CAPILLARITY TESTS ON HISTORIC MORTAR SAMPLES EXTRACTED FROM SITE. METHODOLOGY AND COMPARED RESULTS

Transcription

1 13 t International Brick and Block Masonry Conference Amsterdam, July 4-7, 2004 CAPILLARITY TESTS ON HISTORIC MORTAR SAMPLES EXTRACTED FROM SITE. METHODOLOGY AND COMPARED RESULTS M. Rosário Veiga 1 ; Ana Cristian Magalães 2 ; Violeta Bokan-Bosilikov 3 Abstract Renders and masonry mortars are very exposed to aggressive agents so tey are often te object of repair interventions on monuments and ancient buildings. Te substitution of ancient lime mortars by new mortars, wit different caracteristics, as frequently been te cause of accelerated degradation of old masonry walls because of te cange on teir pysical beaviour. Te knowledge of pysical caracteristics of ancient mortars is terefore very important to select repair mortars and to plan adequately te intervention. Tere is a need to define caracterization test metods applicable to irregular, friable samples to assess tose caracteristics. In te present paper a test metod is described to evaluate water absorption capillarity and some results from Portuguese and Slovenian old mortars are presented and discussed. A comparison wit results obtained for new mortars, using te same metod and also te metod proposed by EN standards elps to extract conclusions of te metod s possibilities. Key Words Conservation, old mortar, test, capillarity. 1 Introduction Interventions on ancient buildings often imply repair or substitution of renders. In tose cases it is very important to use materials compatible wit te old ones. Te knowledge about old mortars caracteristics is necessary to define te caracteristics to accept for repair and substitution materials. Neverteless, pysical and mecanical caracteristics are sometimes difficult to assess, because render samples tat can be extracted from works are usually small, irregular and sometimes wit a low coesion. Terefore, te direct application of existent standardised tests is seldom possible. 1 Laboratório Nacional de Engenaria Civil, Lisbon, Portugal; rveiga@lnec.pt 2 Laboratório Nacional de Engenaria Civil, Lisbon, Portugal; acristian@lnec.pt 3 University of Ljubljana, Faculty of Civil and Geodetic Engineering, Slovenia, vbokan@fgg.uni-lj.si page 1 1

2 To obviate tis problem some special test metods ave to be studied and adapted to old mortar samples collected from site. Water beaviour is particularly relevant for te wole performance of renders applied on old masonry. In fact renders and repointing mortars must protect te masonry against water, preventing te easy entry of significant volumes of water and promoting teir quick drying (VEIGA 2000, VEIGA et al 2001). Besides, water is a major degradation factor for old mortars, because i) it as some capacity of solving mortars constituents ii) it facilitates reactions tat promote te formation of weaker elements iii) it increases volume wen it turns into ice iv) it solves salts and promotes teir movement troug te mortar v) it facilitates fixation of biologic colonization. Te study of old mortars caracteristics concerning water, namely water absorption and drying beaviour, opens te possibility of trying to formulate new mortars wit similar caracteristics, consequently wit similar expected durability and adequate to protect te same masonry. A test metod was prepared and experimented at LNEC and at te University of Ljubljana to assess te capillarity absorption coefficient of old mortar samples extracted from site. To analyse its liability te test was also applied to some lime mortars prepared in laboratory, wit known capillarity absorption coefficients determined by te metod proposed by EN (CEN 1999), wic is based on te immersion in 5-10mm of water of alf prismatic specimens wit dimensions of 40 mm x 40 mm x 160 mm. Tis procedure permitted to compare te results obtained by bot metods and to evaluate te correlation between tem (figs. 1, 2). Figure 1 Standardised capillarity test Figure 2 Capillarity test for old mortars 2 Test metod 2.1 General Te test developed to evaluate te water absorption by capillarity of irregular, friable, samples is based on te placement of te external face of te specimen in contact wit water (instead of in immersion) troug te use of wire basket and wet geotextile gauze. It comprises te following pases: a) Specimens preparation Tree specimens are selected from eac sample taken in situ, wit volumes as similar as possible to tose of te alf prisms of 40 mm x 40 mm x 160 mm required by te standardised metod and wit a rater flat face (te external face), wic will be put in contact wit te water. Tey are brused to remove loose particles and biologic colonisation. page 2 2

3 b) Test Procedure Te geotextile gauze is wet before te beginning of te test, to avoid te influence of a slow initial absorption of tis kind of material. Te dry specimen is weiged (fig. 3) and so are te wire basket and te wet gauze (fig. 4). Te basket wit gauze is placed in a transparent recipient wit water, on a lat, taking care tat te lower face of te gauze barely touces water and te specimen is placed on te basket (fig. 5). Te gauze must not enter te surface of te water more tan 2 mm, so tat te specimen is in contact wit te wet gauze, but not in immersion in water. Tis situation must be kept during te test.te specimen is weiged periodically, togeter wit te wole wire basket and wet gauze, wit intervals of 5 minutes for te first 40 minutes and ten at 60, 90, 180, 300, 480 and 1440 minutes. Te absorbed water is quantified by te weigt increase. After 1440 minutes (24 ), te wole basket, gauze and specimen remain out of water and drying period begins, wit periodical weiging operations to determine te water loss. c) Calculations Te absorbed water is determined troug te difference between te weigts measured periodically and te initial weigt. To measure te specimen irregular surface in contact wit water, tis surface is placed on a seet of millimetric paper and its contour is drawn (fig. 6). Te obtained area is divided into geometric sapes and te total area is determined as te sum of tose regular easily calculated areas. Te capillarity coefficient by contact between instants t 1 and t 2 (Ccc t2-t1 ) is ten calculated troug te following expression: Ccc t2-t1 = (M 2 - M 1 )/(t 2 -t 1 ). Figure 3 Weiging te dry specimen Figure 4 Weiging basket and gauze Figure 5 Placement in a recipient wit water Figure 6 Measurement of te absorption area page 3 3

4 2.2 Samples description Samples of renders from five monuments were included in te study and some specimens of eac one were subjected to te water absorption test adapted for irregular, friable samples. Simultaneously, six different lime based mortars prepared in laboratory wit known capillarity coefficients determined by te standardised metod were also subjected to te same test. Tose laboratory prepared mortars were used to evaluate te correlation between te standardised test and te proposed one. Te main caracteristics of laboratorial mortar specimens are syntesized on table 1 and te monuments and main caracteristics of te tested old mortar samples are summarized on table 2. Mortar Table 1 Caracteristics of laboratory prepared mortars used for validation. Volumetric dosage Composition Constituents Mass of te specimen (g) Density (kg/m 3 ) Caracteristics at 28 days Compressive resistance (MPa) Capillarity coefficient C (90-10) (kg/m 2. 1/2 ) L 1:3 Air lime : river sand L-Cl 1:0.2:2.8 Lls 1:3 AH 1:3 L-AH 1:1:6 L-Ce 1:1:6 Air lime : clay : river sand Air lime : mixed sand from te Lisbon region Artificial Hydraulic lime : mixed sand from te Lisbon region Air lime : artificial ydraulic lime : mixed sand from te Lisbon region Lime : cement : mixed sand from te Lisbon region Table 2 Caracteristics of samples collected in situ. page 4 4

5 Monument Caracteristics Average Mortar Compressive Type Age Composition specimens resistance (MPa) mass (g) CSP1 Ligt brown lime mortar 248 Not determined Castle joints XII CSP2 wit wite lime grains, (Sout Portugal) century CSP3 siliceous sand and clay 356 Not determined Arab Wall Wite lime mortar, very VIII AWP joints (Sout ard, wit wite lime 147 Not determined century Portugal) grains and small sells CLis CSI1R CSI2J CSI3R CSI4P RSI5J WSI6R Convent in te Lisbon area - render Castle-render (Sout-East Slovenia) Castle-joint (Sout-East Slovenia) Castle-render (Sout-East Slovenia) Castle-plaster (Sout-East Slovenia) Roman settlement-joint (Sout Slovenia) Rendered clay brick wall for field studies (Ljubljana) XVI century XIII century XIII XVI XVI-XVIII I XX Ligt brown lime mortar, very ard, wit wite lime grains Ligt brown lime render mortar wit dolomitic crused aggregate and siliceous sand. Degradation process is visible by cemical analysis (SANTOS SILVA 2004) Ligt brown lime joint mortar wit dolomitic crused aggregate and siliceous sand. Degradation process is visible by cemical analysis (SANTOS SILVA 2004) Ligt brown lime render mortar, wit wite lime grains, dolomitic crused aggregate and siliceous sand Wite lime plaster mortar wit wite lime grains dolomitic crused aggregate and siliceous sand Wite lime joint mortar, very ard, wit wite lime grains, aggregate from Sava river and clay bricks particles Wite lime mortar (1:3) made from traditionally produced lime putty and siliceous sand Not determined 332 Not determined 268 Not determined page 5 5

6 2.3 Test results Tables 3 and 4 and grapics 1 to 9 present te test results. Table 3 Capillarity Coefficients of laboratorial mortars by immersion (C) and by contact (Ccc) for comparison Mortar Area in contact wit water (mm 2 ) Specimen mass (g) Area/mass (mm 2 /g) At 5 min Capillarity coefficient (kg/m 2. 1/2 ) Between 90 and 10 min Ccc5 Ccc90-10 C90-10 Factor C/Ccc L L-Cl Ls AH L-AH L-Ce Mortar Table 4 Capillarity Coefficients of ancient mortars by contact Area of specimens in contact wit water (average) (mm 2 ) Specimens mass (g) Area/mass (mm 2 /g) Capillarity coefficient (kg/m 2. 1/2 ) Ccc5 Ccc90-10 Ccc24 CSP CSP CSP AWP CLis CSI1R CSI2J CSI3R CSI4P RSI5J WSI6R* * about five years old lime render from clay brick wall for field studies page 6 6

7 ,00 9,00 1,00 9,00 1,00 3,00 5,00 TIME ( 1/2 ) 0,50 1,00 1,50 2,50 3,00 TIME ( 1/2 ) S1 S2 S3 L (AVERAGE) Grap 1 L mortar Lls L-Cl AH L-AH L-Ce Grap 2 Oter laboratorial mortars ,00 1,00 9,00 1,00 3,00 5,00 TIME ( 1/2 ) 1,00 3,00 5,00 TIME ( 1/2 ) S1 S2 S3 CSP1 (AVERAGE) Grap 3 CSP1 mortar S1 S2 S3 CSP2 (AVERAGE) Grap 4 CSP2 mortar page 7 7

8 ,00 9,00 1,00 9,00 1,00 3,00 5,00 TIME ( 1/2 ) 1,00 3,00 5,00 TIME ( 1/2 ) S1 S2 S3 CSP3 (AVERAGE) S1 S2 S3 AWP (AVERAGE) Grap 5 CSP3 mortar Grap 6 AWP mortar ,00 9,00 1,00 9,00 1,00 3,00 5,00 1,00 3,00 5,00 TIME ( 1/2 ) TIME ( 1/2 ) S1 S2 S3 CLis (AVERAGE) S1 S2 CSL1R (AVERAGE) Grap 7 CLis mortar Grap 8 CSL1R mortar 1 1 1,00 9,00 1,00 3,00 5,00 TIME ( 1/2 ) S1 S2 CSL2J (AVERAGE) Grap 9 CSL2J mortar page 8 8

9 3 Conclusions 3.1 Validity of te metod Te comparison between capillarity coefficients obtained by te contact metod and capillarity coefficients determined by te standardised immersion metod for laboratorial mortars sows tat tere is some correlation, wit a ratio C/Ccc varying from 2 to 3.3. Tis variation seems acceptable, considering tat bot metods ave some variability. In particular, capillarity of recent lime mortars canges quickly wit age and te consequent increase in te degree of carbonation. So tere is a possibility tat wen te new test was applied to tose samples, some time after te accomplisment of te standardised one, tey ad already actually reduced sligtly teir capillarity. Dispersion of results for te different specimens of te same mortar, wic can be evaluated by grapics 1 and 3 to 9, is not too ig, considering te variability of several factors. Te comparison between compressive resistance and capillarity of old mortars seems also consistent, sowing tat generally to iger strengt (mortars from a Slovenian Castle and from a Convent in te Lisbon area) corresponds lower capillarity during te first minutes as expected. Similarly, mortars wit lower strengts (mortars from a Portuguese Castle and from an experimental Slovenian wall) sow iger early capillarity. To increase te liability of te metod it seems important to ave a more or less constant area/volume ratio, for example about 25 m -1, similar to tat used for new mortars tests. Because te weigt of an old mortar is muc easier to measure tan its volume and it implies a lower risk of deterioration, tat relation can be substituted by te relation area/weigt, wic is a sufficient indicator of te sape similarity. Te volume of te samples, or teir weigt to simplify, must also be similar. 250 g migt be a good average mass to consider (similar to te alf prism of 40 x 40 x 160 mm used for te standardised test). In te present study tese rules were not entirely respected, especially for te first tested samples. 3.2 Analysis of te results for old mortars Te stabilisation of water absorption appens muc before 24, consequently Ccc24 is meaningless for traditional lime mortars, or at least it doesn t sow an absorption rate. Te old lime mortars studied present capillarity coefficients rater similar to new lime mortars between 90 and 10 minutes. Neverteless, tey ave muc slower absorptions during te first minutes, as is sown by Ccc5 significantly lower tan new mortars and rater close to Ccc Tis means, apparently, tat te absorption rate of old mortars remain almost constant during te first 90 minutes, wile new mortars ave a muc iger rate of absorption during te first minutes and tan stabilize. Te grapics 1 to 9 confirm tis effect. It is suggested by literature tat a ig initial absorption ratio indicates a weak mortar, wit larger pores, more subject to deterioration and consequently less durable (THORBORG 1997). Determinations of porosimetry must be performed to study tese differences. On site beaviour confirms tose conclusions about durability, as still existent old lime mortars, especially in monuments and important civil and military buildings, sow a iger resistance tan lime mortars recently applied. Te reason for tis may lay on te careful application, some small but influent difference in constituents, or simply on te evolution troug time, due to complete carbonation and possibly oter slow reactions. page 9 9

10 3.3 Utility of te metod Wen it is not possible to use te standardised metod, te described metodology seems to furnis a definite possibility of assessing te capillarity of old mortars, wit resource to samples extracted from site. As referred before, tis caracteristic, in particular te evolution of absorption along time, is an important indicator of te resistance to water effects and consequently of durability of te mortar. Tese results must be complemented by mercury porosimetry to understand te causes of te differences obtained. On te oter side, te verification of a quick drying is very important to avoid damage to te wall due to a long retention of water inside it. In conclusion, te test presented as an expressive utility to old mortars study and conservation and to te selection and preparation of new compatible mortars. Te fact tat tere are not so many tests to assess te pysical caracteristics of old mortars increases its interest. Acknowledgements Te autors tank to te tecnicians João Júnior, Bento Sabala and Dora Santos, of LNEC, wo gave a precious elp to prepare te specimens and to accomplis te tests. References EUROPEAN COMMITTEE OF STANDARDIZATION (CEN), 1999, Metods of test for mortar for masonry Part 18: Determination of water absorption coefficient due to capillary action of ardened rendering mortar, Brussels, April 1999, EN SANTOS SILVA, 2004, Caracterização de argamassas de reboco e de juntas de um Castelo na Eslovénia, Lisbon, LNEC, February 2004, NT 11/2004. THORBORG, Konow, 1997, Restaurering oc reparation med puts-oc murbruk. ABO AKADEMIS FORLAG ABO AKADEMI UNIVERSITY PRESS. Abo, Finland, VEIGA, M. Rosário, 2000, Protecção contra a água de paredes de edifícios antigos. Avaliação experimental da capacidade de protecção de argamassas de reboco com base em cal. Proc. of Encontro Nacional sobre Conservação e Reabilitação de Estruturas - REPAR Lisbon, Portugal, LNEC, June VEIGA, M. Rosário et al, 2001, Metodologies for caracterisation and repair of mortars of ancient buildings. Proc. of International Seminar Historical Constructions Guimarães, Universidade do Mino, November page 10 10