the surface tension of silica [24,25], the
viscosity of dental porcelain is temperature dependent
[26,27], and the attainment of pore sphericity is thus a
matter of exposure-time at temperature, driven by the
surface energy minimization. Evidently, the over-red
condition, i.e. exceeding the upper acceptable bound, has
no internal correlate, but it is certainly of importance as an
observational means of determining when ow is excessive.
Taken together with the lower bound observation, this
suggests that it is a reasonable eld-calibration method
that may be easily applied by users, that is, to allow for
variation in temperature calibration, heating rates, timing
and even porcelain, given that state-of-the-art equipment
and conditions are not universally available. Even so, it is
to be noted that appearance is a supercial attribute.
The process of sintering, as monitored by the change in
appearance, is essentially dependent on diffusive processes,
ow driven by the surface tension of the matrix material and
limited by its viscosity. Because it is diffusive, the overall
process is associated with an activation energy. The accept-
able appearance limits represent for a given temperature
two separate point estimates of the rate at which sintering
is occurring. Thus, writing the general Arrhenius equation
as:
D Ae
2E
a
=RT
3
where D is the `diffusivity' or diffusion coefcient, which is
expressed as a rate, i.e. m
2
s
21
. It is not certain at present
how this needs to be scaled, but since it must have the same
dimensions as the pre-exponential factor, A, and since the
next step eliminates the scaling from consideration, it is of
no consequence for the moment. E
a
is then the activation
energy, T the absolute temperature, and R the molar gas
constant, 8.31451 J K
21
mol
21
. Thus, taking logs:
log D log A 2
E
a
RT
4
from which it is seen that a plot of (arbitrarily scaled) log
rate vs. 1/T yields a plot of slope 2E
a
/R, whence the activa-
tion energy is obtained directly. Applying this to the bound-
ary data extracted from Fig. 2, the plots shown in Fig. 5 were
obtained, in which, for example, the good linearity of the
high temperature boundary plots for the aluminous porce-
lains over three orders of magnitude in time is taken as
evidence for the underlying applicability of the model. It
is also seen in the case of the aluminous porcelains and for
Omega and VMK-68 that there are two clearly separate
segments to the low-temperature lines in addition to the
abrupt change of slope at the low-temperature boundary.
Such slope discontinuities are strongly suggestive that
they may be correlated with phase changes such as decom-
positions (as for example, occurs with feldspar, although the
temperature for this is too high at 11508C, 0.70 kK
21
) and
melting, whether of pure substances or eutectics. However,
it is not yet possible to make any such assignment, no
obvious candidates being identiable without details of
the formulations. Such assignments are implicitly asso-
ciated with increasing ease of diffusion, presumably by
decreasing the connectivity of the glass network. On the
other hand, there is no fundamental reason why a reaction
should not increase the connectivity, increasing the activa-
tion energy for the process, that is, by changing cation
concentration, for example. Suggestions that there might
be increases in slope in one or two places in the present
data are too weak, however, at present to take this further.
To put the estimated activation energies in context,
comparison might be made with the Si-O bond energy,
which is of the order of 368 kJ mol
21
[28]. Hence the
present observations are consistent with bond exchange
through anionic attack permitting the ow. It may thus be
possible to approach the design of dental porcelains for
sintering processes, or glasses for inltration techniques,
by reference to such activation energy and rate data, the
advantage being that the proxy observations of appearance
are cheap and easy to obtain, requiring no specialized equip-
ment beyond a suitable furnace. Furthermore, the role of
various additives in the formulation can be studied in an
objective fashion, with direct reference to the chemistry of
the processes being possible.
4.3. Porosity
Porosity is expected to decrease with sintering time, and
with increasing temperature, as trapped gases escape by
diffusion through the matrix, driven by the dissolution
occurring under the Laplacian excess bubble pressure
according to Henry's Law [29], especially as the viscosity
of the matrix decreases. This is indeed seen for the region
between the formation of a closed pore system, correspond-
ing roughly to the 30% porosity contour (Figs. 3 and 4), and
the lower acceptable appearance bound, that is, where the
contours slope down to the right (Fig. 4). However, there is
no evidence for this process being dominant in the present
data in the acceptable appearance region, i.e. above the
lower bound, except for Omega, where a distinct minimum
is present at about 9508C and 3 min, and similarly for
Carmen at 8508C at about the same time. Otherwise the
conclusion is that there is little effect, but with the notable
K.C. Cheung, B.W. Darvell / Dental Materials 18 (2002) 163173 170
K.C. Cheung, B.W. Darvell / Dental Materials 18 (2002) 163173 171
Fig. 5. Relative sintering rate vs. reciprocal temperature. Points plotted are the outermost `acceptable appearance' points from Fig. 2. The slopes of tted line
segments represent estimates of the activation energy (kJ mol
21
) for the corresponding process. (Reciprocal time on a logarithmic scale.)
exception of VMK-68. Here, there is a steady increase of
porosity with time under all `acceptable appearance' condi-
tions, a pattern that starts to appear for Carmen after about
10 min.
Porosity could be increased in two different ways during
sintering. Bubbles would tend to increase in size on heating
due to expansion towards pressure equilibrium, since the
viscosity of the porcelain decreases with temperature
[9,10], the pressure opposing the surface tension-driven-
contraction. This effect must be traded off kinetically
against the dissolution of the gas in the matrix. Neverthe-
less, the viscosity is still high enough, of the order of
3 MPa.s at 10008C [26], to limit the effect of such pressure
expansion and only a slight increase in porosity is obtained,
if anything, in most cases.
A more probable cause is the generation of a gaseous
product by a chemical reaction occurring during sintering
[11]. It would seem unlikely that carbonates could be
included to any extent in the frit formulation, or form by
adsorption from the atmosphere during storage, although
this is a testable hypothesis. However, where frits are
formed by rst thermally-shocking a melt or sintered mass
by pouring into water, reactions are possible. For example, it
is known that silica glasses hydrolyze in contact with atmo-
spheric moisture, leaving hydroxyl groups on the surface
[30]. This reaction would presumably occur more readily
in water or water vapour at high temperatureas during
shocking. Thus, on subsequent heating dehydration might
be anticipated, generating gas and thus increased porosity.
Again, this is testable and clearly would have implications
for the manufacturing process or the composition of the frit.
Kaolin, of course, is an already hydrous mineral,
Al
2
O
3
.2SiO
2
.2H
2
O, although its decomposition occurs at
4508C, too low to be relevant here even if it is present,
and in any case, for all except Carmen, would have occurred
during the pre-heat stage (Table 1) and so not be an issue
now.
Although the porosity is increased by that process, two
points may be made. Firstly, there was no evidence of an
increase in the number of pores, merely their total volume.
This speaks of a slow rather than abrupt process of decom-
position at a specic temperature, allowing diffusion to
bubbles to occur, the partial pressure of the putative water
being greater in the matrix than in the bubble, whose
Laplacian excess pressure decreases rapidly as 1/r as they
swell. Secondly, this effect was only signicant for Omega
and Carmen at sintering times that must be considered too
long to be relevant to practice. Even so, in the case of VMK-
68 it appears to be operating even at very short times at all
temperatures above about 8508C. This cannot be ignored.
A further notable feature is the rapid decrease in porosity
on increasing the sintering time from 24 to 30 s for all
products, at least for the temperatures tested. It is assumed
that the principal controlling factor here is the conductivity
of the porcelain, it requiring a certain time for the central
region to reach the intended temperature. Since the gradient
change in going from 900 to 10008C, say, is of the order of
10% increase only, timing variation will not be readily
detectable in this experiment. The implication nevertheless
is of a minimum sintering time if the porcelain is to be
heated sufciently uniformly for the porosity reduction to
be effective overall.
4.4. Standards
According to the relevant British Standard [31], the maxi-
mum tolerable porosity is 3%, presumably reecting what
was then considered a reasonable minimum attainable
condition. Thus, the region meeting this requirement (see
Fig. 4) and meeting the acceptable appearance criteria in
Fig. 2 represents the range of appropriate sintering condi-
tions. In all cases, the 3% boundary is seen to be remarkably
lax in comparison with the attainable minimum porosity.
It should also be remembered that there are other limita-
tions. Porcelain fused to metal requires, in the rst instance,
that the alloy solidus is above the ring temperature (for
example, in current alloy ranges, solidus temperatures
from 920 to 12908C at least are known), but in addition to
this some leeway is required to avoid creep. This should be
taken into account in future work.
The manufacturers' recommended ring conditions are
also shown in Fig. 4 (solid circles). It can be seen that in
the case of the aluminous porcelains these t the limits set
by appearance rather exactly and minimize the porosity for
the process as studied. For the feldspathic porcelains the
situation is less clear. Neither Omega nor Carmen attained
minimum porosity for the recommended conditions,
although this could be done easily by ring for 3 or
2.5 min, respectively instead of 1. For VMK-68 it would
appear that better results could be obtained at a slightly
lower temperature (say 9108C) and shorter time (say,
40 s). Whether or not there are other issues to consider in
setting the ring conditions, it would appear from this that a
more detailed appraisal is capable of being made and greater
accuracy attained in nding an optimum.
5. Conclusions
The loss of the connectedness of the porosity of the
compact corresponds to the low biscuit stage of sintering,
while attainment of pore sphericity, the lower boundary of
the acceptable appearance region, may be identied as the
high biscuit stage. There is a clear minimum temperature for
sintering to acceptable appearance, and above this a wide
range of sintering times is available. There is also a mini-
mum sintering time even at high temperatures. However,
minimum porosity is obtained only under a very narrow
range of conditions, essentially high temperature and short
sintering time. Some renement of the recommended times
and temperatures may be made when a system has been
systematically mapped.
K.C. Cheung, B.W. Darvell / Dental Materials 18 (2002) 163173 172
References
[1] McGeary RK. Mechanical packing of spherical particles. J Amer
Ceram Soc 1961;44(10):51322.
[2] Rasmussen ST, Nagaji-Okumu W, Boenke K, O'Brien WJ. Optimum
particle size distribution for reduced sintering shrinkage of a dental
porcelain. Dent Mater 1997;13(1):4350.
[3] Gill JR. Methods and results in condensation of dental porcelain.
J Amer Dent Assoc 1932;19:114754.
[4] Evans DB, Barghi N, Malloy CM, Windeler AS. The inuence of
condensation method on porosity and shade of body porcelain.
J Prosthet Dent 1990;63(4):3809.
[5] Mackenzie JK, Shuttleworth R. A phenomenological theory of sinter-
ing. Proc Phys Soc B 1949;62:83352.
[6] Herring C. Surface tension as a motivation for sintering, Chap. 8. In:
Kingston WE, editor. The physics of powder metallurgy: a symposium
held at Bayside, 1st ed. New York: McGraw-Hill, 1951. p. 14379.
[7] Frenkel J. Viscous ow of crystalline bodies under the action of
surface tension. J Tech Phys 1945;9(5):38591.
[8] Ivensen VA. Theory and technology of sintering, thermal and chemi-
cothermal treatment processes. Powder Metallurgy (USSR) 1970;
4(88):2025.
[9] Vines RF, Semmelman JO, Lee PW, Fonvielle FP. Mechanisms
involved in securing dense, vitried ceramics from preshaped partly
crystalline bodies. J Amer Ceram Soc 1958;41(8):3049.
[10] Jones DW, Wilson HJ. Porosity in dental ceramics. Brit Dent J
1975;138:1621.
[11] Rosenkranz V, Frischat GH. Shrinking of a gas bubble in a glass melt
under microgravity conditions. In: Material Sciences under Micro-
gravity Results of Spacelab-1: Fifth European Symposium Papers.
Schloss Elmau, Germany, 1984. p. 3537.
[12] Wilson HJ, Whitehead FIH. Comparison of some physical properties
of dental porcelains. Dent Practit 1967;17(10):3504.
[13] Shaler AJ, Wulff J. Mechanism of sintering. J Indust Eng Chem
1948;40(5):83842.
[14] Bottinga Y, Weill DF. The viscosity of magmatic silicate liquids: a
model for calculation. Amer J Sci 1972;272:43875.
[15] Anusavice KJ, Lee RB. Effect of ring temperature and water
exposure on crack propagation in unglazed porcelain. J Dent Res
1989;68(6):107581.
[16] Hodson JT. Some physical properties of three dental porcelains.
J Prosthet Dent 1959;9(2):32535.
[17] Olorunfemi BO. Sintering aws in Flexoceram inlay porcelain. Afri
Dent J 1996;10:911.
[18] McLean JW, Hughes TH. The reinforcement of dental porcelain with
ceramic oxides. Brit Dent J 1965;119(6):25167.
[19] Semmelman JO. A method of measuring porosity of porcelain teeth.
J Dent Res 1957;36(6):9459.
[20] Haynes JM. Porosity: characteristics and techniques of investigation.
In: Bever MB, editor. Encyclopedia of materials science and engi-
neering, vol. 5. Oxford: Pergamon, 1986. p. 38309.
[21] The VitaVMK Technique: working instructions. Vita Zahnfabrik, H
Rauter, Bad Sackingen, Germany [undated].
[22] McLean JW. The science and art of dental ceramics I: the nature
of dental ceramics and their clinical use. Chicago: Quintessence,
1979.
[23] Coble RL. Sintering crystalline solids: I. Intermediate and nal state
diffusion models. J Appl Phys 1961;32(5):78792.
[24] Kingery WD. Introduction to ceramics. 2nd ed. New York: Wiley,
1976.
[25] Cambier F. Vitrication, Chap. 15. In: Brook RJ, editor. Materials
science and technology, vol. 17B: processing of ceramics Part II.
Weinheim: VCH, 1996. p. 12344.
[26] Asaoka K, Kon M, Kuwayama N. Viscosity of dental porcelains in
glass transition range. Dent Mater J 1990;9(2):193202.
[27] Bertolotti RL, Shelby JE. Viscosity of dental porcelain as a function
of temperature. J Dent Res 1979;58(10):20014.
[28] Cotton FA, Wilkinson G. Advanced inorganic chemistry. 5th ed. New
York: Wiley, 1988.
[29] Pitzer KS, Brewer L. Thermodynamics. 2nd ed. New York: McGraw-
Hill, 1961.
[30] Charles RJ. Static fatigue of glass I. J Appl Phys 1958;29(11):1549
53.
[31] BS 5612: Specication for dental porcelains for jacket crowns. British
Standards Institution, London, 1978.
K.C. Cheung, B.W. Darvell / Dental Materials 18 (2002) 163173 173