1 

DECAY OF THE RESTORATION RENDER MORTAR OF THE 

CHURCH OF SAN MANUEL AND SAN BENITO, MADRID, SPAIN: 

RESULTS FROM OPTICAL AND ELECTRON MICROSCOPY 

 

 
M.J. Varasa, M. Alvarez de Buergob, E. Perez-Monserratb, R. Fortb  

a. Departamento de Petrología. Facultad de Ciencias Geológicas. Universidad 

Complutense de Madrid, C/ José Antonio Nováis 2, 28040 Madrid. SPAIN 

mjvaras@geo.ucm.es 

b. Instituto de Geología Económica (CSIC-UCM), Facultad de Ciencias Geológicas, 

Universidad Complutense de Madrid, C/ José Antonio Nováis 2, 28040 Madrid. 

SPAIN alvarezm@geo.ucm.es   empmon@geo.ucm.es   rafort@geo.ucm.es 

 
Abstract 

 

The main aim of this research is to determine the reasons for the detachment 

of a mortar rendering applied to the stone masonry of the Church of San Manuel and 

San Benito (1903-1910), Madrid, Spain. Two building materials were analysed using 

optical and electron microscopy: the original limestone dressing affected by intense 

flaking, and the restoration render mortar used in a relatively recent rehabilitation to 

´consolidate´ the decayed stone. Results obtained by microscopic techniques and 

from analysis of the environmental conditions to which the building is exposed made 

it possible to conclude that 1) the mortar is a good material and compatible with the 

stone, 2) the render mortar failure was mainly due to its incorrect application and the 

unsuitable repair technique chosen, and last, 3) the materials decay was worsened by 

the adverse exposure conditions of the upper section of the building (pounding rain 

and sun radiation). 

 

Keywords: restoration render mortar, stone decay, exposure conditions, optical 

microscopy, SEM 

 



 
2 

1. Introduction 

 

 The Church of San Manuel y San Benito (Madrid, Spain, Fig. 1) was built 

between 1903 and 1910 by the prestigious architect Fernando Arbós y Tremanti. The 

Neo-Byzantine structure, an unusual style for Madrid, is characterised by the great 

variety of natural building materials used, brought from different places around 

Spain by the country’s newly constructed railway network [1]. The building, with an 

almost Latin-cross floor-plan, is surrounded by eight slender ornamented towers and 

finished by a copper dome supported by nearly circular drum (16 sides) (Fig. 2). The 

church has been used for worship, except during the Spanish Civil War (1936-1939), 

when it saw service as a military headquarters. It currently houses the provincial seat 

of the Augustinian Order in Spain. The property was declared of Cultural Interest in 

1982.  

Building materials include: granite, found in the socles of the façades and in 

the grating that surrounds the building; brickwork walls with a external limestone 

dressing 13 cm thick; and marble, found in balusters and columns of the windows 

(white marble), and in some decorative panels (white, black, red marble). The 

interior of the church is dressed in Venetian mosaic, and the dome is copper finished.    

Since 1975, the building has undergone several restorations. The last (1990-

1995) consisted mainly in consolidating the severely decayed limestone external 

façades by covering them with a restoration render mortar, followed by the 

application of a paint and a chemical-protective surface treatment. At the present 

time, this mortar has severe stability problems and has fallen off in some sections 

(Fig. 3). 

The decay of the limestone is more severe in the upper parts of the building, 

mainly around the drum, terrace area and small towers. The stone surface is 



 
3 

weakened due to intense flaking and spalling, which have led to surface stone 

fragments becoming detached (Fig. 4). This deterioration process has penetrated to 

5-6 cm beneath the stone surface.  

Restoration render mortar was used to consolidate the severe decayed stone 

and to rebuild decorative and architectural components such as mouldings, imposts, 

cornices, etc., affected by significant volume loss. Between one and six layers of 

mortar were applied, depending on the amount of stone surface loss in need of 

rebuilding; its thickness ranges from 4 mm to 3 cm (Fig. 5). To the naked eye it 

appears very porous, with circular-shaped voids. This porosity defines the different 

layers in the mortar. The render mortar was directly placed on the stone surface, 

without any steel net or anchorages to support its weight. Both mortar and stone are 

finished with a yellowish paint and a protective treatment. In many cases, mortar 

panels or plaques were placed following and coinciding with the limestone dressing 

joints. 

At present, the restoration render mortar is extremely degraded in the drum 

area, with bulging and fissuring that sometimes lead to the complete detachment of 

mortar fragments of significant volume (Fig. 3). These breakages are severe in the 

upper areas oriented W, SW, S and SE (Fig. 2) and are mainly located under the 

windows level of the drum (Fig. 6) and at the slope change with the roofing of the 

side chapels (grader area, Fig. 3). 

Environmental conditions play an important role as the more deteriorated 

upper zones coincide with the zones which are more exposed to sun radiation (S and 

SW) and become significantly hotter than environmental temperatures (Fig 2).  

The main aim pursued with this research is to determine the causes that have 

led to the failure of the restoration or consolidation of the external walls of the 

Church of San Manuel y San Benito, leading to the destabilisation and detachment of 



 
4 

large fragments of mortar renderings applied to the stone masonry in the previous 

restoration project. To that aim, the materials, the deterioration processes, the causes 

of decay (especially the influence of the microclimatic conditions), and the design 

and execution of the previous rehabilitation have been studied. 

In general, the restoration of the church of San Manuel y San Benito 

involved many different aspects, including recent decay and decay of previous 

restorations. This study aims to provide data on original and repair materials in order 

to better specify future repair materials and application methods. 

 

2. Materials and methods 

Environmental conditions play an important role in the decay processes of 

this case study. The most decayed areas of the building are the highest ones, around 

the church drum, which are the most exposed to sun radiation, pounding rain and 

wind. Those that are worst preserved are oriented towards the SW-S-SE and the S 

(Fig 2), corresponding to the areas most exposed to sun radiation. To analyse the 

effect of microclimatic conditions on the materials, temperature and humidity data 

loggers were installed. Data were collected from both the environment (air 

temperature and relative humidity) and directly from the materials´ surfaces (stone 

surface temperature; temperature and humidity content in the stone-render mortar 

contact). These data loggers were placed in three areas with different orientations 

during winter (from 13 December 2003 to 26 January 2004) and summer season ( 

from 3 June 2004 to 3 August 2004). The equipment used was: 

- Onset Hobo Temperature Data Logger (1996, Computer Corp. S/N 

140470), with a standard temperature range of -20°C to +70°C for the logger and of 

-40°C to +120°C for the sensor; sampling interval ranged from 0.5 seconds to 9.0 h 

hours, and an accuracy of 0.2ºC-0.4ºC.  



 
5 

- Onset Hobo Relative Humidity Data Logger (1993, Computer Corp. S/N 

111466), with a standard relative humidity range of 25 to 95%, and an accuracy of 

±5%.  

Both Hobo Data Loggers had a capacity of 1800 measurements readings, 

and a BoxCar 2.06 software was used for data acquisition and processing.  

Wind data (dominant bearing and speed) were provided by station #8 of the 

weather and pollution stations of the Surveillance Automatic Network (air quality) 

of the City Council, Madrid, which is located two hundred metres from the building 

(from 1 January 2003 to 31 July 2004). 

 

This study mainly sampled two types of materials: the limestone dressing 

the brickwork structure of the church, and the mortar rendering used to repair the 

decayed stone in an earlier restoration dated 1990-1995. The paint applied on the 

stone in the rehabilitation was also sampled, together with a chemical-protective 

surface treatment visible with microscopic techniques but not to the naked eye. 

Sampling was carried out by taking flakes (paint), small fragments (stone, mortar) 

and some mortar-stone cores from earlier adherence tests in the drum area. Artefacts 

were reduced to the minimum during the sampling process following 

recommendations suggested by [2] to avoid unnecessary pressure during their 

extraction (no core drilling was used, either hammer or chisel).  

Of a total of 26 samples (stone, render mortar and stone+render), 14 were 

taken from the SW-oriented façade, 8 from NW, 2 from NE and 2 from SE. Most of 

the samples were taken from the upper and most deteriorated section of the building 

(drum and surroundings). For the analysis of the deterioration of both materials, 

render mortar and stone substratum, samples from altered and unaltered areas were 

taken in order to compare the results obtained.  



 
6 

Although a wide range of analytical techniques exist that can give 

information on this specific type of study, microscopic techniques are very useful 

and representative [3].  

Thin section analysis was done in an Olympus BX 51 petrographic 

microscope using polarized light technique. Images were acquired using and 

Olympus DP 12 (6V/2,5Å) digital camera and proprietary software. Magnifications 

of the images were ×4 and ×10 (1 mm and 500 µm scale bar in the micrograph, 

respectively). For the thin sections preparation [4,5], hand specimens were used. 

When samples were not sufficiently consolidated (mostly stone and stone-mortar 

samples), they were vacuum impregnated with an epoxy resin. Sections of these 

samples were polished (the face to be observed), and the polished face was 

fixed/glued to a glass slide. The section was then cut again up to 2 mm thick and 

further reduced on a thin section grinder to a minimum thickness of 50-100 µm 

(according to the different interference colours of the existing minerals). The next 

step was the polishing of the thin section with silicon carbide (for polarising 

microscope) and alumina (for SEM) of different grain sizes to obtain section 

thickness as low as 30 µm.   

Alizarin Red S staining was used to differentiate calcite from dolomite [6-8].  

Petrographic description of thin sections was performed according to 

Tucker´s descriptive worksheets [9].  

The Scanning Electron Microscope (SEM) used was a JEOL JSM 6400, 0.2 

to 40 kV acceleration voltage, a beam current of 6 x 10-10 A, 10-5 Torr vacuum 

conditions, 35 Å SEM resolution, at a working distance of 8 mm and 35 kV. Images 

were acquired at 20 kV of acceleration voltage. A coupled X ray energy dispersive 

spectrometer (EDS) was used (OXFORD-INCA), with a nominal resolution of 133 



 
7 

eV at 5.39 kV. Samples were graphite sputtered (15 nm thick cover) using a Balzers 

Med010 sputter coater (Balzers, Liechtenstein).  

Both fragments (around 12 × 12 × 12 mm) and polished, thin sections were 

studied by SEM-EDS, in secondary (SE) and backscattered electrons (BSE) mode, 

respectively. Morphology, structure and texture of the different constituents of the 

sample were analysed in SE mode (surface rugosity and relief), while BSE mode 

allowed the determination of the elemental composition of the constituents based on 

their atomic number.  

Cross sections of the samples were analysed: perpendicular to the external 

surface of both render mortar and stone. 

Porosity from both mortar and stone samples was determined through 

mercury intrusion porosimeter – MIP- (Micromeritics Autopore IV 9500). This 

automatic instrument measures pore diameters from 0.0055 to 360 µm, with 

measurement conditions from atmospheric pressure to 33,000 psia (228 MPa). 

Samples cores of 10 mm and at least 30 mm tall were length.  

 

 3. Results 

 

3.1. Environmental measurements 

The environmental temperature in the southern areas of the building can 

reach a maximum of 49 ºC during summer, and there are daily temperature 

variations over 33ºC. On these days, the mortar surface can rise to maximum 

temperatures of 54ºC and undergo thermal variations over 39ºC. At the mortar-stone 

contact, the temperatures and the thermal variations are similar to those of the 

environment, although the humidity is slightly higher (31% in summer) than the 

environmental humidity (29%). Although the wind is stable (8 kph-1), the dominant 



 
8 

bearing varies between 220º (SSW) and 130º (ESE). The wind directly affects the 

more deteriorated area of the drum as on rainy days these renderings are soaked by 

pounding rain. Another factor to be taken into account is the fact that the wind 

bearing is highly changeable throughout the day, varying between 180º (S) and 250º 

(WSW) during the time of highest environmental temperature (1 p.m. – 6 p.m.). 

 

3.2. Stone substratum 

Under the polarising microscope (OM), the limestone is defined as a 

biomicrite [10], due to the high percentage of bioclasts. The distribution of such 

percentage is the following: 60vol% of bioclasts (foraminifer, echinoderms, 

bryozoan, mollusca and coralline algae), 10vol% of siliciclastic particles (quartz), 

20-25vol% of micritic matrix and 5-10vol% of pores (Fig. 7). Some neogenic quartz 

is observed in pores. The total porosity of the stone obtained by MIP is 16.2% 

Macroscopically visible flaking continues at microscopic scale, leading to 

delamination and increased permeability (Fig. 7a). Fissure width varies from 0.37 

mm at depth, to 0.9 mm near the surface. This fissuring process affects all the stone 

constituents (bioclasts, matrix and cement). 

The stone is not always covered by the mortar rendering but by a yellowish 

paint. It is a homogeneous, compact, dark-coloured layer that covers the 

irregularities of the stone surface and has a thickness ranging from 0.1 to 0.15 mm 

(Fig. 7b).  

The SEM-EDS study of this calcareous substratum confirms what was 

observed under the OM. The chemical composition detected is mainly Ca, followed 

by Si. SEM also made it possible to detect the micro-fissures on the stone surface 

(Fig. 8). The paint layer, which covers the stone, appears as a poorly consolidated 

layer with significant porosity in its interior and in contact with the underlying stony 



 
9 

substratum. Its chemical composition is mainly constituted by Ti and smaller 

amounts of Fe and Si. In addition, there are remains of a thin, superficial film, 

containing Si, Mg, Al and Ca. At the contact between paint and substratum, porosity 

is higher. 

 

3.3. Restoration render mortar  

With the OM it is possible to recognise dolomitic marble aggregates floating 

within a micro-crystalline mass with a massive aspect and undefined mineralogy. 

These aggregates are very angular and have a well-defined grain-sized selection 

consisting of three dominant types: 0.1, 0.3 and 1 mm. The morphology of the air 

voids is very regular and circular. The voids are located within the micro-crystalline 

matrix (Fig. 9a). Porosity measured by MIP is 22.2%.  

As mentioned earlier, the rendering may have a single layer, although multi-

layer rendering is more frequent. In the latter case (Fig. 9b), the successive layers 

which appear are very similar to each other but differ slightly in colouring and 

thickness, as well as in the amount and morphology of the voids. The more 

superficial layers are thinner and more compact. The aggregate is in contact and its 

main axis, together with the axis of the voids, is oriented parallel to the surface of the 

mortar. The lower, thicker layers are not as compact as the upper layers; therefore, 

the aggregate is very loose and does not appear in contact with each other. No type 

of orientation of the main axis can be observed.  

The stacking of several layers of mortar is marked by a dark-coloured film, 

which may be due to a lower porosity than the surrounding material, also linked to 

the appearance of voids on this film oriented with lengthened morphology (Fig. 9b). 

The contact between the successive layers is straight but discontinuous due to the 



 
10 

substantial porosity observed in these areas. This can lead to the selective 

detachment of the layers of mortar. 

A thin film of yellow paint always appears on this mortar, giving the surface 

a uniform appearance (Fig. 10a). In thin section, the yellow topcoat appears as dark 

colored layer 150-500 µm thick, covering the irregular mortar surface. Bonding of 

the topcoat is incomplete, as demonstrated by a bead-chain of air voids on the 

interface with the substratum. Also, there is a much more substantial porosity which 

separates the layer of paint on the underlying rendering mortar and is particularly 

higher in the superficial part of this mortar. 

The rendering mortar adheres directly to the stone, which has a high degree 

of deterioration with superficial cracking (Fig 10b). Voids with lengthened 

morphology, made up of the coalescence of previous underlying rounded voids, are 

observed. Fundamentally, these voids appear at the base of the rendering in contact 

with the stone. Cracking, horizontal and parallel with the surface of the stone, affects 

the base of the rendering.  

The analysis of the single layer and multi-layer mortar with the SEM+EDS, 

confirms that the aggregate is dolomite (Ca-Mg carbonate) and the binder, whose 

mineralogy was not detected with the OM, is gypsum, confirmed by both the EDS 

spectrum indicating the presence of Ca and S combined with crystals morphology 

(dense mass of needle and arrow shaped crystals). This mass contains isolated 

dolomite crystals which are larger and have more irregular shapes. Pigment and 

stabilizing agents occur interstitially in the gypsum fabric (Fig. 11). In both cases, 

compounds containing, on one hand, Si, Al and Mg have been detected in specific 

areas, while, on the other hand, a uniform distribution of the Ti and smaller amounts 

of Fe have been seen (Fig. 12). They are detected in the area of contact with the 

underlying stone substratum (Fig. 8). In the case of the multilayer mortar, they 



 
11 

appear in the contact zone of the layers of mortar. When there are two layers, these 

chemical composites (mainly containing Si, Al and Mg) impregnate the surface of 

the lower layer and especially affect all the binder (Fig. 12). In this area of contact 

between layers, the line of lengthened voids previously detected with the OM again 

appears and affects the base of the layer which is applied last (Fig. 12). These 

impregnations correspond to the extended dark stains which appear in the OM and 

which make it possible to differentiate the layers of mortar (Fig. 9).  

The layer of paint which covers the surface of the rendering mortar was 

studied in greater detail through the SEM+EDS. This layer has a variable thickness 

(50-100 µm), a granulated appearance and a Ti dominant chemical composition, 

containing also Fe, Si, Al and Ca traces. A thin film (20-30 µm) containing Si, Al 

and Mg covers its surface. The lower interface of the paint layer is quite irregular as 

it adapts to the irregularities of the surface of the mortar. Substantial porosity was 

detected in the paint mortar contact area (Fig. 13).  

 

4. Discussion 

 

In theory, deteriorated walls can be consolidated by covering them with one 

or several thin layers of fresh mortar composed of a specific binder, small-sized 

aggregate and added colouring pigments (mainly containing Ti and Fe). Primer coat 

impregnations were applied on the mortar layers interfaces to improve mortar 

properties (adherence, rigidity, etc). The wall is previously reinforced with a re-

habilitating steel net in order to provide support for the weight of this rendering [11]. 

This type of intervention leads to a relative improvement in the strength and stiffness 

of the wall. These actions can be negative if the mortar is not well-designed and, 

especially, if the execution of the repair technique is incorrect [12,13]. The 



 
12 

mechanical performance of a wall consolidated with this technique will depend on 

the distribution and the amount of rendering mortar used and on the number and 

thickness of the layers of mortar applied. Both these parameters determine the 

maximum load possible (maximum tolerable weight). However, the effect of this 

reinforcing technique relies on the bond strength at the interface of the wall 

substratum with the mortar render [12, 13]. 

One important question not addressed when the intervention technique was 

designed was why deterioration was located in the upper areas of the W, SW, S and 

SE of the building and affected all the natural construction material in these zones. 

As an in-depth study was not carried out on these forms of deterioration, the 

problems reoccurred in the same zones and affected the new materials (rendering 

mortars) applied during the 1990-1995 rehabilitation. 

In short, these deterioration processes are influenced by a number of 

environmental and material factors, which together affect all the construction 

materials. As the materials are quite porous, continuous exposure magnifies the 

deleterious effects of weather in this area. The construction materials located in the 

W, SW, S and SE undergo severe deterioration as they coincide with the zones of 

highest temperatures, strong thermal variations, prevailing winds and pounding rain. 

Under these conditions, the original stone and the artificial stone used in the 

restoration deteriorate more quickly. 

Furthermore, the types of deterioration which affected the original stone 

were not taken into account. Therefore, the new rendering mortar was applied to an 

original stone which had not been restored (i.e. removing the decayed surface or 

consolidating it). This led to a very weak area between the layer of mortar and the 

original healthy stone, as confirmed by microscopic techniques. The weakened zone 



 
13 

of natural stone is intensely cracked, more porous and often comes down when the 

mortar covering becomes detached.  

Consequently, due to the prevailing environmental conditions in the zone 

and the poor state of the supporting stone, this mortar did not respond properly and 

underwent cracking, bulging, collapse and detachment. The strong rigidity and 

resistance provided by this rendering were overcome by the continuous episodes of 

hydric and thermal expansion experienced since 1995. The combined action of 

dampness in the mortar and high temperatures caused the mortar to undergo thermal 

and hydric expansion, which, although not intense, were sufficient to generate 

stresses which led to processes of deformation and displacement of the mortar 

rendering. 

In addition, although the rendering mortar used (dolomite marble aggregate 

and gypsum binder) is material of acceptable quality as regards petrographic, 

mineralogical and petrophysical characteristics, the execution was not properly 

carried out: 

a) This mortar was applied in one or several layers with a thickness which 

varied from 4 mm to 3 cm. The 3 cm thickness brings substantial weight to bear on 

the mortar, adhering directly to the original altered stone with no kind of mesh or 

steel or fibreglass anchoring to enable secure fixing. In areas of maximum thickness, 

the surface of the altered stone does not support the weight of the mortar, and the 

rendering mortar together with the most superficial part of the supporting stone 

become detached. This detachment of the part of the stone stuck to the rendering 

mortar is attributable to the high degree of adherence and rigidity of the mortar-stone 

interface, probably due to the application of primer coats in such interface (detected 

in the dark layers containing Si, Al and Mg by means of EDS analysis). That is why 



 
14 

the detachment, mainly due to continuous episodes of hydric and thermal expansion, 

does not occur on the interface between the two materials, but rather in the original 

stone substratum weakened by alteration.  

b) In many cases, the design of the rendering follows the division of the 

stone support, thus, the coincidence of the joints with the areas where the mortar is 

thinner (joints of the facing panels) implies the presence of stress. This stress leads 

to the appearance of horizontal and vertical cracks through which rainwater 

infiltrates. The internal dampness of the mortar, together with high temperatures, 

cause expansion which leads to the bulging of the rendering in the central area, the 

complete detachment of the stone support, superficial cracking of the most deformed 

area of the rendering and, finally, the detachment of rendering plaques which leaves 

the stone uncovered. 

c) Microscopic techniques show the presence of primer coats applied in the 

stone-mortar and mortar-mortar interfaces and the existence of substantial porosity 

(entrapped-air) have prejudiced the proper conservation of the material. This was the 

working method used for the mortar rendering of the stone wall in the rehabilitation 

of this building.  

Two weak factors have been detected at the base of this mortar and these 

could lead to its detachment. First, the existence of the lengthened voids at the base 

of the rendering in contact with the stone, due to the large number of air bubbles 

which were trapped during the application of the mortar on the stone (entrapped air). 

The second factor is cracking, horizontal and parallel with the surface of the stone, 

which affects the base of the rendering. The cracks are related to the peeling which 

appears on all the ashlars. Under these conditions, the rendering mortar on this stone 

substratum becomes structurally and texturally unstable. This leads to serious 

weakening of the stone surface and the mortar base, which can entail materials loss 



 
15 

due to the detachment of the rendering mortar and the superficial part of the stone. 

These observations verified that no measures had been taken to restore the stone 

before the rendering mortar was applied. Regarding the paint on the mortar, its 

external section is also affected by flaking, with a weakened area in contact with the 

stone due to the dissolution processes of the stone constituents. This section leads to 

stone breakage and the paint layer breaking off (Fig. 7b, 10a).  

With respect of environmental conditions affecting the materials decay, 

rainwater penetrating through cracks in some parts of the rendering dampens the 

interior of the mortar. The spread of this dampness throughout the interior of the 

mortar is the result of a huge porosity network, located between the contact areas of 

the layers of mortar and between the mortar and the surface of the supporting stone. 

This porosity is due to the entrapped air inside the layers of fresh mortar during its 

application. 

Furthermore, this dampness is retained inside the mortar by a number of 

impregnable barriers (primer coats). These barriers impede the evaporation of the 

humidity from the deepest layers. Thus, the humidity retained increases in relation to 

the thickness of the rendering and increases its structural destabilisation.  

Consequently, the excess weight (the humidity retained in the interior of the 

mortar plus the substantial thickness of the mortar) which must be borne by a very 

deteriorated stone surface, together with hostile environmental conditions 

(principally the high temperatures which can be reached on the surface of the mortar 

itself, with temperatures above those of the environment.), conditioned the hydric 

and thermal expansion processes which affected the exterior rendering from 1990-

95. These deterioration processes have led to bulging, cracking, collapse and 

detachment of the rendering mortar since 1995.  



 
16 

 

5. Conclusions 

- The most decayed area of the building is the Southern façade (from W to SE) of the 

circular drum. The original stone, as well as the mortar render used in the 

rehabilitation project, deteriorate here more quickly due to the fact that this zone 

undergoes the highest temperatures, strong thermal variations, prevailing winds and 

pounding rain. The existing fissures in the mortar render favour rainwater infiltration 

and expansion processes, with following bulging, cracking and plaques detachment.  

- The masonry limestone is a biomicrite (porosity of 16.2%). Its macroscopically 

visible flaking continues at microscopic scale. The stone was rendered to consolidate 

it and to stop its severe decay. When it is not covered by this mortar render, it is 

finished with a yellowish paint.  

- The mortar render adheres directly to the decayed stone with no anchorage system; 

it is a gypsum mortar with marble aggregates (dolomite) colouring pigments 

(containing Fe), and a porosity of 22.2% due to the great amount of air voids. The 

render is usually a multi-layer mortar with a thickness ranging from 4 mm to 3 cm; 

along the mortar layers interfaces there is an alignment of oriented air voids that 

creates weakness planes (possible detachment planes), as well as dark areas that 

correspond to primer coat applications, which also favour detachment through the 

underneath decayed stone. A thin film of yellow paint always appears on top of the 

mortar render (Ti and Fe composition), with a high number of air voids in the paint-

mortar and paint-stone interfaces. A thin film attributable to a finishing protective 

impregnation (water-repellent) covers the paint layer.   

From the study carried out mainly using microscopic techniques, it can be 

concluded that the failure of the rehabilitation of the Church of San Manuel y San 



 
17 

Benito was due not only to the flawed design of the operation but also to the faulty 

execution of the repair technique chosen (fresh rendering mortar). 

To sum up, although the mortar is of good quality and is compatible with the  

stone support, it has performed poorly due to the inadequate treatment and 

preparation of the stone support (not removing the decayed surface or consolidating 

it before the mortar application), and adverse environmental conditions. The damage 

directly affects the modern mortar (also the decayed stone surface coming down with 

the detached of the covering mortar), and the general state of the building, and 

deterioration has generally worsened since the rehabilitation in 1995, with bulging, 

cracking, collapse and detachment of the rendering mortar since then.  

 

Acknowledgements 

This study has been financed by a research contract with the Parish-Church 

of San Manuel y San Benito, Madrid (Augustinian Order), and by a “Ramon y Cajal” 

contract (MAdB) (Ministry of Education and Science).  

Thanks are given to the “Luis Bru” Electron Microscopy Research Support 

Centre of the Universidad Complutense de Madrid (UCM), and also to the Linguistic 

Office of the Centre of Modern Languages of the Complutense University of 

Madrid, for the proofreading and revising of the paper. 

 



 
18 

References 

[1] Gomez-Heras M, Fort R. Location of quarries of non-traditional stony materials 

in the architecture of Madrid: The crypt of Santa María la Real de la 

Almudena. Materiales de Construccion 2004;54:31-47.  

[2] Crumbie A.K. SEM Microstructural studies of cementitious materials: sample 

preparation of polished sections and microstructural observations with 

backscattered images. Artefacts and practical considerations. Proc 23rd Int 

Conf on Cement Microscopy, Albuquerque, NM, April 29 - May 3; 2001. 

[3] Hughes JJ, Cuthbert SJ. The petrography and microstructure of medieval lime 

mortars from the west of Scotland: Implications for the formulation of 

repair and replacement mortars. Materials and Structures 2000;33:594-600.  

[4] Humphries DW. The preparation of thin sections of rocks, minerals and 

ceramics. Oxford, UK: Royal Microscopical Society, Oxford Science 

Publications, Microscopy Handbooks 24; 1992, 83 p. 

[5] Ques J, Tritlla J. Introducción a las técnicas de preparación de láminas delgadas. 

In: Melgarejo JC, editor. Atlas de Asociaciones Minerales en Lámina 

Delgada. Publicaciones de la Universidad de Barcelona & Fundación 

Folch. Barcelona; 1997, p. 33-7. 

[6] Friedman GM. Identification of carbonate minerals by staining methods. Journal 

of Sedimentary Petrology 1959;29:87-97. 

[7] Evamy B.D. The application of a chemical staining technique to the study of 

dedolomitisation. Sedimentology 1963;2:164-170. 

[8] Miller J, editor. Microscopical techniques: I. Slices, slides, stains and peels. In: 

Tucker ME, editor. Techniques in Sedimentology. Oxford: Blackwell 

Scientific Publications; 1988, p. 86-107. 



 
19 

 [9] Tucker ME, editor Sedimentary Petrology: an introduction to the origin of 

sedimentary rocks. 3rd ed. Oxford: Blackwell Scientific Publications; 2001: 

11-20, 110-114. 

[10] Folk R.L. Spectral subdivision of limestone types. In: Ham WE, editor. 

Classification of Carbonate Rocks. American Association of Petroleum 

Geologists; 1962, 62-85. 

[11] Valluzi RM, Binda L, Modena C. Experimental and analytical studies for the 

choice of repair techniques applied to historic buildings. Materials and 

Structures 2002;35:285-292. 

[12] Ashurst J, Ashurst N. Mortars, Plasters & Renders. Practical Building 

Conservation Vol 3. English Heritage Technical Handbook. Aldershot, 

Hampshire, United Kingdom: Gower Technical Press; 1989.  

[13] Binda L, Gambarotta L, Lagomarsino S, Modena C, Segnalini O. Montesanto e 

Roccanolfi centri storici a rischio (available only in Italian). Bollettino 

Italia Nostra 1999;355:5-8. 



 
20 

 
Fig. 1. General view of the building. South façade. 

 



 
21 

 

Fig. 2. Ground plan of the building. 

 

 

 

Fig. 3. The fissure pattern (a) in the rendering mortar follows the original stonework 
joints. The increasing bulging that results on mortar cracks (b) and, later on, on 
themortar detachment (c).  

 



 
22 

 

Fig. 4. The stone under the rendering mortar is severestronglyly decayed. The most 
frequent deterioration forms are flaking and spalling, involving a great amount of 
surface volume loss. 

 

 



 
23 

 

Fig. 5. Multilayered mortar applied on the surface of the decayed stone. The 
interface among the different layers is characterized by the presence of a high 
number of air voids. 

 

 

Fig. 6. The mortar applied on the surface of the decayed stone has been used to 
reproduce many architectural and ornamental elements that were lost. 



 
24 

 

 

 

Fig. 7. Micrographs under OM (plane polarized light, PPL) of the stone. a) The 
presence of fissures parallel to the external surface is the responsible for the flaking 
of the stone surface. c) Weakness area (B; high porosity) under the paint layer (A) 
that may occur directly on the stone (C). 



 
25 

 

Fig. 8. SEM micrographs (secondary electrons mode) (a) and backscattered mode (b) 
of a polished section of the contact between the highly fissured limestone (right) and 
the rendering mortar (left). The stone contains Ca and Si. The mortar is constituted 
by a dolomitic aggregate (Ca and Mg) and a binder (A –point analysis) composed of 
gypsum (Ca and S), prime coat (Si, Al and Mg) and colouring pigments (Fe). 



 
26 

 

 

Fig. 9. Micrographs under OM (plane polarized light, PPL) of the rendering mortar. 
a) General appearance of the mortar (A-dolomitic aggregate, B-binder, C-air voids). 
b) The different layers of the mortar are limited by the presence of elongated voids 
(D) and by a dark layer (E). 



 
27 

 

 

Fig. 10. Micrographs under OM (plane polarized light, PPL) of: a) The paint layer 
(A) that is also on the modern rendering mortar (C). There is a weakness area in the 
interface between the paint layer and the most external section of the mortar (B; high 
porosity). b) The contact area between the strongly fissured natural stone (B) and the 
rendering mortar (A). There are elongated voids along this contact plane (C).  



 
28 

 

Fig. 11. SEM micrographs (secondary electrons mode) and EDS spectrum of a 
fragment of the rendering mortar. The gypsum crystals present growing enlargement 
due to the impregnation with prime coats (Si, Al and Mg) and to the pigments (Ti 
and Fe) added to the binder. Point analysis (A). 



 
29 

 

Fig. 12. SEM micrographs (secondary electrons mode) of a fragment of the 
rendering mortar. A-Gypsum crystals without prime impregnation (mortar, upper 
layer), B-Gypsum crystals with prime impregnation (mortar, lower layer) and C-Air 
voids in between both layers. 

 



 
30 

 

Fig. 13. SEM micrographs (secondary electrons mode) of a fragment of paint, 
surface layer. A-Film corresponding to a water-repellent treatment, B-Paint layer, C-
Voids and D-Rendering mortar.