Original Paper

Received: July 23, 2002

Caries Res 2004;38:67–74
Accepted after revision: August 28, 2003
DOI: 10.1159/000073923

Scanning Electron Microscopy of

Softened Enamel
M. Eisenburger a R.P. Shellis b M. Addy b
a Department of Prosthetic Dentistry, Medical University of Hanover, Hanover, Germany;
b Division of Restorative Dentistry, University of Bristol Dental School, Bristol, UK

Key Words
Demineralisation W Enamel softening W Replicas W
Scanning electron microscopy
Abstract
After exposing enamel specimens to 0.3% citric acid at
pH 3.2 for various times, the acid was titrated to pH 7
before rinsing the specimens in water. After freeze-dry-
ing the specimens were examined by scanning electron
microscopy. This procedure eliminates artefacts due to
drying and mineral precipitation. The results showed
that the outer region of softened enamel is much more
delicate than previously thought, even after short (5- to
20-min) etching times. Mineral was lost from both prism
boundaries and the prism bodies, resulting in a surface
presenting thin, separate crystal bundles. In further stud-
ies, replicas of subsurface pores, created by resin im-
pregnation, showed the softening depth to be several
times greater than is suggested by techniques based on
removing the softened enamel by physical forces. The
results point to a need for improved methods of measur-
ing softening depth. More importantly, it appears that
the outer region of the softened layer remaining after an
erosive challenge might be too fragile to resist frictional
forces in vivo.
Copyright © 2004 S. Karger AG, Basel
Acid erosion causes a bulk loss of surface enamel and
leaves a thin layer of softened, partly demineralised sur-
face enamel. After exposure to citric acid, the softened
layer seems to be about 2–4 Ìm thick [Schweizer-Hirt et
al., 1978; Eisenburger et al., 2000]. Owing to its reduced
mineral content this layer has a lowered physical resis-
tance and can be easily removed by mechanical forces
such as abrasion [Schweizer-Hirt et al., 1978; Davis and
Winter, 1980; Attin at al., 1999; Jaeggi and Lussi, 1999]
or ultrasonication [Eisenburger et al., 2000]. There is
great interest in the possibility of remineralising softened
enamel, since this would result in its rehardening and
preservation. Softened enamel has been remineralised
experimentally by exposure to supersaturated solutions in
vitro [Koulourides, 1968; Feagin et al., 1971; White and
Nancollas, 1988; Collys et al., 1990; Eisenburger et al.,
2001a] and by exposure to the oral environment in situ
[Gedalia et al., 1991; Collys et al., 1993; Attin et al.,
2000].
The aim of the work reported here was to examine
enamel softened in vitro by scanning electron microscopy
(SEM) using two approaches. We directly examined the
enamel surface exposed to the acid solution. However,
this approach can give only limited information about
subsurface changes so, in addition, we used a replica tech-
nique as an indirect means of visualising the internal pore
structure of softened enamel. Such techniques have been

© 2004 S. Karger AG, Basel Dr. Michael Eisenburger

ABC 0008–6568/04/0381–0067$21.00/0 Medical University of Hanover, Department of Prosthetic Dentistry
Fax + 41 61 306 12 34 Carl-Neuberg-Strasse 1
198.143.54.65 - 1/27/2016 4:51:19 AM

E-Mail karger@karger.ch Accessible online at: DE–30625 Hannover (Germany)

www.karger.com www.karger.com/cre Tel. +49 511 532 4769, Fax +49 511 532 4790, E-Mail M.Eisenburger@gmx.de
Kungliga Tekniska Hogskolan
Downloaded by:
used in previous studies in cariology [Shellis and Halls-
worth, 1987; Shellis, 1996; Gray and Shellis, 2002].
There have been numerous SEM studies of acid-etched
or softened enamel surfaces but all used a preparative pro-
cedure in which, following exposure to acidic solution, an
enamel specimen is rinsed in water. Except in a few stud-
ies which used critical point drying [Meurman and Frank,
1991a, b], the washed specimens were air-dried, either
directly or after removing the water by washing in alcohol.
Specimens prepared in this way are prone to two kinds of
artefact. Firstly, there may be precipitation of mineral at
the surface, resulting from a sudden rise in pH at the sol-
id/solution interface when samples are transferred from
the erosive medium to water [Boyde et al., 1978]. Sec-
ondly, the surface of softened enamel is likely to be unsta-
ble because of thinning of the crystals. Boyde et al. [1978]
suggested that the SEM appearance of etched enamel is
influenced by collapse of the surface structure, due to sur-
face tension effects during the drying process. Therefore,
it was necessary, as part of this study, to develop a tech-
nique for avoiding these artefacts.
Materials and Methods
Sample Preparation
Enamel specimens were cut from buccal and lingual surfaces of
unerupted human third molar teeth and embedded in epoxy resin
blocks as described previously [Eisenburger et al., 2000, 2001a, b].
For the current study blocks were produced in two sizes: small blocks
in moulds measuring 8 ! 5 ! 2 mm and large blocks in moulds
measuring 30 ! 10 ! 7 mm. All specimens were polished with
1,000-grit abrasive papers to produce flat exposed enamel surfaces.
Parallel strips of self-adhesive PVC tape were applied to the polished
surface of each block to leave an exposed window of enamel approxi-
mately 2 mm wide.
Acid Treatment
Samples were exposed at 35 B 0.2 ° C to 0.3% citric acid adjusted
to pH 3.2 with NaOH. The samples were placed at the bottom of a
1-litre glass beaker 10 cm in diameter, containing 300 ml citric acid
stirred with a two-blade propeller stirrer (RW 11 basic, IKW-Werke,
Staufen, Germany) at 270 rpm. Most samples were exposed to acid
for 2 h, as in previous studies, but 6 further samples were exposed for
5, 10 and 20 min (2 samples each).
Samples for Direct Observation
Softened Enamel
To assess the effect of preparative technique on the SEM appear-
ance of the softened enamel surface, four techniques were applied to
68 Caries Res 2004;38:67–74
12 specimens exposed to acid for 2 h (3 by each technique). In addi-
tion, the 6 specimens exposed for 5–20 min were processed by tech-
nique 1:
(1) Titrated/Freeze-Dried. To reduce the chances of artefactual
reprecipitation on transferring specimens from the citric acid to
water, the acid was carefully neutralised after the exposure to acid
before removing the specimens. In the case of the specimens exposed
to acid for 2 h, 0.5 ml 1 mol/l NaOH was added every 30 s until the
solution reached pH 7.0. The specimens exposed for 5–20 min were
titrated somewhat more quickly; 0.5 ml 1 mol/l NaOH was added as
soon as the pH approximately stabilised after addition of the pre-
vious aliquot, resulting in a rate of addition of about 0.5 ml every
20 s. After titrating to neutrality, all samples were placed in de-
ionised water. To avoid collapse of surface structure through surface-
tension effects, the samples were freeze-dried. Specimens were
removed from water and directly submersed in iso-pentane cooled to
near its freezing point in liquid nitrogen. The frozen specimens were
then placed on the stage of an EPTD-2 tissue drier (Edwards High
Vacuum, Crawley, Sussex, UK), pre-cooled to –60 ° C, and freeze-
dried for 24 h.
(2) Titrated/Air-Dried. The citric acid was titrated as in (1) above,
then rinsed by transferring them to de-ionised water. After rinsing,
the specimens were removed from the water and allowed to air-dry.
(3) Rinsed/Freeze-Dried. At the end of the exposure to acid, spec-
imens were rinsed by transferring directly from citric acid to de-
ionised water. The specimens were then freeze-dried as above.
(4) Rinsed/Air-Dried. Samples were rinsed in de-ionised water
directly after exposure to acid, then removed from the water and
allowed to air-dry.
Softened/Ultrasonicated Enamel
Six specimens exposed to acid for 2 h were rinsed with distilled
water. To remove the softened layer, the specimens were ultrasoni-
cated for 30 s at 20 ° C in 70 ml 0.9% saline in a plastic beaker which
was in turn placed in 300 ml tap water in an ultrasonic bath (Pul 55,
Kerry Ultrasonics Ltd., Hitchin, Hert. UK: power output 100 W at
38 kHz). The specimens were then air-dried.
Replicas
Specimens
Eight specimens (4 small, 4 large blocks) were exposed to acid for
2 h, rinsed with distilled water and the tape strips removed. Two
specimens of each type were ultrasonicated for 30 s as above and
rinsed with distilled water.
Resin Impregnation
The specimens were air-dried. The enamel window was then con-
tinuously moistened with drops of acetone for a period of 1 min and
the specimens left to air-dry again. A dentine adhesive (Scotchbond:
3M Dental Products, St. Paul, Minn., USA) was used as the impreg-
nation resin. A first coat of the bonding agent was applied using a
disposable mini-sponge applicator and a second coat applied after
30 s. The resin was light-cured for 20 s using a mains-powered light
curing unit.
Preparation of Complete Replicas
Small blocks were bonded using fissure sealant (Conseal F, South-
ern Dental Industries Ltd., Bayswater, Australia) to a Perspex block
Eisenburger/Shellis/Addy
198.143.54.65 - 1/27/2016 4:51:19 AM
Kungliga Tekniska Hogskolan
Downloaded by:
approximately 1 ! 1 cm square with the enamel window contacting
the Perspex. The fissure sealant was light-cured for 60 s. The surface
of the block facing away from the Perspex was ground on a lapping
machine to reduce the thickness to less than 1 mm. The specimens
were then exposed to 1 mol/L HCl until the enamel had been dis-
solved completely (approximately 4 h). Finally the specimens were
gently rinsed with distilled water and air-dried.
Preparation of Cross Section Replicas
On the large blocks, a thin layer of fissure sealant was applied on
top of the dentine adhesive and light-cured for 60 s. This measure
was taken to stabilise the bonding agent during the following section-
ing and polishing procedures. Specimens were then sectioned in two
across the acid-treated area using a water-cooled diamond-edged
blade in a sectioning machine (Microslice 2: Malvern Instruments,
Malvern, UK). The cut surfaces of both halves were hand-polished
with 3-Ìm diamond paste on a polishing disc (Engis, Maidstone,
UK). The lengths of the halved blocks were then reduced by cutting
off the excess on the sectioning machine, to fit the SEM stubs. The
resulting specimens were etched in 1 mol/l HCl for 5 s, rinsed with
distilled water and immediately freeze-dried.
Scanning Electron Microscopy
All samples were glued to aluminium SEM stubs and then sput-
ter-coated twice for 30 s with gold and examined at 20 kV in the
secondary emission mode in a PC-controlled ISI 60 scanning electron
microscope (SEM Tech Ltd., Bonsall, Derbyshire, UK).
Results
Direct Observation of Enamel Surface
Titration of the citric acid solution to pH 7.0 took 5.5–
6 min. The pH reached 4.5 after 2–2.5 min. In specimens
exposed to acid for 2 h, then titrated to neutrality before
rinsing and freeze-drying, the prisms appeared as brush-
like bundles of delicate crystals or groups of crystals sepa-
rated by etched prism junctions (fig. 1). In the prism bod-
ies the crystals were mostly parallel with the prism direc-
tion, although some deviation from this orientation was
seen at the outermost part of the surface. The difference in
crystal orientation between the prism bodies and the
prism tails was easily distinguished (fig. 1).
Specimens that had been exposed to acid for shorter
times showed thin, delicate crystals projecting from the
surface, even after 5 min (fig. 2), but the crystals were
somewhat shorter than after 2 h of acid treatment
(fig. 1, 2).
Various combinations of treatments are compared in
figures 3–5. Titration to neutrality, followed by rinsing in
water and air-drying, produced a surface in which the
prism bodies were elevated and separated from each other
Microstructure of Softened Enamel
by grooves marking the prism boundaries. However, the
crystals were clumped together to give each prism body a
pointed appearance (fig. 3). A crystalline enamel structure
could be recognised in specimens that had been rinsed
with water and then freeze-dried (fig. 4), but in this case
the surface was partly obscured by a network of material
that consisted partly of sheet-like material and partly of
rod-shaped crystals (fig. 4). The extraneous crystals were
scattered widely over the surface and also formed chain-
like aggregates. Finally, specimens that were rinsed with-
out prior titration and then air-dried showed surfaces cov-
ered in a feltwork of randomly orientated crystals lying
parallel with the specimen surface. The specimen surfaces
showed round lighter areas, apparently associated with
prisms, where the crystals appeared to be more closely
packed (fig. 5).
The surfaces of specimens that had been ultrasonicated
after 2-hour exposure to acid showed less relief than in
specimens that had only been acid-treated. The prism
bodies were demarcated by a change in crystal orienta-
tion, and sometimes by a shallow depression, from the
prism tails (fig. 6). All parts of the surface were obviously
crystalline, but the exposed crystals were much shorter
and more robust than in surfaces that had not been ultra-
sonicated (fig. 1, 6).
Complete Replicas
These presented a view of the internal pores, viewed
from beneath the enamel surface. In specimens which had
simply been acid-treated for 2 h before replication, a hon-
eycomb was formed by replicas of sheet-like, intercon-
nected pores at the prism boundaries (fig. 7). The upper-
most edges of the prism-boundary replicas appeared
rounded. In the intervening pits which were the sites of
the prism bodies, the surface of the replica was often
rough and perforated by numbers of holes (fig. 7). In spec-
imens which had been ultrasonicated after exposure to
acid, the replicas again had a honeycomb-like appearance
but the pits were shallower (fig. 8). A commonly observed
feature was the presence of a row of prominent holes at
the junction between the prism body and the prism
boundary.
Cross Section Replicas
These gave a lateral view of the morphology of the
internal pores within the softened layer. In specimens that
had been only acid-treated, most of the pore replicas
appeared as sheet-like structures which, from their tube-
like form and their spacing, represented the prism bound-
aries (fig. 9). The prism-boundary replicas extended for
Caries Res 2004;38:67–74 69
198.143.54.65 - 1/27/2016 4:51:19 AM
Kungliga Tekniska Hogskolan
Downloaded by:
 70Caries Res 2004;38:67–74
Eisenburger/Shellis/Addy
Downloaded by:
Kungliga Tekniska Hogskolan
198.143.54.65 - 1/27/2016 4:51:19 AM
Fig. 1. Surface of enamel softened by acid treatment for 2 h.
Titrated/freeze-dried specimen, showing prisms consisting of bun-
dles of fine crystals. The change in crystal orientation at the prism
boundaries can be seen. Bar = 2 Ìm.
Fig. 2. Surface of enamel softened by acid treatment for 5 min.
Titrated/freeze-dried specimen, showing prisms from the surface of
which project thin, delicate crystals. Bar = 5 Ìm.
Fig. 3. Surface of enamel softened by acid treatment for 2 h.
Titrated/air-dried specimen. The crystalline structure of the prism
bodies and tails is evident but there is obvious aggregation, and the
fine crystals visible in freeze-dried specimens (fig. 1) cannot be dis-
tinguished. Bar = 2 Ìm.
Fig. 4. Surface of enamel softened by acid treatment for 2 h. Rinsed/
freeze-dried specimen. The finely crystalline structure of the true sur-
face (see fig. 1) can be seen, but this is overlaid by a network-like
deposit which is partly amorphous and sheet-like, and partly crystal-
line. Bar = 2 Ìm.
Fig. 5. Surface of enamel softened by acid treatment for 2 h. Rinsed/
air-dried specimen. The true enamel surface is covered by randomly
arranged rod-shaped crystals. Round, light areas where crystal densi-
ty appears to be highest appear to correspond in size and spacing to
prisms. Bar = 2 Ìm.
Microstructure of Softened Enamel
Fig. 6. Surface of enamel softened by acid treatment for 2 h, follow-
ing ultrasonication and air-drying. Crystal aggregates, much coarser
than in figure 1, are seen. The division into prisms is marked by
changes in orientation of the crystal bundles. Bar = 2 Ìm.
Fig. 7. Complete replica of enamel softened by acid treatment for
2 h, viewed from beneath the original surface. A continuous, pseudo-
hexagonal network of prism-boundary replicas is prominent. The
intervening surfaces, representing intraprismatic pores, are rough
and present holes which are presumably the sites of large crystal
aggregates. Bar = 4 Ìm.
Fig. 8. Complete replica of enamel softened by acid treatment for
2 h, then ultrasonicated. Prism-boundary replicas are much shorter
than before ultrasonication (see fig. 6). Pronounced holes in the intra-
prismatic regions tend to be concentrated next to the prism bound-
aries. Bar = 4 Ìm.
Fig. 9. Cross section replica of enamel softened by acid treatment for
2 h, disclosed by etching with HCl. The resin layer is at the top of the
field. Replicas of prism boundaries (examples marked by asterisks)
extend into the enamel. Between these, rough profiles represent the
surface of replicas of intraprismatic pores. Bar = 4 Ìm.
Fig. 10. Cross section replica of enamel softened by acid treatment
for 2 h, then ultrasonicated, disclosed by etching with HCl. The resin
layer is at the top of the field. Prism-boundary replicas (examples
marked by asterisks) are very short (see fig. 8). Bar = 4 Ìm.
Caries Res 2004;38:67–74 71
198.143.54.65 - 1/27/2016 4:51:19 AM
Kungliga Tekniska Hogskolan
Downloaded by:
9–12 Ìm from the inner boundary of the Scotchbond into
the enamel, becoming thinner with depth. Between the
prism-junction replicas the surface of the replica appeared
jagged (fig. 9). In specimens which had been ultrasoni-
cated after exposure to acid, the prism-boundary replicas
were distinctly shorter and thicker (fig. 10).
Discussion
The concentration and pH of the citric acid solution
used here were selected on the basis of analyses of soft
drinks tabulated by Hughes et al. [1999]. Under the con-
ditions used here, the thickness of the softened layer pla-
teaus after 1 h, so is at a maximum after the 2-hour acid
exposure time [Eisenburger et al., 2000, 2001a, b], pre-
sumably having reached a steady state in which the rate of
enamel loss at the surface equals the rate of acid penetra-
tion into the tissue. The shorter acid exposure times were
chosen to be more representative of single intakes of ero-
sive drinks.
As noted in the ‘Introduction’, Boyde et al. [1978]
identified two possible artefacts in acid-treated enamel
surfaces and we have adopted techniques intended to
eliminate or minimise them. Our comparative experi-
ments support the hypothesis [Boyde et al., 1978] that
simple, conventional preparation methods tend to intro-
duce artefact. The crystals which covered, partially or
completely, the surfaces of specimens which had been
rinsed without prior neutralisation of the citric acid seem
to be extraneous, since there seems to be no reason, if they
were a genuine part of the acid-treated surface, why they
should not also appear on the specimens which had been
titrated before washing. The presence of the crystals is
explicable by the hypothesis [Boyde et al., 1978] that
where an acidic, saturated solution diffusing out from the
subsurface enamel contacts water at the enamel surface,
the resulting rapid neutralisation leads to precipitation.
Whereas rinsed/air-dried specimens were covered with a
quite dense layer of crystals, rinsed/freeze-dried speci-
mens presented a network of material which appeared to
be only partly crystalline. We suggest that the partly amor-
phous appearance is due to the inhibition of crystal
growth by the low temperature during freeze-drying,
while the discontinuous distribution is the result of a
phase separation during freezing, solute-containing solu-
tion becoming concentrated at the interfaces between
growing ice crystals. The clumping of the partly dissolved
enamel crystals seen in specimens air-dried after titration
of the citric acid is attributable to surface tension forces
72 Caries Res 2004;38:67–74
associated with the air/water interface as it moves through
the tissue during drying [Boyde et al., 1978].
We do not, of course, claim that the images presented
here are free of all artefact, only that we have minimised
two kinds of artefact of which we were aware. For
instance, the titration/freeze-drying technique introduces
some error into the etching time, which is extended by the
titration process. However, demineralisation rate falls
logarithmically with pH [Davis and Winter, 1980] and is
negligible at pH 4.5, which is reached during the titration
procedure after only about 2 min. Therefore, the increase
in demineralisation time is probably relatively small,
even for short acid exposure times, e.g. 5–10 min.
In the mouth, pH does not rise abruptly after ingestion
of an erosive drink [Millward et al., 1997], but rises pro-
gressively, at about the same rate used in our titrations.
Moreover, in vivo softened tooth surfaces normally re-
main moist and are not exposed to drying processes which
would expose them to forces tending to collapse the sur-
face structure. On these grounds we believe that the titra-
tion/freeze-drying procedure provides a reasonably accu-
rate image of a surface freshly softened in vivo. Specimens
prepared using both the titration technique and freeze-
drying show that the surface of softened enamel left after
exposure to acid under the conditions used here, with
partly dissolved bundles of crystals separated by relatively
large spaces, is much more delicate than previous SEM
images have suggested. The outermost layer will clearly be
extremely vulnerable to mechanical forces. This has ob-
vious implications for intra-oral wear, which are dis-
cussed later in the paper. However, our observations on
softened surfaces also have implications for accuracy of
measuring the thickness of the softened layer. First, dry-
ing will cause the outer softened layer to collapse, result-
ing in underestimation by any technique. Secondly, the
mineral content of the outermost softened layer might
well be too low for microradiography. Thirdly, profilom-
eter measurements will be affected by the tendency of the
stylus to penetrate the outer, most demineralised layer.
The replication technique used here is quick and sim-
ple and, as described below, provides useful information
about subsurface changes in structure. However, it is
important to be aware of potential artefacts. Both com-
plete and cross section replicas of softened enamel give
the impression that most of the subsurface porosity was
associated with the prism boundaries. While this is con-
sistent with the higher porosity and raised solubility at
these sites [Shellis, 1996], and with the etch pattern
observed at the surface [Meurman and ten Cate, 1996], it
seems likely that the present replica techniques under-
Eisenburger/Shellis/Addy
198.143.54.65 - 1/27/2016 4:51:19 AM
Kungliga Tekniska Hogskolan
Downloaded by:
represent the extent of intraprismatic demineralisation.
The extremely fragile outer layer will probably contain
such a high proportion of resin that it would appear as
part of the supporting resin, rather than of the enamel.
Replication of deeper, less demineralised intraprismatic
enamel might have been incomplete because the contact
time between resin and enamel was very short, and
because infiltration was not assisted by thorough dehydra-
tion and slow polymerisation as used in more exacting
methods [Shellis and Hallsworth, 1987]. Finally, it was
found in pilot experiments that when complete replicas
were freeze-dried, the thin terminal portions of the prism-
boundary replicas collapsed and formed a film which
obscured the surface of the specimen. For this reason, the
complete replicas were air-dried, but this causes the prism
boundaries to appear shorter and more rounded than in
the cross section replicas.
Despite these limitations, the replicas provided valu-
able information. The complete replicas showed that sof-
tening had resulted in confluence and enlargement of the
prism boundaries, which would have removed support
from the prisms. The cross section replicas show that acid
damage extends to a depth of 9–12 Ìm below the surface
and even this depth will be an underestimate if, as sug-
gested above, the most demineralised part of the softened
layer is not detectable within the resin. These observa-
tions indicate that the depth of 2–4 Ìm measured by ultra-
sonication after 2 h exposure to acid [Eisenburger et al.,
2000] is a considerable underestimate of the softening
depth. The discrepancy is probably accounted for by two
factors. First, the outer, most demineralised part of the
softened layer probably offers so little resistance to a pro-
filometer stylus (see above) that it does not register. Sec-
ondly, the deeper parts of the softened enamel, where
demineralisation appears to be mostly confined to the
prism boundaries, might offer the same resistance as
sound enamel to ultrasonication, thus causing the sof-
tened/sound enamel interface to appear higher than it is.
After ultrasonication of softened enamel, the reduction
in length of the prism-boundary replicas to 3–4 Ìm from
9–12 Ìm showed that most of the softened enamel is vul-
nerable to this physical force. Figure 5 suggests that enam-
el left behind by ultrasonication is still partly demineral-
ised, showing that this technique does not remove all sof-
tened enamel but only that fraction which has passed a
certain threshold of mineral loss. Nevertheless, the differ-
ence in depth of the prism-boundary replicas between
acid-treated and acid-treated/ultrasonicated enamel (5–
9 Ìm) still exceeded previous estimates of softening depth
after 2-hour exposure to acid [Eisenburger et al., 2000].
Microstructure of Softened Enamel
The discrepancy is probably due to penetration of the out-
ermost softened layer by the profilometer stylus, so that
the height of the trace is inaccurate.
Problems in measuring the softening depth, such as
collapse of the outer layer on drying, could in principle be
overcome by the use of confocal laser scanning microsco-
py of hydrated samples. However, so far confocal laser
scanning microscopy has only been applied to early stages
of softening [Duschner et al., 2000] and its application to
advanced softening needs to be tested. In addition, stud-
ies to correlate the CLSM image with mineral loss, as
measured by microradiography, and resistance to physi-
cal forces such as ultrasonication, indentation or abrasion
are required.
It has been shown that etch patterns at the surface of
acid-treated enamel can persist for long periods in the
mouth [Garberoglio and Cozzani, 1979; Collys et al.,
1991; Kuroiwa et al., 1994]. However, the relevance of
these observations in the context of softening is limited,
since in all these studies the enamel was etched with 30–
50% phosphoric acid for 15–120 s, as in clinical practice.
This etching technique is much more aggressive than an
erosive challenge and is usually followed by thorough
washing to remove secondary calcium phosphate precipi-
tates, so results in a mechanically robust surface, capable
of resisting intra-oral forces. In contrast, it seems highly
unlikely that the delicate outermost layer of citric acid-
softened enamel revealed by the titration/freeze-drying
procedure would withstand even mild intra-oral frictional
forces for long. If this were the case, the layer would prob-
ably be lost irreversibly soon after being created by an ero-
sive challenge and would stand little chance of being con-
solidated by remineralisation. To explore further the fate
of softened enamel, further studies to examine specimens
exposed to erosive challenges in situ are necessary.
Caries Res 2004;38:67–74 73
198.143.54.65 - 1/27/2016 4:51:19 AM
Kungliga Tekniska Hogskolan
Downloaded by:
 References
Attin T, Buchalla W, Gollner M, Hellwig E: Use of
variable remineralisation periods to improve
the abrasion resistance of previously eroded
enamel. Caries Res 2000;34:48–52.
Attin T, Deifuss H, Hellwig E: Influence of acidif-
ied fluoride gel on abrasion resistance of
eroded enamel. Caries Res 1999;33:135–139.
Boyde A, Jones SJ, Reynolds PS: Quantitative and
qualitative studies of enamel etching with acid
and EDTA; in Johari O (ed): Scanning Electron
Microscopy 1978. Chicago, SEM Inc., 1978,
vol 2, pp 991–1000.
Collys K, Cleymaet R, Coomans D, Michotte Y,
Slop D: Rehardening of surface softened and
surface etched enamel in vitro and by intraoral
exposure. Caries Res 1993;27:15–20.
Collys K, Cleymaet R, Coomans D, Slop D: Acid-
etched enamel surfaces after 24 h exposure to
calcifying media in vitro and in vivo. J Dent
1991;19:230–235.
Collys K, Slop D, Coomans D: Interaction of mag-
nesium and fluoride in the rehardening and
acid resistance of surface-softened bovine
enamel in vitro. Magnes Trace Elem 1990;9:
47–53.
Davis WB, Winter PJ: The effect of abrasion on
enamel and dentine after exposure to dietary
acid. Br Dent J 1980;148:253–256.
Duschner H, Götz H, Walker R, Lussi A: Erosion
of dental enamel visualized by confocal laser
scanning microscopy; in Addy M, Embery G,
Edgar WM, Orchardson R (eds): Tooth Wear
and Sensitivity. London, Dunitz, 2000, pp 67–
73.
Eisenburger M, Hughes J, West NX, Jandt KD,
Addy M: Ultrasonication as a method to study
enamel demineralisation during acid erosion.
Caries Res 2000;34:289–294.
74 Caries Res 2004;38:67–74
Eisenburger M, Hughes J, West NX, Shellis RP,
Addy M: The use of ultrasonication to study
remineralisation of eroded enamel. Caries Res
2001a;35:61–66.
Eisenburger M, Addy M, Hughes JA, Shellis RP:
Effect of time on the remineralisation of enam-
el by synthetic saliva after citric acid erosion.
Caries Res 2001b;35:211–215.
Feagin FF, Pate PR, Koulourides T, Pigman W:
Study of the effect of calcium, phosphate, fluo-
ride and hydrogen ion concentrations of the
remineralisation of partially demineralised hu-
man and bovine enamel surfaces. Arch Oral
Biol 1971;16:535–548.
Garberoglio R, Cozzani G: In vivo effect of oral
environment on etched enamel: A scanning
electron microscopic study. J Dent Res 1979;
58:1859–1865.
Gedalia I, Ionat-Bendat D, Ben-Mosheh S, Shapiro
L: Tooth enamel softening with a cola type
drink and rehardening with hard cheese or
stimulated saliva in situ. J Oral Rehabil 1991;
18:501–505.
Gray GB, Shellis RP: Infiltration of resin into white
spot caries-like lesions of enamel: An in vitro
study. Eur J Prosthodont Restor Dent 2002;10:
27–32.
Hughes JA, West NX, Parker DM, Newcombe RG,
Addy M: Development and evaluation of a low
erosive blackcurrant juice drink. 3. Final drink
and concentrate, formulae comparisons in situ
and overview of the concept. J Dent 1999;27:
345–350.
Jaeggi T, Lussi A: Toothbrush abrasion of erosively
altered enamel after intraoral exposure to sali-
va: An in situ study. Caries Res 1999;33:455–
461.
Koulourides T: Experimental changes of mineral
density; in Harris RS (ed): Art and Science of
Dental Caries Research. New York, Academic
Press, 1968, pp 355–378.
Kuroiwa M, Kodaka T, Kuroiwa M, Abe M: Brush-
ing-induced effects with and without a non-
fluoride abrasive dentifrice on remineraliza-
tion of enamel surfaces etched with phosphoric
acid. Caries Res 1994;28:309–314.
Meurman JH, ten Cate JM: Pathogenesis and mod-
ifying factors of dental erosion. Eur J Oral Sci
1996;104:199–206.
Meurman JH, Frank RM: Scanning electron micro-
scopic study of the effect of salivary pellicle on
enamel erosion. Caries Res 1991a;25:1–6.
Meurman JH, Frank RM: Progression and surface
ultrastructure of in vitro caused erosive lesions
in human and bovine enamel. Caries Res
1991b;25:81–87.
Millward A, Shaw L, Harrington E, Smith AJ: Con-
tinuous monitoring of salivary flow rate and
pH at the surface of the dentition following
consumption of acidic beverages. Caries Res
1997;31:44–49.
Schweizer-Hirt CM, Schait A, Schmid R, Imfeld T,
Lutz F, Mühlemann HR: Erosion und Abra-
sion des Schmelzes: Eine experimentelle Stu-
die. Schweiz Monatsschr Zahnmed 1978;88:
497–529.
Shellis RP: A scanning electron-microscopic study
of solubility variations in human enamel and
dentine. Arch Oral Biol 1996;41:473–484.
Shellis RP, Hallsworth AS: The use of scanning
electron microscopy in studying enamel caries.
Scanning Microsc 1987;1:1109–1123.
White DJ, Nancollas GH: Kinetic and physical
aspects of enamel remineralisation – a constant
composition study. Caries Res 1988;22:11–19.
Eisenburger/Shellis/Addy
198.143.54.65 - 1/27/2016 4:51:19 AM
Kungliga Tekniska Hogskolan
Downloaded by: