Journal

of
Dentistry
Journal of Dentistry 27 (1999) 89–99

Review

Thermal cycling procedures for laboratory testing of dental restorations

M.S. Gale a,*, B.W. Darvell b
a
Conservative Dentistry, The University of Hong Kong, Faculty of Dentistry, Hong Kong
b
Dental Materials Science, The University of Hong Kong, Faculty of Dentistry, Hong Kong
Received 20 November 1997; revised 20 January 1998; accepted 12 May 1998

Abstract

Objectives: Exposure of restorations in extracted teeth to cyclic thermal fluctuations to simulate one of the many factors in the oral
environment has been common in many tracer penetration, marginal gap and bond strength laboratory tests. Temperature changes used
have rarely been substantiated with temperature measurements made in vivo and vary considerably between reports. Justification and
standardization of regimen are required.
Data, sources and study selection: An assessment of reports describing temperature changes of teeth in vivo is followed by an analysis of
130 studies of laboratory thermal cycling of teeth by 99 first authors selected from 25 journals. A clinically relevant thermal cycling regimen
was derived from the in vivo information, and is suggested as a benchmark standard.
Conclusions: Variation of regimens used was large, making comparison of reports difficult. Reports of testing the effects of thermal
cycling were often contradictory, but generally leakage increased with thermal stress, although it has never been demonstrated that cyclic
testing is relevant to clinical failures. However, should this be done, the standard cyclic regimen defined is: 35⬚C (28 s), 15⬚C (2 s), 35⬚C
(28 s), 45⬚C (2 s). No evidence of the number of cycles likely to be experienced in vivo was found and this requires investigation, but a
provisional estimate of approximately 10 000 cycles per year is suggested. Thermal stressing of restoration interfaces is only of value when
the initial bond is already known to be reliable. This is not the case for most current restorative materials. 䉷 1999 Elsevier Science Ltd. All
rights reserved.

Keywords: Thermal cycling; Microleakage; Standardization; Dental restorations

1. Introduction Laboratory simulations of clinical service are often

Restorative materials are routinely used to obturate dental
cavities but later pain, marginal staining and caries often
occur. These conditions may be associated with an
inadequate cavity seal [1–17], which is thought to be
exacerbated by the effects of thermal changes [18]. Intraoral
temperature changes may be induced by routine eating [19],
drinking [18,20–24] and breathing [25]. Thermal stresses
can be pathogenic in two ways. Firstly, mechanical stresses
induced by differential thermal changes can directly induce
crack propagation through bonded interfaces [18,26,27].
Secondly, the changing gap dimensions are associated
with gap volume changes which pump pathogenic oral
fluids in and out of the gaps. This cyclical flow has been
incorrectly termed ‘percolation’ [18].
* Corresponding author. Address for correspondence: Endodontic Unit,
School of Dental Science, Faculty of Medicine, Dentistry and Health
Sciences, University of Melbourne, 711 Elizabeth Street, Melbourne,
Victoria 3000, Australia.
0300-5712/99/$ - see front matter 䉷 1999 Elsevier Science Ltd. All rights reserved.
PII S0 30 0 -5 7 12 ( 98 ) 00 0 37 - 2
performed because clinical trials are costly and time-
consuming. Thermal cycling is an in vivo process often
represented in these simulations, but the regimens used
vary considerably and, with few exceptions [28–33], are
always proposed without reference to in vivo observations.
Standardization of conditions is necessary to allow com-
parison of reports. The aim now is to analyse the limited
number of reported in vivo observations, summarize pre-
vious thermal cycling regimens, and recommend a single
substantiable regimen to enable comparability between the
results of future tests.
2. Temperatures recorded in vivo
Some experimental work has attempted to measure the
routine limits of temperature change induced by eating and
drinking. It is difficult to be precise about such events, as
eating and drinking are very erratic habits and large varia-
tions are expected between occasions, subjects [22] and
90 M.S. Gale, B.W. Darvell / Journal of Dentistry 27 (1999) 89–99
locations in the mouth [22,24]. Air temperature, humidity
and air velocity when breathing can radically alter even
resting mouth temperature [25]. However, with no thermal
load and no mouth breathing, intraoral temperature has been
measured at 35.2( ⫾ 2.1)⬚C [22], and as ‘approximately’
35⬚C [34]. With respect to the marginal gap, temperatures
from the tooth surface down to the cavity floor are of
interest.
2.1. Temperatures measured at the tooth surface
The temperature changes at the outer surface of teeth
have been measured with thermocouples [21–24] or
thermistors [20] during the drinking of hot and cold liquids.
Several confounding factors have been identified. Firstly,
large variations may occur across the mouth [24], but most
in vivo investigations have used single sensors in specific
locations (Table 1), making comparison difficult when sites
differ. Palmer et al. [23] reported temperatures at two
different locations, and demonstrated a mean difference
between the peak temperatures of the palatal surfaces of
incisors and molars of 4.5⬚C. Spierings et al. [22] used
two separate thermocouples on the upper first molar and
first premolar, and even when so close together there were
notable differences. Also, tolerance to extreme temperature
is likely to vary with location, amount of gum recession and
the presence and type of restoration. Secondly, the volume
of fluid taken into the mouth has a large effect on the tem-
perature change, as well as its duration [24]. Spierings et al.
[22] used a fixed volume of 30 mL of liquid, but even within
the same person in the same location with the same liquid
type, great variation was noted, with differences between
occasions of as much as 23⬚C.
Table 1 gives examples of temperatures of imbibed
liquids and the resultant mean minimum and maximum
tooth surface temperatures. Other reports [18,34,36]
describe application of thermal loads to teeth in vivo, but
there was no aim to identify the tolerable temperature range
allowed by the subjects. The most extreme tolerable tem-
peratures experienced are dependent upon time of exposure,
and in the tests in Table 1, except for cold application by
Palmer et al. [23], it was therefore up to the subject to
control exposure of the liquid to the teeth via their normal
Table 1
Temperatures recorded at the tooth surface
Location Volume drunk Hot liquid
(mL) temp. (⬚C)
Incisor labial — 60
Incisor palatal — ⬍ 61
Molar occlusal — ⬍ 61
— 63.5
— 58
— 55
Molar palatal — 60
30 60
eating habits. The lowest possible temperature likely at the
tooth surface is probably 0⬚C as it is unusual to eat or drink
anything colder than melting ice, and eating ice is not
uncommon [23]. Although Palmer et al. recorded temper-
atures around 1⬚C, subjects were instructed to hold the ice
cube between their teeth for 2 min. A minimum temperature
of 0⬚C was recommended, but it was not reported whether
this was comfortable or not. Therefore, the lowest com-
fortable temperature may be higher than this, as reported
by Peterson et al. [21] who noted that 10⬚C was tolerable but
uncomfortable, but 15⬚C was without discomfort. However,
one proviso to this different finding was that the labial
incisor site he used may be more thermally sensitive than
the occlusal molar site used by Palmer et al.
The highest comfortable temperature is also subjective
and, because of the risk of injury, some studies applying
heat to teeth have allowed subjects to define their own
upper temperature limits [20,23]. Others, surprisingly,
gave no indication of subjects being able to refuse the pre-
determined hot liquid temperatures (e.g., 55⬚C [34], 60⬚C
[18,21,22,35]) and one test was performed under local
anaesthesia (0⬚C, 54⬚C, 10 s) [36]. Clearly, to identify the
highest tolerable temperature, the subject must be given
control of the imbibed liquid temperature, even though
this tolerance varies considerably (42.0–57.3⬚C, molar
palatal region, 12 subjects [23]). Plant et al. [20] determined
that coffee was too hot to sip above 68⬚C in the cup, but
subjects could sip it with discomfort between 60⬚C and
68⬚C; could drink it, although it was considered relatively
hot, between 55⬚C and 60⬚C; and could drink it freely in
large amounts between 50⬚C and 55⬚C. This cup temper-
ature of 55⬚C produced the maximum comfortable tooth
surface temperature of 47⬚C. However, even coffee initially
at 47⬚C held at the tooth surface for more than 15 s produced
discomfort and pain [20], and probably temperature changes
of up to only several seconds occur routinely [22]. Never-
theless, the manner of drinking was dependent on the cup
temperature so that mouth temperature was always limited
by sipping and simultaneous intake of air [20], or by
protecting teeth with soft tissues [37]. Tooth surface
temperature during a hot food meal has also been
reported by Crabtree and Atkinson [19] as between 43⬚C
and 53⬚C.
Max. Cold liquid Min. tooth Reference
tooth temp. (⬚C) temp. (⬚C) temp. (⬚C)
45 0 15 Peterson et al., 1966 [21]
58.5 — — Palmer et al., 1992 [23]
53.1 0 1.0 Palmer et al., 1992 [23]
53.5 — — Plant et al., 1974 [20]
50 — — Plant et al., 1974 [20]
47 — — Plant et al., 1974 [20]
48.5 — — Spierings et al., 1986 [35]
44.86 0 21.63 Spierings et al., 1987 [22]
 M.S. Gale, B.W. Darvell / Journal of Dentistry 27 (1999) 89–99
2.2. Temperatures under restorations
Detailed analyses of the rates at which heat crosses the
food/tooth interface, and then the rate at which it penetrates
tooth and filling materials, have been made [22,25,34–
36,38–46]. However, in the context of the present work
such detail is probably unnecessary given that no minimum
pathogenic gap dimension is known [47–52] and damage to
the seal even only at the surface must already be considered
to be a failure.
2.3. Number of thermal cycles per unit time in vivo
No reports have been found of the number of thermal
cycles per unit time in vivo, and this requires further investi-
gation. Brown et al. [27] postulated, without citing
evidence, that 10 cycles per day occurred, but it was not
clear whether this applied to cows or humans and it may be
an underestimate for humans with a prepared diet. Kim et al.
[53] gave no explanation for cycling in vitro specimens for
three short periods each day, with each period containing 10
cycles, but presumably this thrice daily cycling was
intended to simulate three meals per day. Lloyd et al. [37]
noted a similarity in enamel crack lengths in teeth in vivo
after several years’ service and newly erupted extracted
teeth after several thousand thermal cycles in vitro. He
therefore suggested that several thousand thermal cycles
might occur in vivo in several years, but this conclusion
must be confounded by mechanical stresses both during
service and extraction.
3. Temperature regimens previously used for in vitro
tests
Thermal cycling is common in tracer penetration, shear
bond strength and tensile bond strength tests of dental
materials. Some 130 thermal cycling experimental reports
from 25 journals with 99 first authors made up the sample
reviewed here (Table 2). Of these, 110 involved a tracer
penetration test, 26 a shear bond strength test, nine a tensile
bond strength test and one recorded enamel crack length.
The mean low-temperature point was 6.6⬚C (range 0–36⬚C,
median 5.0⬚C). The mean high-temperature point was
55.5⬚C (range 40–100⬚C, median 55⬚C). The majority of
reports quoted used just hot and cold temperature points,
but 27 studies also used an intermediate temperature, all
but one [29] at 37⬚C. The number of cycles used varied
from 1 (e.g., [57]) to 1 000 000 (e.g., [85]) cycles, with a
mean of about 10 000 and median of 500 cycles. Dwell
times were sometimes not stated [5,7,10,65,74,
79,87,97,104,114,141,149,154,157], but the mean stated
dwell time was 53 s, the median 30 s, with a range of 4 s
[69] to 20 min [132]. Sometimes a longer dwell time (e.g.,
23 s [3]) was used with an intermediate temperature of
37⬚C, and a shorter dwell time (e.g., 4 s [3]) for the
91
temperature extremes, presumably in an attempt to mimic
expected intraoral timings. In nearly [28–33] all reports, no
reasoning was given for the choice of temperature and
timing conditions [171]. The most rigorous thermal
treatment seen (not included in the above analysis)
involved boiling specimens for 8 h before testing [55].
This appears to be an unreasonable requirement in normal
service.
Even without the intention to thermal cycle, temperatures
for interface tests are often unstated. Some exceptions
include work involving air tracers [162–164] or marginal
gap [57,121,165,166] measurements for which isothermal
cabinets [162–164] or rooms [57,121,165,166] were used to
maintain 37⬚C.
4. Investigation of effects of thermal cycling regimens
Various aspects of a thermal cycling regimen have been
tested experimentally. The dependent variable was often
some measure of tracer penetration. Comparison of cycled
and uncycled specimens [56,62,69,77,83,86,109,121,
128,167], temperature range [168], number of temperatures
in the cycle [57,69], number of cycles [3,9,49,53,99,
118,168], dwell times [57,69] and whether the cycles were
in tracer [9] have all been made. Gage and Clarke [87] used
an artificial saliva during thermal cycling. It is disappointing
that all but a few reports using thermal cycling as a test
variable lack any apparent justification from in vivo mea-
surements for the temperatures, number of cycles or dwell
times used in the experiments, e.g., [3,9,49,69,77,83,86,109,
128,140,169]. It is therefore difficult to understand the
choices of regimen. Justification is still more rare in reports
which use but do not test the effect of thermal cycling. In all,
only three reports [29,31,32] have been found which refer to
previous in vivo measurements.
The results of these tracer tests are also inconsistent.
Some have reported gap widening from just one thermal
cycle [26,168,170], whilst others have shown that seal
integrity is reduced by increasing the number of thermal
cycles [49,53,99,118]. Still other papers have shown no
significant increase in leakage with the use of thermal
cycling [56,63,69,77,83] or increasing number of cycles
[3,128], whilst Kidd et al. [169] reported a decrease in leak-
age with cycling. Asmussen [57] found little difference
between 2 s and 5 s dwell times except after 1000 cycles,
and Mandras et al. [9] reported no significant difference
between the test specimens being in and out of tracer during
cycling. Krejci and Lutz [109] found that thermal cycles had
little effect on bonds to enamel, but caused the deterioration
of some dentine margins. Both Mandras et al. [9] and
Rigsby et al. [140] showed that thermal cycling with, but
not without, load cycling increased tracer penetration.
Munksgaard and Irie [121] concluded that thermal cycling
was detrimental after early polishing (at 10 min), but after
polishing at 24 h there was no effect observed. This may
92 M.S. Gale, B.W. Darvell / Journal of Dentistry 27 (1999) 89–99

Table 2
Thermal cycling regimens: experimental conditions used in tracer and bond strength tests. Some reports describe the use of two periods of thermal cycling or
two different dwell times

First author Year Reference Low (⬚C) High (⬚C) Int. a (⬚C) No. of cycles Dwell time (s)

1st 2nd 1st 2nd

Abdalla 1993 [54] 5 55 500 180

Amsberry 1984 [55] 15 56 150 45
Arcoria 1991 [56] 5 50 540 30
Asmussen 1974 [57] 2 60 1 180
Atta 1990 [58] 5 60 37 500 15
Ben Amar 1987 [1] 4 55 200 30
Ben Amar 1993 [59] 5 55 100 10
Brown 1972 [27] 36 60 7000 30
Buonocore 1973 [60] 0 100 1 (boiled for 8 h)
Chan 1994 [61] 6 60 200 30
Chan 1994 [62] 15 45 37 150 60 240
Chan 1992 [63] 15 45 37 150 60 240
Chan 1993 [31] 4 45 37 150 15 60
Charlton 1992 [64] 5 45 2500 30
Chohayeb 1992 [65] 5 55 3780 No data
Chow 1980 [66] 4 60 25 60
Cooley 1991 [67] 6 60 800 60
Cooley 1991 [68] 6 60 800 30
Crim 1981 [28] 5 60 37 1500 4 23
Crim (method 1) 1985 [69] 12 60 37 1500 4 23
Crim (method 2) 1985 [69] 12 60 1500 30
Crim 1987 [3] 12 54 37 100 4 23
Crim 1990 [4] 12 54 37 100 4 23
Crim 1991 [70] 12 54 37 700 4 23
Crim 1992 [71] 12 54 37 100 4 23
Crim 1993 [72] 12 54 37 100 4 23
Crim 1993 [32] 12 54 37 100 4 23
Crim 1994 [73] 12 54 37 100 4 23
Davidson 1993 [74] 5 55 300 No data
Davis 1989 [75] 5 54 37 3600 15
DeWet 1980 [76] 15 60 5745 30
Derand 1990 [5] 10 45 80 No data
Derkson 1986 [77] 4 56 100 120
Douglas 1989 [78] 5 55 500 30
Dumsha 1984 [79] 5 55 3360 No data
Dutton 1993 [80] 6 60 500 30
Eakle 1992 [81] 5 55 240 30
Eliades 1985 [82] 4 60 1300 15
Eriksen 1976 [83] 0 60 30 60
Finger 1988 [84] 5 55 250 15
Fukuda 1992 [85] 2 72 1 000 000 15
Fusayama 1989 [86] 4 60 37 100 60 120
Gage 1991 [87] 5 55 No data No data
Garcia-Godoy 1993 [88] 5 55 37 4000 30
Garcia-Godoy 1987 [89] 6 60 100 30
Gilpatrick 1994 [90] 4 58 100 60
Gordon 1991 [7] 5 55 100 No data
Grieve 1980 [91] 15 45 37 150 60 240
Hadavi 1993 [92] 4 60 180 30
Haller 1993 [93] 5 55 1440 30
Hammesfahr 1987 [94] 10 48 540 60
Harashima 1992 [49] 4 60 600 1500 15
Hasegawa 1992 [95] 8 50 500 15 (in dye)
Hasegawa 1993 [96] 8 50 500 15 (in dye)
Hayakawa 1989 [97] 5 55 1440 No data
Hirschfeld 1992 [98] 4 60 200 60
Iwase 1989 [99] 4 60 37 28 800 60 120
Jamil 1992 [100] 5 55 250 15
Judes 1982 [101] 4 60 20 120
 M.S. Gale, B.W. Darvell / Journal of Dentistry 27 (1999) 89–99 93

Table 2 (continued )

First author Year Reference Low (⬚C) High (⬚C) Int. a (⬚C) No. of cycles Dwell time (s)

1st 2nd 1st 2nd

Kamel 1990 [102] 8 50 500 15

Kanca 1989 [8] 10 50 500 30
Kanca 1989 [103] 5 55 500 30
Kanca 1990 [104] 5 55 500 No data
Kanca 1991 [105] 5 55 500 30
Kaplan 1992 [106] 4 58 100 60
Katoh 1993 [107] 6 46 37 2500 30
Kim 1992 [53] 4 60 300 180
Kingsford-Smith 1992 [108] 8 48 300 25
Krejci 1990 [109] 5 55 1250 120
Krejci 1993 [110] 5 55 1250 120
Lacy 1992 [111] 5 55 120 60
Lee 1970 [112] 4 50 100 222 10
Litkowski 1991 [113] 5 55 5000 9
Lloyd 1978 [37] 24 52 5000 30
Lutz 1986 [114] 5 55 2500 No data
Mandras 1991 [9] 8 55 250 1000 15 (in dye)
Martin 1993 [115] 5 55 100 60
McCaghren 1990 [116] 8 50 250 15
McInnes 1990 [10] 5 60 500 No data
Mixson 1992 [117] 5 55 100 30
Momoi 1990 [118] 4 60 37 9000 60 120
Morley 1977 [119] 4 60 No data 240
Mount 1992 [120] 5 55 500 30
Munksgaard 1987 [121] 15 50 500 15
Nakabayashi 1982 [122] 4 60 60 60
Nara 1992 [123] 4 60 125 30
Nolden 1985 [124] 4 65 60 60
O’Brien 1988 [125] 5 55 250 30
Pearson 1987 [29] 5 65 32 13 056 4 30 (in
dye)
Perdiago 1993 [126] 5 55 500 30
Prati 1989 [127] 4 55 37 250 10 30
Prati 1994 [128] 4 55 240 1689 30
Prevost 1982 [129] 15 45 30 100 No data
Rebitski 1993 [130] 10 50 400 30
Rees 1990 [131] 5 60 830 30
Reid 1991 [132] 5 55 37 6 1200
Reilly 1992 [133] 4 54 37 37 500 15
Retief 1988 [134] 8 55 250 500 15
Retief 1990 [135] 8 50 500 15
Retief 1992 [136] 8 50 500 15
Retief 1992 [137] 8 50 250 15
Retief 1993 [138] 8 50 500 15
Rigsby 1990 [139] 8 50 500 15
Rigsby 1992 [140] 8 50 500 15
Roulet 1989 [141] 5 55 2500 No data
Rossomando 1995 [33] 5 55 5000 1060 (in dye)
Roydhouse 1967 [15] 4 65 50 100 60
Rupp 1986 [142] 5 55 3780 80
Saunders 1990 [143] 5 55 2110 10
Scott 1992 [144] 5 55 37 500 10
Sidhu 1991 [145] 5 55 250 15
Sidhu 1992 [146] 5 55 500 30
Souza 1993 [147] 8 50 500 15 (in dye)
Srisawasdi 1998 [148] 5 55 500 30
Staninec 1988 [149] 4 60 100 No data
Staninec 1993 [150] 4 60 1000 60
Swift 1991 [151] 5 55 300 30
Swift 1992 [152] 5 55 300 30
Swift 1993 [153] 5 55 300 10 30
(continued on next page)
94 M.S. Gale, B.W. Darvell / Journal of Dentistry 27 (1999) 89–99

Table 2 (continued )

First author Year Reference Low (⬚C) High (⬚C) Int. a (⬚C) No. of cycles Dwell time (s)

1st 2nd 1st 2nd

Thordrup 1994 [154] 12 62 2500 No data

Tjan 1991 [13] 5 55 300 60
Torstenson 1988 [26] ⬍0 80 1 45 75
Wendt 1990 [155] 8 50 250 15
Wendt 1992 [30] 5 50 250 15 (in dye)
White 1992 [156] 5 50 1500 30
Williams 1983 [157] 4 40 500 1000 No data
Wu 1983 [158] 5 55 3780 80
Yoshida 1990 [159] 4 60 20 000 60
Yu 1991 [160] 4 54 37 50 000 15
Yu 1991 [161] 4 54 37 2000 15
a
Intermediate temperature.

have been due to a confounding water sorption effect. A
lack of negative controls (i.e., no cycling at all) in some
of these experiments [3,9,29,49,53,99] has prevented
definitive conclusions. Importantly, although the non-
destructive test methods [49,53,99,118] tend to show
increasing leakage with number of cycles (except [128]),
the increase may be related to other time-dependent effects
rather than cycling (see below). Momoi et al. [118] did have
uncycled controls, and these showed some increase in tracer
penetration with time, but not as great as with cycled teeth.
Such conflicting results are hardly surprising given that
no two authors used exactly the same protocol, and even the
same author varied fundamental test conditions such as the
extreme temperatures as well as the main test variable in
different papers [3,69]. Another possible reason for
inconsistent results is that perhaps tracer penetration may
not be directly linked to gap dimensions in the presence of
other factors, for example water sorption expansion of
materials.
5. Recommendations for thermal cycling simulation for
in vitro interface testing
Past thermal cycling regimens are almost all unreferenced
to in vivo observations [171], despite the general expecta-
tion that methods are substantiated in publications. In
addition, despite great variations in temperature fluctuations
and their tolerance in vivo, a standard thermal cycling
simulation is required to allow comparison of materials
and procedures between reports. The essential element is
ordinarily assumed to be the generation of mechanical
stresses and fluid flows. The thermal gradient and rate of
change of temperature also control the mechanical stresses
through the effects of differential thermal conductivity. But
the thermal masses involved are small, so that thermal
changes are moderately rapid [20,24]. Therefore, long
resting periods appear to be unnecessary [28,30,33],
permitting accelerated simulation.
Excessive stresses will have poor discriminatory power
and would unreasonably condemn materials which may
serve perfectly well in practice. Setting a test criterion
high when the inadequacy of a lower value has not been
demonstrated is pointless, and does not meet the purpose of
a minimum acceptable performance standard, which is pre-
sumably one goal of using the present type of test to predict
expected clinical service.
On the basis of the literature reviewed, no definitive state-
ment of a relevant regimen can be made and, in any case, the
choice must be arbitrary to a greater or lesser extent because
of wide variation. However, it is concluded from the
available data that while extreme temperatures could con-
ceivably be encountered, they cannot reasonably be taken as
typical or representative. We therefore opt for temperatures
of 15⬚C and 45⬚C. Likewise, on the available evidence, we
opt for 35⬚C as the reference resting temperature, although
in absolute terms this is unlikely to be critical. Dwell times
are also arbitrary, but equilibrium at either extreme is not
expected; a thermal gradient over the depth of the restora-
tion is probably more relevant. In addition, sufficient time
must be allowed for a return to the reference resting
temperature after an extreme if an effective transition
from one to the other extreme is to be avoided, a possible
but uncommon event in life.
Thus, in the interests of test standardization, and with the
recognition that a specific regimen does not represent the
natural variability in vivo, the sequence of temperatures:
35⬚C, 15⬚C, 35⬚C and then 45⬚C, with a corresponding
dwell sequence of 28 s, 2 s, 28 s and 2 s, is suggested to
be sufficiently clinically relevant and is therefore
recommended here as a suitable discriminatory challenge.
The frequency of cycles in vivo remains undetermined at
present and requires formal estimation. In the absence of
this information, it is proposed that on the basis that such
cycles might occur between 20 and 50 times in a day, some
10 000 cycles might represent a service year. For the
sequence described above, it would take about 3.5 days to
complete such a test cycle and this may conveniently be
done automatically [9,21,37,49,118,119,169]. Again, this
is arbitrary but modest.
 M.S. Gale, B.W. Darvell / Journal of Dentistry 27 (1999) 89–99
6. The usefulness of thermal cycling
It is to be noted that there is no concrete evidence that
failures in practice occur because of thermal stresses, not-
withstanding the theoretical expectation. However, the
distinction must be made between the equivalent static
stress test (i.e., increase steadily until collapse occurs) and
fatigue failure, where repeated loading to a stress below the
static strength occurs. A less severe test would in fact
improve discrimination of this point, so long as stresses
were below those which would cause immediate collapse.
Equally, it has never been determined whether it is the time
at temperature (i.e., cumulative duration under stress), as
opposed to true fatigue, that is the relevant factor. In other
words, whether failure occurs due to flow in one or other of
the layers in the bonded structure is unknown. This must be
dependent on the glass transition temperatures of those
bonded components (which do not appear to have been
determined). Such flow would lead to collapse in a truly
static test at a stress below the ordinary static strength.
Cycling tests cannot of themselves discriminate these
issues, and failure to allow for them might account in part
for the inconsistency of test results so far reported. The test
outcome clearly depends on the stress generated and the
failure mechanism.
Certainly, if fatigue as such is not the mechanism, a cyclic
test is irrelevant and wasted effort. In the light of this, and
simply put, all previous results are uninterpretable. We
conclude that cycling has been adopted as a superficial
‘simulation’ of supposed surface conditions, without asking
the right question [172]. Indeed, as far as can be seen, only
one author [57,168] has made experimental observations in
any detail to determine the mode of failure. However, his
findings may not apply to the materials currently available.
A further point that has previously been overlooked by all
but a few [30,173] in the context is that bond breakdown
may be enabled or facilitated by hydrolysis of the bonding
agent and its union with tooth tissue. This type of reaction
may also be stress-aided, as occurs in other systems. This is
a potential mechanism for fatigue failure, but equally time at
temperature may be the relevant condition. Accordingly,
distilled water is considered to be an entirely inadequate,
if benign, medium for such tests. It is therefore proposed
that, as a minimum challenge, the immersion medium be an
artificial saliva (cf. [87]), specifically that of Leung and
Darvell [174], adjusted to pH 6, equilibrated with 3.5%
CO 2 in air, saturated with respect to hydroxyapatite.
It is obvious that there are many other points relating to
the test design that have a bearing on the relevance and
standardization of a test [175]. However, these are beyond
the present scope and will have to be considered elsewhere.
Nevertheless, it is necessary to point out that the principles
adduced here need to be taken into account. The decision to
use thermal cycling for restoration interface tests requires
critical thought, even assuming that the points made above
are taken into account. Relevance of the degree of seal to
95
pathology is not well understood [49] and, in the absence of
clinically predictive gap criteria [47–53], only the presence
or absence of a seal is a sure measure of gap pathogenicity.
Only when materials consistently create an initially intact
seal, and the intention of the laboratory test is to predict
clinical service, then representative simulation of thermal
stresses might improve test relevance. In this situation, such
as when testing the reportedly effective [7,9,10,12,
67,94,108,109] bond of resins to acid-etched enamel, the
cumulative duration of stress or the number of cycles to
the beginning of leakage has been proposed as a possibly
clinically relevant criterion [49].
However, if materials (e.g., dentine-bonded filled resin
[7,103,109,152,176,177]) routinely do not achieve this
initial complete seal in the absence of applied stress, there
is no point in simulating in vivo thermal cycles to further
damage an already defective interface. Any degree of failure
in seal is already a failure of the interface, and the extra
effort required is not worthwhile or necessary. This point
will remain pertinent until the clinical relevance of leakage
and gap tests has been convincingly determined. Thus, we
come to the conclusion that thermal cycling is an
irrelevancy with spurious legitimacy, or is at best premature,
because the validating preconditions have yet to be met.
Even so, if the intention of the test is not to determine
material serviceability but to investigate the mode of failure,
then thermal cycling might be appropriate, even with
materials which consistently fail.
Acknowledgements
Financial assistance in the form of a postgraduate student-
ship for Martin Gale from The University of Hong Kong is
gratefully acknowledged. This work was done in partial
fulfilment of the requirements for the degree of Ph.D. for
the first author.
References
[1] Ben Amar A, Nordenberg D, Liberman R. et al. The control of
marginal microleakage in amalgam restorations using dentin
adhesive: a pilot study. Dent Mater, 1987;3:94–96.
[2] Christen AG, Mitchell DF. A fluorescent dye method for demonstrat-
ing leakage around dental restorations. J Dent Res, 1966;45:1485–
1492.
[3] Crim GA, Garcia-Godoy F. Microleakage: the effect of storage and
cycling duration. J Prosthet Dent, 1987;57:574–576.
[4] Crim GA. Prepolymerization of Gluma 4 sealer: effect on bonding.
Amer J Dent, 1990;3:25–27.
[5] Derand T. A lining system for composite resin fillings. J Prosthet
Dent, 1990;63:134–137.
[6] Duke ES, Lindemuth J. Polymeric adhesion to dentin: contrasting
substrates. Amer Dent J, 1990;3:264–270.
[7] Gordon M, Wasserstein A, Gorfil C, Imber S. Microleakage in three
designs of glass ionomer under composite resin restorations. J Oral
Rehab, 1991;18:149–154.
96 M.S. Gale, B.W. Darvell / Journal of Dentistry 27 (1999) 89–99
[8] Kanca J. The effect on microleakage of four dentin–enamel bonding
systems. Quintessence Int, 1989;20:359–361.
[9] Mandras RS, Retief DH, Russell CM. The effects of thermal and
occlusal stresses on the microleakage of Scotchbond 2 dentinal
bonding system. Dent Mater, 1991;7:63–67.
[10] McInnes P, Perkins E, Weinberg R. Microleakage of glass ionomer/
composite laminate class V restorations. Amer J Dent, 1990;3:21–
24.
[11] Pashley DH. Clinical correlations of dentin structure and function.
J Prosthet Dent, 1991;66:777–781.
[12] Setcos JC. Dentine bonding in perspective. Amer J Dent,
1988;1:173–175.
[13] Tjan AHL, Tan DE. Microleakage at gingival margins of class V
composite resin restorations rebonded with various low-viscosity
resin systems. Quintessence Int, 1991;22:565–573.
[14] Wilson NHF. The evaluation of materials: relationship between
laboratory investigations and clinical studies. Oper Dent,
1990;15:149–155.
[15] Roydhouse RH, Weiss ME. Penetration around the margins of
restorations: 1. Review and experiments. J Can Dent Assoc,
1967;33:680–689.
[16] Kidd EAM, O’Hara JW. The caries status of occlusal amalgam
restorations with marginal defects. J Dent Res, 1990;69:1275–1277.
[17] Shortall AC. In vitro microleakage of class V composite resin
restorations with and without light-cured glass–ionomer (poly-
alkenoate) cement lining. Eur J Pros Rest Dent, 1992;1:7–11.
[18] Nelsen RJ, Wolcott RB, Paffenbarger GC. Fluid exchanges at the
margins of dental restorations. J Amer Dent Assoc, 1952;44:288–
295.
[19] Crabtree MG, Atkinson HF. A preliminary report on the solubility of
decalcified dentine in water. Austral J Dentistry, 1955;55:340–342.
[20] Plant CG, Jones DW, Darvell BW. The heat evolved and temper-
atures attained during setting of restorative materials. Brit Dent J,
1974;137:233–238.
[21] Peterson EA, Phillips RW, Swartz ML. A comparison of the physical
properties of four restorative resins. J Amer Dent Assoc,
1966;73:1324–1336.
[22] Spierings TAM, Peters MCRB, Bosman F, Plasschaert AJM.
Verification of theoretical modelling of heat transmission in teeth
by in vivo experiments. J Dent Res, 1987;66:1336–1339.
[23] Palmer DS, Barco MT, Billy EJ. Temperature extremes produced
orally by hot and cold liquids. J Prosthet Dent, 1992;67:325–327.
[24] Longman CM, Pearson CJ. Variation in temperature of the oral
cavity during the imbibition of hot and cold fluids [special issue].
J Dent Res, 1984;63:521. (Abstract 283).
[25] Boehm RF. Thermal environment of teeth during open mouth
respiration. J Dent Res, 1972;51:75–78.
[26] Torstenson B, Brannstrom M. Contraction gap under composite resin
restorations: effect of hygroscopic expansion and thermal stress.
Oper Dent, 1988;13:24–31.
[27] Brown WS, Jacobs HR, Thompson RE. Thermal fatigue in teeth.
J Dent Res, 1972;51:461–467.
[28] Crim GA, Mattingly SL. Evaluation of two methods for assessing
marginal leakage. J Prosthet Dent, 1981;45:160–163.
[29] Pearson GJ, Longman CM. The effect on marginal leakage, in vitro,
of curing a composite material at elevated temperatures with or
without marginal etching of the cavity. J Dent, 1987;15:171–174.
[30] Wendt SL, McInnes PM, Dickinson GL. The effect of thermal
cycling in microleakage analysis. Dent Mater, 1992;8:181–184.
[31] Chan MFWY, Glyn-Jones JC. Marginal sealing ability of four
restorative materials placed in root surfaces. Eur J Pros Rest Dent,
1993;2:23–28.
[32] Crim GA, Schmidt ED. Microleakage resistance of glutaraldehyde-
containing adhesives. Amer J Dent, 1993;6:142–144.
[33] Rossomando KJ, Wendt SL. Thermocycling and dwell times in
microleakage evaluation for bonded restorations. Dent Mater,
1995;11:47–51.
[34] Tibbets VR, Schnell RJ, Swartz ML, Phillips RW. Thermal diffusion
through amalgam and cement bases: comparison of in vitro and in
vivo measurements. J Dent Res, 1976;55:441–451.
[35] Spierings TAM, Van Der Varst PGT, Peters MCRB. Thermal load-
ing conditions in the oral cavity [special issue]. J Dent Res,
1986;65:777. (Abstract 463).
[36] Harper RH, Schnell RJ, Swartz ML, Phillips RW. In vivo measure-
ments of thermal diffusion through restorations of various materials.
J Prosthet Dent, 1980;43:180–185.
[37] Lloyd BA, McGinley MB, Brown WS. Thermal stress in teeth. J Dent
Res, 1978;57:571–582.
[38] McCabe JF. Resin based filling materials. In: Applied dental
materials. 7th edn. Oxford: Blackwell Scientific Publications,
1990:45–156.
[39] Tani Y. Current status of posterior composites resins in Japan. In:
Vanherle G, Smith DC, editors. Posterior composite resin dental
restorative materials. USA: Minnesota Mining and Mfg Co.,
1985:185–97.
[40] Jacobs HR, Thompson RE, Brown WS. Heat transfer in teeth. J Dent
Res, 1973;52:249–252.
[41] Spierings TAM, Bosman F, Peters MCRB, Plasschaert AJM. Deter-
mination of the convective heat transfer coefficient. Dent Mater,
1987;3:161–164.
[42] Craig RA, Peyton FA. Thermal conductivity of tooth structures,
dental cements and amalgam. J Dent Res, 1961;40:411–418.
[43] Drummond JL, Robledo J, Garcia L, Toepke TRS. Thermal con-
ductivity of cement base materials. Dent Mater, 1993;9:68–71.
[44] De Vree JHP, Spierings TAM, Plasschaert AJM. A simulation model
for transient thermal analysis of restored teeth. J Dent Res,
1983;62:756–759.
[45] Spierings TAM, De Vree JHP, Peters MCRB, Plasschaert AJM. The
influence of restorative dental materials on heat transmission in
human teeth. J Dent Res, 1984;63:1096–1100.
[46] Spierings TAM, Peters MCRB, Bosman F, Plasschaert AJM. The
influence of cavity geometry on heat transmission in restored teeth.
J Dent, 1986;14:47–51.
[47] Roulet, JF. Marginal integrity: clinical significance. J Dent 1994;22
Suppl 1:S9–S12.
[48] Wu MK, Wesselink PR. Endodontic leakage studies reconsidered.
Part I. Methodology, application and relevance. Int Endod J,
1993;26:37–43.
[49] Harashima I, Uzawa T, Hirasawa T. A new method for the assess-
ment of marginal sealability of dental restorations. Dent Mater,
1992;11:150–157.
[50] Dietschi D, Maeder M, Holz J. In vitro evaluation of marginal fit and
morphology of fired ceramic inlays. Quintessence Int, 1992;23:271–
278.
[51] Qualtrough AJE, Piddock V. Fitting of indirect restorations: a review
of methods of assessment. Eur J Pros Rest Dent, 1992;1:57–61.
[52] Mattison GD, Delivanis PD, Thacker RW, Hassell JK. Effect of
post preparation on the apical seal. J Prosthet Dent, 1984;51:785–
789.
[53] Kim JY, Takahashi Y, Kito M, Morimoto Y. Semi-quantitative
analysis of early microleakage around amalgam restorations by
fluorescent spectrum method: a laboratory study. Jpn Dent Mater J,
1992;11:45–58.
[54] Abdalla AI, Davidson CL. Comparison of the marginal integrity of in
vitro class II composite restorations. J Dent, 1993;21:158–162.
[55] Amsberry W, Von Fraunhofer JA, Hoots J, Rodgers H. Marginal
leakage of several acid-etch composite resin restorative systems.
J Prosthet Dent, 1984;52:647–653.
[56] Arcoria CJ, Fisher MA, Wagner MJ. Microleakage in alloy–glass
ionomer lined amalgam restorations after thermocycling. J Oral
Rehab, 1991;18:9–14.
[57] Asmussen E. The effect of temperature changes on adaption of resin
fillings. Acta Odontol Scand, 1974;32:161–171.
[58] Atta OM, Smith BGN, Brown D. Bond strengths of three chemical
 M.S. Gale, B.W. Darvell / Journal of Dentistry 27 (1999) 89–99
adhesive cements adhered to a nickel–chromium alloy for direct
bonded retainers. J Prosthet Dent, 1990;63:137–143.
[59] Ben Amar A, Cardash HS, Liberman R. Varnish application
technique and microleakage of amalgam restorations. Amer J
Dent, 1993;6:65–68.
[60] Buonocore MG, Sheykholeslam Z, Glena R. Evaluation of an enamel
adhesive to prevent marginal leakage: an in vitro study. J Dent Child,
1973;40:119–124.
[61] Chan KC, Swift EJ. Marginal seal of new generation dental bonding
agents. J Prosthet Dent, 1994;72:420–423.
[62] Chan MFWY, Glyn-Jones JC. Significance of thermal cycling in
microleakage analysis of root restorations. J Dent, 1994;22:292–295.
[63] Chan MFWY, Glyn-Jones JC. A comparison of four in vitro
marginal leakage tests applied to root surface restorations. J Dent,
1992;20:287–293.
[64] Charlton DG, Moore BK. In vitro evaluation of two microleakage
detection tests. J Dent, 1992;20:55–58.
[65] Chohayeb AA. Microleakage comparisons of dentin bonding
systems with glass ionomer. Dent Mater, 1992;8:27–30.
[66] Chow MH. The effects of sealants placed over composite resin
restorations. J Prosthet Dent, 1980;44:531–535.
[67] Cooley RL, Barkmeier WW. Dentinal shear bond strength, micro-
leakage and contraction gap of visible light polymerized liners/bases.
Quintessence Int, 1991;22:467–474.
[68] Cooley RL, Tseng EY, Barkmeier WW. Dentinal bond strengths and
microleakage of a 4-META adhesive to amalgam and composite
resin. Quintessence Int, 1991;22:979–983.
[69] Crim GA, Swartz ML, Phillips RW. Comparison of four thermo-
cycling techniques. J Prosthet Dent, 1985;53:50–53.
[70] Crim GA. Microleakage of three dentinal bonding systems: a 6
month evaluation. Quintessence Int, 1991;22:387–389.
[71] Crim GA. Post-operative bleaching: effect on microleakage. Amer J
Dent, 1992;5:109–112.
[72] Crim GA. Effect of aging on microleakage of restorative systems.
Amer J Dent, 1993;6:192–194.
[73] Crim GA, Chapman KW. Reducing microleakage in class II restora-
tions: an in vitro study. Quintessence Int, 1994;25:781–785.
[74] Davidson CL, Abdalla AI, De Gee AJ. An investigation into the
quality of dentine bonding systems for accomplishing a durable
bond. J Oral Rehab, 1993;20:291–300.
[75] Davis EL, Joynt RB, Weiczkowski G, Laura JC. Bond durability
between dentinal bonding agents and tooth structure. J Prosthet
Dent, 1989;62:253–256.
[76] De Wet FA, Ferreira MR. The durability of dental glazes. J Prosthet
Dent, 1980;44:300–306.
[77] Derkson GD, Pashley DH, Derkson ME. Microleakage measurement
of selected restorative materials: a new in vitro method. J Prosthet
Dent, 1986;56:435–440.
[78] Douglas WH, Fields RP, Fundingsland J. A comparison between the
microleakage of direct and indirect composite restorative systems.
J Dent, 1989;17:184–188.
[79] Dumsha T, Biron G. Inhibition of marginal leakage with a dentin
bonding agent. J Dent Res, 1984;63:1255–1257.
[80] Dutton FB, Summitt JB, Chan DC, Garcia-Godoy F. Effect of a resin
lining and rebonding on the marginal leakage of amalgam
restorations. J Dent, 1993;21:52–56.
[81] Eakle WS, Staninec M, Lacy AM. Effect of bonded amalgam on the
fracture resistance of teeth. J Prosthet Dent, 1992;68:257–260.
[82] Eliades GC, Caputo AA, Vougiouklakis GJ. Composition, wetting
properties and bond strength with dentin of 6 new dentin adhesives.
Dent Mater, 1985;1:170–176.
[83] Eriksen HM, Buonocore MG. Marginal leakage with different com-
posite restorative materials in vitro. Effect of cavity design. J Oral
Rehab, 1976;3:315–322.
[84] Finger WJ. Dentin bonding agents. Relevance of in vitro investiga-
tions [special issue]. Amer J Dent, 1988;1:184–188.
[85] Fukuda K, Nara Y, Katsuyama S. Microleakage of class II composite
97
inlay under repeated load [special issue]. J Dent Res
1992;71:Abstract 1632.
[86] Fusayama A, Kohno A. Marginal closure of composite restorations
with the gingival wall in cementum/dentin. J Prosthet Dent,
1989;61:293–296.
[87] Gage JP, Clarke DF. Comparison between two Procion dyes and
Alizarin Red to assess marginal leakage in resin restorations. A
laboratory study. Austral Dent J, 1991;36:87–93.
[88] Garcia-Godoy F, Finger WJ. Reliability of microleakage evaluation
using dentin bonding agents [special issue]. J Dent Res, 1993;72:308.
(Abstract 1636).
[89] Garcia-Godoy F, Malone WFP. Microleakage of posterior composite
restorations after rebonding. Compend Contin Edu, 1987;8:606–
610.
[90] Gilpatrick RO, Kaplan I, Roach D. Microleakage of composite resin
restorations with various etching times. Quintessence Int,
1994;25:573–576.
[91] Grieve AR, Glyn-Jones JC. An in vitro study of marginal leakage
associated with composite restorations using an acidified agar gel.
J Oral Rehab, 1980;7:215–233.
[92] Hadavi F, Hey JH, Ambrose ER, el Badawy HE. Repair of high
copper amalgam with and without an adhesive system: in vitro
assessment of microleakage and shear bond strength. Gen Dent
1993;(Feb):49–53.
[93] Haller B, Hofmann N, Klaiber B, Bloching U. Effect of storage
media on microleakage of five dentin bonding agents. Dent Mater,
1993;9:191–197.
[94] Hammesfahr PD, Huang CT, Shaffer SE. Microleakage and bond
strength of resin restorations with various bonding agents. Dent
Mater, 1987;3:194–199.
[95] Hasegawa T, Retief DH. Quantitative microleakage of some dentinal
bonding restorative systems [special issue]. J Dent Res, 1992;71:663.
(Abstract 1180).
[96] Hasegawa T, Retief DH. Quantitative microleakage of some dentinal
bonding restorative systems. Dent Mater, 1993;9:114–117.
[97] Hayakawa T, Endo H, Nagatsuka A, Horie K. Studies on adhesion to
tooth substrate. VI. The adhesion of composite resin to dentin
mediated by the dentin primer containing MYTA. Jpn Dent Mater
J, 1989;8:155–163.
[98] Hirschfeld Z, Frenkel A, Zyskind D, Fuks A. Marginal leakage of
class II glass ionomer–composite resin restorations: an in vitro
study. J Prosthet Dent, 1992;67:148–153.
[99] Iwase H, Mamoi Y, Asanuma A. et al. A marginal leakage of
composite resin gradually increased with thermal cycling [special
issue]. J Dent Res, 1989;68:923. (Abstract 455).
[100] Jamil M, Aboush YEY, Elderton RJ. Bond strengths of dentine
bonding agents to dentine. Brit Dent J, 1992;172:344–347.
[101] Judes H, Eli I, Lieberman R, Serebro L, Amar B. Rebonding as a
method of controlling marginal microleakage in composite resin
restorations. New York State Dent J, 1982;52:137–143.
[102] Kamel FM, Retief DH, Mandras RS, Russell CM. A laboratory study
of the Herculite XR system. Amer J Dent, 1990;3:271–277.
[103] Kanca J. Microleakage of five dentin bonding systems. Dent Mater,
1989;5:415–416.
[104] Kanca J. A method for bonding to tooth structure [special issue].
J Dent Res, 1990;69:231. (Abstract 984).
[105] Kanca J. A method of bonding to tooth structure using phosphoric acid
as a dentin–enamel conditioner. Quintessence Int, 1991;22:285–290.
[106] Kaplan I, Mincer HH, Harris EF, Cloyd JS. Microleakage of com-
posite resin and glass ionomer cement restorations in retentive and
non-retentive cervical cavity preparations. J Prosthet Dent,
1992;68:616–623.
[107] Katoh Y, Shinkai K. Adaption of composite fillings to cavity walls in
three different cavities. Asian J Aesthet Dent, 1993;1:3–11.
[108] Kingsford-Smith ED, Martin FE. Microleakage of glass ionomer/
composite resin restorations: a laboratory study. 1. The influence
of glass ionomer cement. Austral Dent J, 1992;37:23–30.
98 M.S. Gale, B.W. Darvell / Journal of Dentistry 27 (1999) 89–99
[109] Krejci I, Lutz F. Mixed class V restorations: the potential of a dentine
bonding agent. J Dent, 1990;18:263–270.
[110] Krejci I, Kuster M, Lutz F. Influence of dentinal fluid and stress on
marginal adaption of resin composites. J Dent Res, 1993;72:490–
494.
[111] Lacy AM, Wada C, Du W, Wantabe L. In vitro microleakage at the
gingival margin of porcelain and resin veneers. J Prosthet Dent,
1992;67:7–10.
[112] Lee HL, Swartz ML. Scanning electron microscope study of com-
posite restorative materials. J Dent Res, 1970;49:149–158.
[113] Litkowski LJ, Swierczewski M. Root surface marginal microleakage
of composites: comparison of cavosurface finishes. Oper Dent,
1991;16:13–16.
[114] Lutz F, Krejci I, Oldenburg TR. Elimination of polymerization
stresses at the margins of posterior composite restorations. A new
restorative technique. Quintessence Int, 1986;17:777–784.
[115] Martin FE, O’Rourke M. Marginal seal of cervical tooth coloured
restorations. A laboratory investigation of placement techniques.
Austral Dent J, 1993;38:102–107.
[116] McCaghren RA, Retief DH, Bradley EL, Denys FR. Shear bond
strength of light-cured glass ionomer to enamel and dentin. J Dent
Res, 1990;69:40–45.
[117] Mixson JM, Eick JD, Moore DL, Tira DE. Effect of two dentin
bonding agents on microleakage in two different cavity designs.
J Prosthet Dent, 1992;67:441–444.
[118] Momoi Y, Iwase H, Nakano Y. et al. Gradual increases in marginal
leakage of resin composite restorations with thermal stress. J Dent
Res, 1990;69:1659–1663.
[119] Morley F, Stockwell PB. A simple thermal cycling device for testing
dental materials. J Dent, 1977;4:39–41.
[120] Mount GJ, Papageorgiou A, Makinson OF. Microleakage in the
sandwich technique. Amer J Dent, 1992;5:195–198.
[121] Munksgaard EC, Irie M. Dentin–polymer bond established by
Gluma and tested by thermal stress. Scand J Dent Res,
1987;95:185–190.
[122] Nakabayashi N, Kojima K, Masuhara E. The promotion of adhesion
by the infiltration of monomers into tooth substrates. J Biomed Mater
Res, 1982;16:265–273.
[123] Nara Y, Fukuda K, Katsuyama S, Dogon IL. Microleakage of direct
and indirect class II composite restoration. J Dent Res,
1992;71:1628.
[124] Nolden R. Bonding of restorative materials to dentine: the present
status in the Federal Republic of Germany. Int Dent J, 1985;35:166–
172.
[125] O’Brien JA, Retief DH, Bradley EL, Denys FR. Shear bond strength
of a new dentin bonding restorative system. Dent Mater,
1988;4:179–183.
[126] Perdiago J, Swift EJ, Cloe BC. Effects of etchants, surface moisture
and resin composite on dentin bond strengths. Amer J Dent,
1993;6:61–64.
[127] Prati C, Nucci C, Montanari G. Effects of acid and cleansing agents
on shear bond strength and marginal microleakage of glass–ionomer
cements. Dent Mater, 1989;5:260–265.
[128] Prati C, Tao L, Simpson M, Pashley DH. Permeability and
microleakage of class II resin composite restorations. J Dent,
1994;22:49–56.
[129] Prevost AR, Fuller JL, Peterson LC. The use of an intermediate resin
in the acid–etch procedure: retentive strength, microleakage and
failure mode. J Dent Res 1982;61:412–418.
[130] Rebitski G, Donly KJ. Dentin pretreatment and caries inhibition by a
fluoride-releasing resin. Amer J Dent, 1993;6:204–206.
[131] Rees JS, Jacobsen PH, Koliniotou-Kubia E. The current status of
composite materials and adhesive systems. Part 4. Some clinically
related research. Rest Dent 1990;(Aug):4–8.
[132] Reid JS, Saunders WP, Chen YY. The effect of bonding agent and
fissure sealant on microleakage of a composite resin restoration.
Quintessence Int, 1991;22:295–298.
[133] Reilly B, Davis EL, Joynt RB, Quevedo J. Shear strength of resin
developed by four bonding agents used with cast metal restorations.
J Prosthet Dent, 1992;68:53–55.
[134] Retief DH, O’Brien JA, Smith LA, Marchman JL. In vitro investiga-
tion and evaluation of dentin bonding agents [special issue]. Amer J
Dent, 1988;1:176–183.
[135] Retief DH, Mandras RS, Russell CM, Denys FR. Extracted human
versus bovine teeth in laboratory studies. Amer J Dent, 1990;3:253–
258.
[136] Retief DH, Mandras RS, Russell CM, Denys FR. Phosphoric acid as
a dentin etchant. Amer J Dent, 1992;5:24–28.
[137] Retief DH, McCaghren RA, Russell CM. Microleakage of vitrebond/
P-50 class II restorations. Amer J Dent, 1992;5:130–132.
[138] Retief DH, Mandras RS, Russell CM, Denys FR. Evaluation of the
Syntac bonding system. Amer J Dent, 1993;6:17–21.
[139] Rigsby DF, Retief DH, Russell CM, Denys FR. Marginal leakage
and marginal gap dimensions of three dentinal bonding systems.
Amer J Dent, 1990;3:289–294.
[140] Rigsby DF, Retief DH, Bidez MW, Russell CM. The effect of axial
load and temperature on microleakage of resin restorations. Amer J
Dent, 1992;5:155–159.
[141] Roulet J, Reich T, Blunk U, Noack M. Quantitative marginal
analysis in the scanning electron microscope. Scan Elect Microsc,
1989;3:147–159.
[142] Rupp NW, Venz S, Cobb EW. Sealing the gingival margin of
composite restorations [special issue]. J Dent Res, 1986;62:254.
(Abstract 765).
[143] Saunders WP, Grieve AR, Russell EM, Alani AH. The effects of
dentine bonding agents on marginal leakage of composite
restorations. J Oral Rehab, 1990;17:519–527.
[144] Scott JA, Saunders WP, Strang R. Microleakage of a composite inlay
system. Amer J Dent, 1992;5:177–180.
[145] Sidhu SK, Soh G, Henderson LJ. Effect of dentin age on effective-
ness of dentin bonding agents. Oper Dent, 1991;16:218–222.
[146] Sidhu SK. Sealing effectiveness of light-cured glass ionomer cement
liners. J Prosthet Dent, 1992;68:891–894.
[147] Souza MHS, Retief DH, Russell CM, Denys FR. Shear bond strength
and microleakage of All-Bond. Amer J Dent, 1993;6:148–154.
[148] Srisawasdi S, Boyer DB, Reinhardt JW. The effect of removal of the
smear layer on microleakage of class V restorations in vitro. Dent
Mater, 1988;4:384–389.
[149] Staninec M, Holt M. Bonding of amalgam to tooth structure: tensile
adhesion and microleakage tests. J Prosthet Dent, 1988;59:397–402.
[150] Staninec M, Kawakami M. Adhesion and microleakage tests of a
new dentin bonding system. Dent Mater, 1993;9:204–208.
[151] Swift EJ. Microleakage of dentin adhesive systems. J Esthet Dent,
1991;1:91–94.
[152] Swift EJ, Le Valley BD. Microleakage of etched-dentin composite
resin restorations. Quintessence Int, 1992;23:505–508.
[153] Swift EJ, Denehy GE, Beck MD. Use of phosphoric acid etchants
with Scotchbond Multi-Purpose. Amer J Dent, 1993;6:88–90.
[154] Thordrup M, Isidor F, Horsted-Bindslev P. Comparison of marginal
fit and microleakage of ceramic and composite inlays: an in vitro
study. J Dent, 1994;22:147–153.
[155] Wendt SL, Jebeles CA, Leinfelder KF. The effect of two smear
layer cleansers on shear bond strength to dentin. Dent Mater,
1990;6:1–4.
[156] White SN, Sorensen JA, Kang SK, Caputo AA. Microleakage of new
crown and fixed partial denture luting agents. J Prosthet Dent,
1992;67:156–161.
[157] Williams PT, Hedge GL. Thermally induced degradation of dental
amalgam margins [special issue]. J Dent Res, 1983;62:172. (Abstract
S23).
[158] Wu W, Cobb EN, Dermann K, Rupp NW. Detecting marginal leak-
age of dental composite restorations. J Biomed Mater Res,
1983;17:37–43.
[159] Yoshida K, Matsumura H, Atsuta M. Monomer composition and
 M.S. Gale, B.W. Darvell / Journal of Dentistry 27 (1999) 89–99
bond strength of light-cured 4-META opaque resin. J Dent Res,
1990;69:849–851.
[160] Yu XY, Joynt RB, Davis EL. Scanning electron microscopic and
energy dispersive x-ray evaluation of two smear layer-mediated
dentinal bonding agents. Quintessence Int, 1991;22:305–310.
[161] Yu XY, Davis EL, Joynt RB, Wieczkowski G. Bond strength
evaluation of a class V composite resin restoration. Quintessence
Int, 1991;22:391–396.
[162] Nielsen TH. Sealing ability of chelate root filling cements: capillary
physical concepts applied to leakages in root-filled teeth. Part 2.
J Endod, 1980;6:777–780.
[163] Nielsen TH. Sealing ability of chelate root filling cements) the triple-
CA test using compressed air. Part 3. J Endod, 1980;6:835–841.
[164] Granath LE, Svensson A. Studies of microleakage with restorative
materials. A new air pressure method. Scand J Dent Res,
1970;78:353–366.
[165] Hansen EK. Visible light-cured composite resins: polymerization
contraction, contraction pattern and hygroscopic expansion. Scand
J Dent Res, 1982;90:329–335.
[166] Munksgaard EC, Irie M. Effect of load cycling on bond between
composite fillings and dentin established by Gluma and various
resins. Scand J Dent Res, 1988;96:579–583.
[167] Hirsch L, Weinreb M. Marginal fit of direct acrylic restorations.
J Amer Dent Assoc, 1958;56:13–21.
[168] Asmussen E. The effect of temperature changes on adaption of resin
fillings. II. Acta Odontol Scand, 1974;32:291–297.
99
[169] Kidd EAM, Harrington E, Grieve AR. The cavity sealing ability of
composite restorations subjected to thermal stress. J Oral Rehab,
1978;5:279–286.
[170] Fiasconaro J, Sherman HA. A quantitative method for evaluating the
cavity sealing properties of restorations. J Dent Res, 1952;31:503.
(Abstract 114).
[171] Eliades G. Clinical relevance of the formulation and testing of
dentine bonding systems. J Dent, 1994;22:73–81.
[172] Darvell BW. Guest Editorial: The First Three Questions. Austral
Dent J, 1995;40:397–398.
[173] Carracho AL, Chappell RP, Glaros A. Storage and thermocycling
effects on bond strength of dentin adhesives. J Dent Res,
1990;69:285. (Abstract).
[174] Leung VWH, Darvell BW. The calcium phosphate system in a
saliva-like medium. J Chem Soc Faraday Trans, 1991;87:1759–
1764.
[175] Soderholm KJM. Correlation of in vivo and in vitro performance of
adhesive restorative materials: a report of the ASC MD156 Task
Group of test methods for the adhesion of restorative materials.
Dent Mater, 1991;7:74–83.
[176] Douglas WH. Clinical status of dentine bonding agents. J Dent,
1989;17:209–215.
[177] Jendresen MD, Allen EP, Bayne SC. et al. Report of the Committee
on Scientific Investigation of the American Academy of Restorative
Dentistry. J Prosthet Dent, 1992;68:137–190.