See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/280570300

Restoration of semi-industrial glazed ceramic tiles by re-firing

Conference Paper · July 2015

CITATION READS

1 415

5 authors, including:

Antonina Chaban Silvia Pereira

Italian National Research Council National Laboratory for Civil Engineering
19 PUBLICATIONS   59 CITATIONS    53 PUBLICATIONS   383 CITATIONS   

SEE PROFILE SEE PROFILE

Lurdes Esteves João M. Mimoso

Museu Nacional do Azulejo, Lisboa, Portugal Laboratório Nacional de Engenharia Civil- LNEC, Portugal, Lisbon
31 PUBLICATIONS   78 CITATIONS    66 PUBLICATIONS   227 CITATIONS   

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Advanced Raman Spectroscopy: searching for the molecular fingerprints in artwork (RS4ART) View project

CerAzul View project

All content following this page was uploaded by Antonina Chaban on 30 July 2015.

The user has requested enhancement of the downloaded file.

RESTORATION OF SEMI-INDUSTRIAL GLAZED
CERAMIC TILES BY RE-FIRING

Antonina Chaban
University of Bologna, Italy, [email protected]
Silvia R. M. Pereira
Laboratório Nacional de Engenharia Civil, Lisbon, Portugal, [email protected]
Maria de Lurdes Esteves
Museu Nacional do Azulejo, Lisbon, Portugal, [email protected]
Fausto Peddis
University of Bologna, Italy, [email protected]
João Manuel Mimoso
Laboratório Nacional de Engenharia Civil, Lisbon, Portugal, [email protected]

SUMMARY: Semi-industrial glazed ceramic tiles lining urban façades represent a

rich heritage asset of Portuguese cities and in particular of Lisbon. However, still a sizeable
part of them is being continuously lost. Whenever the façade tiles are decayed, they are
usually merely removed and replaced by newly manufactured ones.
In this communication we introduce re-firing as an alternative to simple disposal of
damaged original tiles. We assess the technique as a technological option for the restoration
of crazed and spalled tiles. The re-firings are applied at different heatworks and
temperatures and the effects of all tests are studied by visual, macroscopic observation and
through the glaze resistance to a pull-off. The effect of continuous serviceability of the re-
fired tiles is assessed through accelerated ageing tests. The domain of application of the
technique and the shortcomings detected are presented and discussed in this communication.
Dwelling on a new ground, we present the initial results on which further research in
this field can be based.

KEY-WORDS: azulejos; façade tiles; restoration; re-firing of ceramics; acquired calcium

carbonate in ceramics
 1. INTRODUCTION
Since the 16th century azulejos (ceramic tiles) are closely related with the
life and culture of the Portuguese. There is no other country in the world
where we can admire such an abundant heritage of azulejos as in Lisbon,
Oporto, Evora and many other Portuguese towns and even small villages.
They are used wherever large areas are available, in the interiors of churches
and palaces, cloisters, halls, staircase walls of multi-storey buildings and,
since the second quarter of the 19th century, covering whole urban façades
[1]. The use of azulejos in Portugal is arguably its most original contribution
to the artistic heritage of Europe and so they represent a valuable asset.
The historical individually painted tiles are easily conceived as art creation
and so different techniques are developed and applied to their conservation
and restoration [2]. However, the semi-industrial tiles are still considered
expendable and in towns such as Lisbon a sizeable part is lost every year
through decay and neglect [3].
The township of Ovar tries to restore the tiles of its urban façades [4, 5], but
with few exceptions whenever façades are restored, decayed azulejos are
usually removed and replaced by newly manufactured tiles, often of inferior
aspect and of inferior quality as well. The original tiles removed from the
walls are usually simply thrown away. In some cases they may be packed
and stored by the town authorities for future reference and such cases they
may one day be given new life…
The individually painted panels represent integrated art that complements
Architecture and thus hold a set of values (both in themselves and through
their integration) that preclude the impairment of their historical value
through re-firing, unless there were no other means to preserve the image
[6]. The façade tiles, unlike the historical azulejo panels, are repeating units
deriving from a semi-industrial process, all similar and holding the same
intrinsic values. Because during the process of urban restorations the
damaged tiles are simply thrown away, any technique allowing their
continued serviceability would constitute a better option, even if some values
such as the historical record of the original manufacturing technique were
impaired or even lost in the process.
 2. PREVIOUS STUDIES AND OBJECTIVES OF THE
RESEARCH
Three years ago LNEC found out as a by-product of another research that in
at least some cases crazing could be mended and a decaying adhesion could
eventually be restored by re-firing the tiles [7] and proposed to dwell into the
possibility of restoring façade tiles by re-firing them.
The tests were done at the Museu Nacional do Azulejo (National Azulejo
Museum, in Lisbon) in 2012 on three historical hand-painted azulejos of the
17th and 18th centuries and on a single semi-industrial (early 20th century)
tile. Tiles were cut in four probes; one was kept in its original condition and
the other parts were fired at 890°C, 990° and 1100°C during 20 minutes with
a heating rate of 240°C per hour. The first tests showed that at least in some
cases crazing could be reversed and the glaze to biscuit adhesion was not
seen to be impaired. Actually the pull-off strength could potentially be
increased by re-firing to a sufficiently high temperature [7]. The subject has
been pursued within the present research project aiming at developing a
feasible restoration method usable for semi-industrial patterned tiles. The
research was pursued at LNEC in collaboration with the University of
Bologna originating a MSc dissertation “Restoration of semi-industrial
glazed ceramic tiles by re-firing”, defended in Italy in March 2015 [8].

3. RE-FIRINGS
35 semi-industrial samples were selected, representative in terms of the
period (2nd half of the 19th century to early 20th century) and decay forms
(crazing and glaze loss). They were tentatively dated according to their
technical and stylistic features. The industrial provenance of several samples
was known and they were manufactured by Fábrica Roseira, one of the first
and main protagonists of façade tiles production in Lisbon [9]. The residuals
of mortar on the tiles were removed manually as a possible source of
gaseous emissions during the re-firing; the tiles were cleaned and desalinated
according to the standard procedure of immersion in de-ionized water [10].
The whole tiles and larger fragments were first cut in four or five test items,
some being subsequently re-divided into smaller pieces for comparative re-
firings, ageing procedures or characterization purposes.

3.1. Methodology outline

The re-firing technique was applied at several heatworks and maximum
temperatures, and characterized in terms of its efficacy addressing the
following problems: i) glaze crazing; ii) insufficient glaze to biscuit
adhesion; and iii) reintegration of lacunae in areas where the glaze had been
lost.
At the beginning we planned to use only simple cycles made up of
continuous heating and cooling ramps at constant rates which we call here
single-stage re-firings. They have been applied on different samples to 890,
930, 950, 990 and 1100°C with a ROHDE KE-150 L kiln (front loader,
maximum temperature 1280°C). In all cases a heating ramp of 240º/h was
used because it is the standard heating rate used by the Museu Nacional do
Azulejo and it is the fastest recommended by their long experience on the
glazing of reproductions [Personal communication by Dr. Maria de Lurdes
Esteves].

3.2. Single stage re-firing

The single stage re-firing obliterated the crazing, restored the glaze
whiteness and shine (Fig. 1) and showed good promise in terms of lacunae
restoration for most samples. But in three of 35 tiles it failed. In these cases a
restoration was not achieved due to bubbles protruding from the glaze
surface or extensive “orange skin” on the surface.
At the same time, all the samples showed an increase in porosity when re-
fired at the lower range of temperatures but a decrease in open porosity in all
cases when re-fired at 1100°C. The results of mercury intrusion porosimetry
have showed a slight increase in pore size in the samples re-fired at 890 and
950°C (Fig. 2).
The glaze to biscuit adhesion did not show to worsen after re-firings and in
two cases small pieces of detached glaze were verified to have adhered after
re-firing in a way that they could not be separated with the nail. In no case
did a pull-off test detect any evident impairment of the glaze to biscuit
adherence but it made evident that the cohesion of the biscuit decreased on
the firings at lower temperatures but increased in all cases at 1100ºC .
The micro-hardness Knoop tests and nano-indentation tests showed that
there was no systematic alteration and no significant changes in the glaze
mechanical properties due to re-firing [8].
a) b) c)

d) e)
Figure 1. Visual observation of single-stage re-firing results: a) and d) before re-firing;
b) after re-firing at 890°C during 3 hours; c) glossy glaze surface after re-firing e)
after re-firing at 930°C during 20 minutes

b)
a)
Figure 2. Pore size distribution diagrams: a) before re-firing, peak ø 0,3-0,4 μm; b) after re-
firing at 950°C during 40 minutes, peak ø 0,6-0,7 μm
The microscopic observation of the glaze (Fig. 3) in the studied samples has
shown that crazing cracks had been effectively sealed through re-firing
eventually leaving scar-like marks. The number of bubbles in the glaze
decreased and their reduced presence close to the glaze/biscuit interface
makes the re-fired tile potentially more resistant to degradation.

a) b)

Figure 3. OM observations of single-stage re-firings: a) before re-firing; b) after re-

firing at 950°C during 20 minutes

a) b) c)
Figure 4. Pull off tests: a) ultimate force results before and after re-firing (kN);
b) fracture in not re-fired condition; c) facture after re-firing at 950°C during 40
minutes.
The impairment of the glaze surface in some cases was now addressed. The
fact that bubbling through the glaze now compromised the aesthetics of the
decoration in tiles that were originally fired at similar or higher temperatures,
suggested that what was causing this problem had been acquired during the
service lifetime of the tile. Since free water and organic material are lost or
consumed at lower temperatures than that at which the glaze softens [11], a
de-carbonation process was suspected. This hypothesis led to a study of the
calcite content of the samples.
Although the biscuits of recently fired azulejos cannot have a sizeable
content in carbonates, in aged units calcite may eventually have been formed
through re-carbonation of lime (either acquired from the walls or remaining
in the biscuit after the original firing) [11]. And indeed, the presence of
calcite was confirmed by loss on ignition tests performed at LNEC (see
Table 1) confirmed by X-ray diffraction analysis performed at the University
of Bologna [8].
Table 1. Loss on ignition results
Loss of CO2
CaCO3
Sample Re-firing (500-900ºC)*
(%)
(g) (%)
AC 26a None 0,032 3,21 7,30
Single stage firing
AC 26a 0,0015 0,19 0,43
990°C for 5 minutes
Double stage firing
AC 26c2 0,0015 0,15 0,34
700°C during 3 hours at 950°C for 5 minutes
* assumed that all losses in this temperature range are due to release of CO2

An analysis by SEM-EDS of a re-fired sample identified lime or calcite

deposited in the hollows accessible to solutes circulating in the walls and
biscuits. Upon firing the calcite is transformed into lime or combined e.g.
with silica to form gehlenite or calcium silicates such as wollastonite,
releasing carbon dioxide in the process (Figure 5).

Figure 5. SEM observations of a re-fired sample.

The increase in mass by accelerated carbonation testing (at LNEC)
confirmed that: i) there is lime present in the samples indicating that not all
calcium was combined to form new chemical species (such as gehelenite and
wollastonite) during the original firing, or that lime was acquired afterwards
from the mortars; and ii) that lime is accessible and readily carbonates, re-
forming calcite.

3.3. Double stage re-firings

To address the unexpected problem we decided to replicate in part the
manufacture process. We introduced double-stage re-firings and evaluated
their effect on different samples. The heatwork relevant for the re-firing of
the glaze (i.e. above its softening temperature) is the same both in single and
double stage re-firings. The difference is that double-stage re-firing
processes foresee extra time for calcite transformations inside the ceramic
body before the softening of the glaze. In practice it means that the heating
ramp is stopped and maintained for a period at the temperature point when
the carbonates decomposition has already started but the glaze has still not
melted (600-700°C).
The method worked well and the bubbling problem was successfully solved
(Fig. 6). The further tests were aimed at finding the minimal duration and
optimal temperature to let the carbonates decompose and the carbon dioxide
escape from the ceramic body. Heating ramps of 200°C/h and 240°C/h were
tested, as well as the effect of pre-firings on 21 samples at 550, 600, 650 and
700ºC during 1-4 hours. The maximum temperatures evaluated in double-
stage re-firings were 890, 950 and 990°C. However, it was still not possible
to avoid completely the appearance of “orange skin” on the surface of one
tile. We can assume that this anomalous result is caused by the glaze
softening at a lower temperature than in the other tiles so that during the de-
carbonation phase the gas already starts bubbling through the glaze.
 Figure 6. Visual observation of double-stage re-firing results: a) not re-fired; b)
single-stage refiring (990°C 5 minutes); c) double-stage re-firing

4. AGEING TESTS
To evaluate the future behavior of tiles restored by re-firing, two types of
accelerated ageing tests were performed. The tiles were aged artificially by
boiling in water according to the general lines of EN-155 Standard
(maximum moisture expansion stress applied) and by exposure to a NaOH
solution at 60°C during 30 days. Both ageing tests have shown that: i) the
tests used are effective in the sense that they cause decay; ii) The re-fired
samples followed the same ageing trend as their un-refired counterparts.
The re-firing treatment did not influence in any noticeable systematic
manner the degree of pull-off resistance decay after ageing in boiling water.
More revealing: the new crazing detected after the ageing in boiling water
did not coincide with the original crazing means that the re-firing did reverse
the fissures in a way that the scars do not constitute preferential propagation
paths to new crazing phenomena (Figure 7).

a) b)

Figure 7. Comparison of crazing pattern: a) not re-fired; b) re-fired and aged in

boiling water
 5. APPLICATION OF THE TECHNIQUE TO FULL
RESTORATION OF TILES
The technique has been applied and evaluated as a technological option for
full restoration of spalled tiles (Figures 8). The lacunae in 8 spalled tiles have
been reintegrated with new glaze CE VTR- 81 (from Ferro) widely used
nowadays in tiles production, painted and subsequently re-fired. The double-
stage firing at 650°C for 3 hours and then 890°C for 45 minutes proved
efficacious and 890°C is almost surely inferior to the temperature at which
the biscuit has been originally fired [11] and should not lead to the formation
of new chemical species in the ceramic body.
Small patches of minimal orange skin were occasionally reported but they
did not led to unacceptable results. However, one sample returned
completely unexpected results since the new glaze did not adhere to the
biscuit and the original glaze now curled at the edges (Figure 6f). The
problem in this case is not related to calcite decomposition during the firing
and has not been explained so-far.

a) before restoration; b) after restoration by re-firing;

c) before restoration; d) after restoration by re-firing;
Figure 8. Application of double-stage re-glazing and re-firing technique to the restoration
of spalled tiles

e) before restoration; f) after restoration by re-firing

Figure 8. Application of double-stage re-glazing and re-firing technique to the restoration
of spalled tiles

Through the work the most relevant problem that had to be addressed was
the fact that the ceramic body of an old tile removed from a wall now
contains non-original/contaminant chemical species that have been
accumulated with time in the porous network or products of alteration that
release gases upon re-firing. The understanding of the cause of the bubbling
and the development of a technique to address this obstacle were achieved
through this project.

Each case is a case in itself and although a receipt was developed that may
be applied successfully to most cases, there was a definite failure. It is
recommended that, if possible, preliminary tests and trials should be done to
determine the content in carbonates of the samples and, in all cases,
preliminary firings be run to check the response of the azulejos to different
firing cycles so as to determine the most appropriate and efficient one.

Acknowledgements
FCT funding support is acknowledged through the project CerAzul (PTDC/CTM-
CER/119085/2010).

Bibliographic references

BARROS, V., A.J. & Almasqué, I., Azulejos de fachada em Lisboa, Câmara Municipal de
Lisboa, Lisbon, 1989;
2PERREIRA, S.; Esteves, L.; Mendes, M.T.; Musacchi, J.; Rodrigues J.D.; Mimoso J.-M., CerAzul:
Assessment and development of materials and techniques for the conservation of historic azulejos, In
Azulejar 2012, Aveiro, 2012;
3MIMOSO, J-M.; Pereira, S.; Esteves, L.; Santos Silva, A., A tour of façade azulejos decay in
Lisbon, in Proc. Int. Congress Azulejar, Aveiro, Portugal, Oct. 10-11, 2012;
4MOURA Ferreira, I., Azulejos tradicionais de fachada em Ovar: contributos para uma
metodologia de conservação e restauro, Câmara Municipal de Ovar, Portugal, 2009;
5 RIBEIRO, L.; Ferreira, L. M; Coroado, J.; Velosa, A. L. (2012), Conservación y restauración del
azulejo de Ovar – metodología para una intervención sostenible, In Qualicer 2012, Castellón, Spain,
2012;
6 MIMOSO, J-M, Cesare Brandi’s "Theory of Restoration" and azulejos, in Proc. Int. Sem.

Conservation of Glazed Tiles, Lisbon, LNEC April 15-16, 2009;

7 MIMOSO, J-M; Pereira, S; Esteves, L., A matter of strength - measuring the glaze/biscuit adherence
in Portuguese azulejos, in Proc. Int. Congress Azulejar, Aveiro, Portugal, Oct. 10-11, 2012;
8CHABAN, A, Restoration of semi-industrial glazed ceramic tiles by re-firing, Master’s thesis
Dissertation, discussed at the University of Bologna, School of Sciences, March 19, 2015;
 9 MIMOSO, J-M., New stuff about old stuff: the azulejos of the Roseira Factory, lecture given on Jan
14, 2015 at the ARTIS Institute of the University of Lisbon,
https://blogazlab.wordpress.com/2014/12/19/coisas-novas-acerca-de-coisas-velhas-os-azulejos-de-
fachada-da-fabrica-roseira-de-lisboa/;
10Fundação Ricardo do Espirito Santo Silva, Festa barroca a azul e branco: exposição - conservação
e restauro, Lisbon: FRESS, 2003;
11 PEREIRA, S; Mimoso, J-M; Santos Silva, A, Physical-chemical characterization of historic

Portuguese tiles, Relatório LNEC 23/2011, Lisbon, LNEC, 2011.

View publication stats