Épitőanyag 2014. 4. szám

8/9/2019 pitanyag 2014. 4. szm

1/48

ptanyag A Sziliktipari Tudomnyos Egyeslet lapjaA TARTALOMBL:

Variability case study basedon in-situ rebound hardnesstesting of concretePart 2. Statistical analysis ofspatial variability parameters

Formation of Catalyst ModelDispersed of Pdon a thin MgO (100)

Performance of novel typethree dimensionally deformedsteel fibres for concrete

Utilization of repair mortar

for the loss compensation ofHungarian porous limestone

Discrete Element Modellingof uniaxial compression test ofhardened concrete

Image digitalization as a toolfor processing experimentaldata of crack width of concrete

Investigation and optimizationof homogeneity of ceramicinjection molding raw materialto improve crack toughness ofend product

Review of the Glasstec 2014exhibition in Dsseldorf

2014/4

ptanyag ptanyag Journal of Silicate Based and Composite MaterialsJournal of Silicate Based and Composite MaterialsJournal of Silicate Based and Composite Materials

8/9/2019 pitanyag 2014. 4. szm

2/48

Applications

The steel industry produces not only metal, but also a by-productwhich has been successfully used in many construction or agricultureapplications. Nowadays, about 87 wt% of ferrous slag produced inEurope is used for building purposes. Using slag instead of naturalmaterials is a sustainable alternative with high durability in severalapplications.

Slag is...

proven over centuries. From the beginning of the Iron Age, 800 B.C.,through the invention of the blast furnace process in the 18th century,slag products have been successfully marketed and put to bene cialuse.

... ecological and intelligent. With various techniques for furtherprocessing, slag is prepared for a wide range of applications - asbuilding materials or fertilisers. Through recycling, a portion of thesteel slag is reused in the blast furnace as calcium and iron bearingmaterial.

... for roads. Slag is exceptionally well suited for road construction:as porous asphalt and as other construction materials with positivecharacteristics like durability and high skid resistance.

... for bridges. Slag is used for demanding construction projects such asbridges and skyscrapers.

... for waterways. In hydraulic engineering such as in building locks,river bottoms and bank reinforcements slag is a proven material.

... as fertiliser. Slag products have been successfully used as fertilisersfor generations.

The use of slag is ecologically sound and economically smart.

The European Association representingmetallurgical slag producers and processors

Building Materials Institute, FEhS Institut fr Baustoff-Forschung e.V.Bliersheimer Strasse 62, 47229 Duisburg-Rheinhausen, GermanyPhone +49.2065.9945-0 Fax +49.2065.9945-10E-Mail info(at)euroslag.org http://www.euroslag.com

From roads to felds,slag is safely usedin a wide range ofactivities with highquality expectations.

Porous asphalt designed for noisereduction

Rheinkniebrcke in Dsseldorf, Germany

Steel slag asphalt used on heavily traf-cked road

8/9/2019 pitanyag 2014. 4. szm

3/48

S zerkeSztbizottSg e ditorial boardP f. D. Gmze A. L l l /p dD . Borosnyi Ad f / d - - h f Wojnrovitsn D . HrAPkA il a

l b l f l l / d - - h f ttH-AsztALos r a /d g d

t agok MeMberSP f. D . Pa ALizADeH, BocskAy Bal ,P f. D. cske Ba ab , P f. D . ka h t. FABer,P f. D . sa Fiore, P f. D . Da d Hui,P f. D . GLos m l , D . v GriBniAk,P f. D . k isHizAki, D . jzsA z a a,krPti L l, D . kocserHA i ,D . kovcs k f, P f. D . s g N. kuLkov ,mAttyAsovszky zsoLnAy e , D . mucsi G b ,P f. D . oPoczky L d lla, D . PLvLGyi ta ,D . rvAy m l , P f. D . t a SADOWSKI,

P f. D . t h sekino, P f. D . Da d s. smitH,P f. D . B a sreeDHAr, P f. D . szPvLGyi j ,P f. D. szcs i , P f. D. tAms F

t ancSad t eStlet a dviSory boardFintA F , kiss r b , D . mizser j

A f l a f l a th al f d b :ca b dg s f Ab a P Q

A f l a ba l l l g.All p bl h d pap a p - w d.k ad P bl h : s l pa t d eg l (szte)el P d : AsztALos i1034 B dap , B 122124.t l./fax: +36-1/201-9360e- a l: p a ag@ . g.ht d l La d :nmetH Ha al ac lapf c ph :ksA L a k l a

HirdetSi rak 2014 ADVERTISING RATES 2014:B2 b l 76 000 F 304 eurB3 b l 70 000 F 280 eurB4 b l 85 000 F 340 eur1/1 ldal pag l 64 000 F 256 eur1/1 ldal f -f h pag b&w 32 000 F 128 eur1/2 ldal pag l 32 000 F 128 eur1/2 ldal f -f h pag b&w 16 000 F 64 eur1/4 ldal pag l 16 000 F 64 eur1/4 ldal f -f h pag b&w 8 000 F 32 eurA a a f a al a . W h vAt.

A h d g d l l l h a f l a h lap l.o d -f f ad a al abl h w b

f h al.

www .epitoanyag .org .Huo l issn: 2064-4477 P issn: 0013-970x

inDeX: 2 52 50 66 (2014) 93136

a z Szte tMogat tagvllalatai

S upporting coMpanieS of Szte

3B H g a kf . A L q d kf . A kf .

Ba a a t gla kf . B t gla pa kf .

B da t gla z . B dap k a kf .

c l x kf . c la - a kf . el -c d kf .

F l g kf . G&B ela kf . Ge H ga z .

G a kf . G a d a o h a kf . H x kf .

i kf . kk ka B kf . kkA kf .

k kf . kti kh . k a - kf .

La bda s kf . L b l La L gh h kf .

mal ha H ga kf . m H ga g kf .

mFL H g a kf . m alh ld g kf .

mtA kk Aki o-i ma fa g mag a g kf .

omyA kf . P pa t gl. kf . P l -92 kf . Q&L kf .

rAtH H g a kf . r w l H ga kf .

sp l Ba kf . s a ma kf . szikkti Lab kf .

WiteG kp kf . zala a z .

Vol. 66, No. 4

A finomkermia-, veg-, cement-, msz-, beton-, tgla- s cserp-, k- s kavics-, tzllanyag-, szigetelanyag-ipargak szakmai lapjaScientific journal of ceramics, glass, cement, concrete, clay products, stone and gravel, insulating and fireproof materials and composites

TARTALOM94 e a l b

h l g l2. . A h l l g

a a lBorosnyi Ad j

100 Pd d p a al d ll p da mgo (100) gFat a BAArA Abd lbak ChemAm

105 p , h d ba g la l la la d gaKereKes Pt Borosnyi Ad j

109 mag a g p hal al a ha g a agsZemerey-Kiss Balz TrK k

113 m g l d l b l d g g la ad ll d l d l

GyurK Z lt BAGi Katal Borosnyi Ad j

120 B p d ga g- df ld lg a pd g al l l alnAGy rka

125 k a f alapa ag h g ag la a p al l a a g p d

g a f a d beGsZ d GmZe A. L zl

131 B l a 2014. d ld f Gla ll l JAKAB A d nehme K ga moLnr Pt sal G g nehme

CONTENT94 va ab l a d ba d - b d

ha d g f Pa 2. s a al a al f pa al a ab lpa a Ad j Borosnyi

100 F a f ca al m d l D p d f Pd a h mgo (100)

Fat a BAArA Abd lbak ChemAm

105 P f a f l p h d alld f d l f b f Pt KereKes Ad j Borosnyi

109 u l a f pa a f h l p af H ga a p l

Balz sZemerey-Kiss k TrK

113 D el m d ll g f ax alp f ha d d

Z lt GyurK Katal BAGi Ad j Borosnyi

120 i ag d g al a a a l f p gxp al da a f a w d h f

rka nAGy

125 i ga a d p a f h gf a ld g aw a al

p a gh f d p d d eGsZ L zl A. GmZe

131 r w f h Gla 2014 xh b D ld f A d JAKAB K ga nehme Pt moLnr sal G g nehme

ptanyag 2014/4ptanyag 2014/4Journal of Silicate Based and Composite MaterialsJournal of Silicate Based and Composite Materials

8/9/2019 pitanyag 2014. 4. szm

4/48

ptanyag Journal of Silicate Based and Composite Materials

Variability case study based on in-siturebound hardness testing of concretePart 2. Statistical analysis of spatial variability parameters

a dorjn BOROSNYI A . P f., Bme, D p . f c ma al a d e g g G l g b .ad a @ p .b .hr d: 28. 08. 2014. : 2014. 08. 28. h p://dx.d . g/10.14382/ p a ag- b .2014.17

Abstractspa al a g aph a al ha b p f d b d d x da a ll d - a hb fa f a fl 130 2 f g a a. P bl f g a al

h d f a g a , ad g a a d d g a a al wa d a d. iha b al d ha h la al h d f g a a b f h f d f

a fa h pa al a ab l a al f - b d ha d l a d a p bl d wa d h f p al a f x al

ff f p p d.k w d : s al ; n d g (nDt); r b d d x; spa al a ab l ;va g a

Adorjn Borosnyic v l g (ms ), PhD, A a P f

a Bme D p . f c maa d e g g G l gy. ma f ld

: a g a d d fl f , appl a f - all (

f f , d v g f

ppl a y g a al f hp f a , h

m b f h f b ta G p 4.1 s v abm d l , p d g b f riLem

t h al c 249-isc n d g h a f cha a f h szte c D v

1. IntroductionGeostatistical methods are widely used in mining, geology,

soil science, environmental science, hydrology, meteorologyand recently, in engineering sciences or remote sensing,sur ace texture modelling, reliability and risk analysis, serviceli e design and analysis o non-destructive testing [1-10]. Tespatial variability analysis is well-described in the geostatisticalliterature or a long time; the reader may re er to the severaltextbooks available in geostatistics [11-16]. Te use o thegeostatistical methods or concrete structures is, however, still very much limited today [8-10].

In-situ non-destructive testing (ND ) o concrete structuresmay target strength estimation o structural concrete either tocomplete destructive testing or in the absence o drilled cores

or laboratory testing [17-20]. In-situ assessment is requentlyinitiated by corrosion problems [21-23].

Structural concrete is a multiple-level heterogeneouscomposite material [24-25]. Tis multiple-level heterogeneityresults bothinherent (local) variability andspatial (regional) variability o the per ormance properties. Local variability canbe analysed by conventional mathematical statistical methods.Analysis o regional variability needs spatial models that aretypically used in geostatistics.

2. Scope of the studiesIn the rst part o the present series o papers, inherent

variability parameters were analysed corresponding to theindependent test areas o a rein orced concrete slab, in terms ostatistical location parameters, statistical variance parameters,statistical dispersion parameters and normality parameters [26].

In this second part, spatial variability is studied by contourplot, omnidirectional semivariogram (variogram o order2), omnidirectional madogram (variogram o order 1) andomnidirectional rodogram (variogram o order 1/2).

3. ExperimentalTe experimental background was provided by the same

structural element that was analysed in [26], i.e. the bottom

sur ace o the top concrete slab o a ramed, monolithic, subsoilconcrete tunnel was studied, with dimensions o 25.0 m 7.5m and a thickness o 0.48 m [10]. Te measuring region onthe bottom sur ace was 22.0 m 6.0 m. A total number o 42test areas were selected or Schmidt rebound hammer testing.N-type original Schmidt rebound hammer was used. Elevenindividual rebound index readings were recorded at each testarea. Te measurements were per ormed by the same operator.

4. Spatial variability analysis of the statisticalparameters

In a practical situation, when eventually weaker regionso structural concrete within the element are present, it canbe interesting to know that at which extent and in whichdirections the discrepancy is present and has inuence onthe per ormance properties. Geostatistical approaches canbe adapted or the analysis o spatial variation. Geostatisticsdeals with spatially autocorrelated data (autocorrelation =correlation between elements o a series and others rom thesame series separated rom each other by a given interval)and usually assumes that the differences between the values osamples are determined by the relative spatial distance o thesamples and the mean and variance o the differences dependonly on the relative distance [13]. Differentvariograms areintroduced in geostatistical spatial correlation analysis that plotdifferent correlation parameters o samples as the unction othe separation between two spatial locations (re erred to aslag ;indicated withh in the present paper). Empiricalsemivariogram can be composed by the empirical semivariances o order 2(see Appendix ). Empiricalmadogram can be composed by theempirical semivariances o order 1. Empiricalrodogram canbe composed by the empirical semivariances o order 1/2. Itis possible to composeomnidirectional variograms by takinginto account all pairs o data in any possible relative distance,and it is possible to composeunidirectional variograms by

taking those pairs o data into account that correspond toa given direction. Tis latter method was not applied in thepresent studies but would be used to nd anisotropy in the

94 | ptanyag JSBCM 2014/4 v l. 66, n . 4

8/9/2019 pitanyag 2014. 4. szm

5/48


v l. 66, n . 4 2014/4 ptanyag JSBCM | 95

spatial variability in a practical situation. In our nomenclature,the term variogram generally covers either semivariogram, or

madogram or rodogram throughout the present paper.I a variogram levels off (bounded) then a stationaryrandom eld can be used to model the observations [27].

According to the geostatistical modelling nomenclature, thesill is the value o the actual correlation parameter at which

the variogram levels off and therange is the lag distance overwhich the actual correlation parameter is constant. Te rangedistance can be re erred to ascorrelation distance, over which

Fig. 1. Variographic analysis or the normality parameters 1. bra Normalitsi paramterek variograi vizsglata

Fig. 2. Variographic analysis or the statistical location parameters 2. bra Helyzeti statisztikai paramterek variograi vizsglata

8/9/2019 pitanyag 2014. 4. szm

6/48



lag the values o the variable are not correlated (independent).Te correlation distance indicates the degree o similarity othe variable between two points as a unction o the distancethat separates them. Te nugget is the value o the actualcorrelation parameter at distances smaller than the minimumlag o observations [13,16]. Te nugget effect may providein ormation on strong differences in value within very shortdistances, or on a structural discontinuity, or on local materialdeterioration, or on an erroneous measurement [27].

Experimental variograms cannot be used directly ormodelling and simulation. A wide range o variogram modelsare introduced in the literature o geostatistics. ypical modelsare the spherical, exponential, Gaussian, wave, nugget models

[13,16]. Combined models are also used.Figs. 1 to 4 summarize the empirical omnidirectionalsemivariograms, madograms and rodograms corresponding

to the statistical parameters analysed. In the present study, theconcrete slab has dimensions o 25.0 m 7.5 m o which thestudied region is 22.0 m 6.0 m. In geostatistics, a commonrule o thumb is accepted or the maximum lag in a variogramrestricted to hal o the diagonal o data extent [28]. It shouldbe noted that lag is intentionally not limited to about 12 meterssince one dimension o the slab is three times larger than theother. Restricting the variograms at maximum lag about halo the diagonal o data extent would not allow the diagrams to visibly level off and the sill could not be clearly determined. Itshould be also noted that intentionally no ltering was applied

or the calculated correlation parameters but the suspectedoutlier values are indicated with empty markers in the

variograms. Spherical model is tted or all the 36 variogramsthat is also indicated inFigs. 1 to 4.

Fig. 3. Variographic analysis or the statistical variance parameters 3. bra Variancia paramterek variograi vizsglata

Fig. 4. Variographic analysis or the statistical dispersion parameters 4. bra erjedelem paramterek variograi vizsglata

8/9/2019 pitanyag 2014. 4. szm

7/48



5. DiscussionTe strength estimation o concrete by rebound hammer

testing usually applies empirical relationships betweenstatistical location parameters (mostly the mean and themedian) and the compressive strength. It can be realized bystudying the contour plots (Fig. 2) and as was shown earlier

by linear correlations that the statistical location parametersare interrelated and a rather strong correlation is oundbetween the mean and median values or the concrete oorin the present study [10,26]. It was also demonstrated thatno correlation is ound between the statistical locationparameters and the statistical dispersion parameters; nocorrelation is ound between the statistical location parametersand normality parameters; as well as no correlation is oundbetween the statistical dispersion parameters and normalityparameters [26].

It can be realized by studying the contour plots oFig. 3 and4 that the interrelated nature o the range and the standarddeviation demonstrated earlier [26] is visible or the concreteoor in the present study.

Contour plots o Fig. 1 and 4 conrm the earlier ndings[26] about the interrelated nature o the studentized range andexcess kurtosis.

It can be realized in the spatial variability analysis thatthe semivariograms, madograms and rodograms can beconstructed or all the twelve statistical parameters and all variograms are bounded and level off at a more or less clearlyrecognizable sill. Nugget is visible in all cases. It can be generallyconcluded that semivariograms are the most structured androdograms are the least structured. Te relative nugget effect(ratio o nugget and sill) is the strongest or the semivariogramsand the least pronounced or the rodograms. Correlationranges are very much similar or all types o variograms and

or all the twelve statistical parameters studied. Te correlationrange was ound to be about ten meters or the concrete oorin the present study.

Te variograms constructed do not clearly demonstraterecognizable differences in the spatial variabilitybehaviour (ortrend) o the twelve statistical parameters studied, there ore,another direction or the spatial variability analysis is presentedhere towards the analogues o the analyses o max-stablestochastic processes.

6. Outlook Skewed distributions were ound to be t with the best

goodness o t or the vast majority o requency histogramso the local statistical parameters analysed or the concreteoor in the present study [26]. It was also demonstrated thatbest goodness o t o the Fisher- ippett (Generalized extreme value, GEV) distribution could be ound or the individualrebound indices collected at the 42 individual test areas ino separation o data by location is applied [26]. Tere ore,an analogue or the empirical estimation o the extremalcoefficient unction (h) is introduced in the present spatial

variability analysis.Extremal coefficient provides a measure o the degreeo spatial dependence between locations or max-stable

stochastic processes [29]. Similarly to the semivariogramthat can be a tool or measuring dependence o Gaussianstochastic processes (due to the relationship with covariance),the madogram is proposed as a tool or measuring dependencein max-stable stochastic processes (due to the relationshipwith the extremal coefficient) [30]. It was shown that anempirical estimation o the extremal coefficient unction canbe constructed with the introduction o a modied madogram,the F -madogram (h), proposed or the distribution unctionF (values o F -madograms range rom 0 to 1/6 correspondingto complete dependence and independence, respectively) [31].Te empirical estimation o the extremal coefficient unctionbecomes:

where

Since extreme value distributions were ound to providebest goodness o t or the statistical parameters tested in thepresent analyses, a simple analogue o the extremal coefficient

unction is proposed or the empirical madograms as ollows:

where is the empirical semivariance o order 1 (see Appendix ).

Results are indicated inFig. 5. It should be noted that outliers(i any) were ltered and were not plotted in the graphicalrepresentation. It can be realized on the contrary to the quitesimilar semivariogram, madogram or rodogram proles thatthe per ormance o the proposed dependence measure (1)(h)is apparently separated to quite different behaviours.

Tere are statistical parameters or whichno trend or very weaktrend is visible: this is the case ornormality parameters and orstatistical variance parameters. Clearlinear trend is visible orthe statistical dispersion parameters. Anexponential variogrammodel can be tted to thestatistical location parameters. Terelative nugget effect o the proposed dependence measure (1)(h) is ound to be considerably different or the mean, orthe median and or the mode.

At this time it is not clear how the proposed (1)(h)dependence measure can be applicable or the spatial analysiso real in-situ measurements, but it emphasizes that somemore sensitive parameters may be ormulated than thesemivariogram, madogram and rodogram, which behaviourwas not ound to be satis actorily distinctive or the statisticalparameters studied in the present analyses o rebound indexdata collected in-situ on a concrete oor. Future research isneeded in this eld.

7. ConclusionsSpatial variability analyses were carried out on in-situ

rebound hammer test results collected at the bottom sur aceo a concrete oor over 130 m2 o area tested. Te ollowingobservations can be highlighted:

8/9/2019 pitanyag 2014. 4. szm

8/48



1. Semivariogram, madogram and rodogram can besuccess ully constructed or any statistical parameterto visualize spatial variability. Variograms always leveloff and nugget is clearly observed as well. It was ound

that restricting o the variograms at maximum lag abouthal o the diagonal o data extent is not practical or thegeometry o the concrete oor studied.

2. It was ound that semivariograms are the most structuredand rodograms are the least structured. Te relativenugget effect is the strongest or the semivariograms andthe least pronounced or the rodograms. Correlationrange was ound to be about ten meters or the concreteoor in the present study, independently o the statisticalparameter tested. Further studies on the inuence ooutlier ltering in the spatial variability analysis or

rebound index data sets is addressed to uture worksince outliers may considerably increase both nuggetand sill in variograms, as indicated in [32].

3. It was ound that variograms constructed do not clearlydemonstrate recognizable differences in the spatial variability behaviour or different statistical parameters,there ore, a dependence measure was introduced as ananalogue or the empirical estimation o the extremalcoefficient unction (dened orF -madogram), withthe simple adaptation o the idea to madograms. It wasdemonstrated that the per ormance o the proposed

dependence measure is apparently separated to quitedifferent behaviours; providing more sensitivity thansemivariograms, madograms and rodograms.

8. AcknowledgementsAuthor grate ully acknowledges the support o the Hungarian

Scientic Research Fund project Durability and per ormancecharacteristics o concretes with novel type supplementarymaterials (O KA K 109233).

References[1] Matheron, G.: (1971) Te Teory o Regionalized Variables and its

Applications. cole Nationale Suprieure des Mines de Paris[2] Goovaerts, P.: (1999) Geostatistics in soil science: state-o -the-art and

perspectives.Geoderma, Vol. 89, No. 1-2, 1999, pp. 145. http://dx.doi.org/10.1016/S0016-7061(98)00078-0 [3] utmez, B. Hatipoglu, Z.: (2007) Spatial estimation model o porosity.

Computers & Geosciences, Vol. 33, No. 4, 2007, pp. 465475.http://dx.doi.org/10.1016/j.cageo.2006.07.008

[4] Balaguer, A. Ruiz, L. A. Hermosilla, . Recio, J. A.: (2010)Denition o a comprehensive set o texture semivariogram eatures andtheir evaluation or object-oriented image classication.Computers &Geosciences, Vol. 36, No. 2, 2010, pp. 231240.http://dx.doi.org/10.1016/j.cageo.2009.05.003

[5] Mlynarczuk, M.: (2010) Description and classication o rock sur aces bymeans o laser prolometry and mathematical morphology.International Journal o Rock Mechanics and Mining Sciences, Vol. 47, No. 1, 2010, pp.138149. http://dx.doi.org/10.1016/j.ijrmms.2009.09.004

[6] Stewart, M. G. Mullard, J. A.: (2007) Spatial time-dependent reliabilityanalysis o corrosion damage and the timing o rst repair or RCstructures.Engineering Structures, Vol. 29, No. 7, 2007, pp. 14571464.http://dx.doi.org/10.1016/j.engstruct.2006.09.004

[7] Li, J. Masia, M. J. Stewart, M. G., Lawrence, S. J.: (2014) Spatial variabilityand stochastic strength prediction o unrein orced masonry walls in verticalbending.Engineering Structures, Vol. 59, 2014, pp. 787797.

http://dx.doi.org/10.1016/j.engstruct.2013.11.031[8] Nguyen, N. . Sbarta, Z. M. Lataste, J. F. Breysse, D. Bos, F.:

(2013) Assessing the spatial variability o concrete structures using ND

Fig. 5. Analogue or the empirical estimation o the extremal coefficient unction or the statistical parameters studied 5. bra apasztalati szlsrtk egytthat ggvny analgijnak illusztrlsa a vizsglt statisztikai paramterekre vonatkozan
http://dx.doi.org/10.1016/S0016-7061(98)00078-0http://dx.doi.org/10.1016/S0016-7061(98)00078-0

8/9/2019 pitanyag 2014. 4. szm

9/48



techniques Laboratory tests and case study.Construction and Building Materials, Vol. 49, 2013, pp. 240250.http://dx.doi.org/10.1016/j.conbuildmat.2013.08.011

[9] Gomez-Cardenas, C. Sbarta, Z. M. Balayssac, J. P. Garnier, V. Breysse, D.:(2012) Spatial sampling optimization o non-destructive testing measurements.31mes Rencontres Universitaires de lAUGC, ENS Cachan, 29-31 May 2013.http://augc2013.ens-cachan. r/Data/Articles/Contribution1187.pd

[10] Borosnyi, A. Szilgyi, K.: (2013) Studies on the spatial variability o

rebound hammer test results recorded at in-situ testing. ptanyag-JSBCM , Vol. 65, No. 4, 2013, pp. 102106.http://dx.doi.org/10.14382/epitoanyag-jsbcm.2013.19

[11] Cressie, N. A. C.: (1993) Statistics or Spatial Data. John Wiley & Sons ISBN 978-0-471-00255-0

[12] Chils, J. P. Delner, P.: (1999) Geostatistics - Modeling SpatialUncertainty. John Wiley & Sons ISBN 0-471-08315-1

[13] Clark, I.: (2000) Practical Geostatistics.Geostokos Limited ISBN 9780-9548911-1-4

[14] Wackernagel, H.: (2003) Multivariate Geostatistics - An Introduction withApplications.Springer ISBN 978-3-662-05294-5

[15] Diggle, P. Ribeiro, J. P.: (2007) Model-based Geostatistics.Springer ISBN978-0-387-48536-2

[16] Webster, R., Oliver, M. A. (2007) Geostatistics or EnvironmentalScientists. John Wiley & SonsISBN 978-0-470-02858-2

[17] Breysse, D. Martinez-Fernandez, J. L.: (2013) Assessing concretestrength with rebound hammer: review o key issues and ideas or morereliable conclusions. Materials and Structures, Vol. 46, 2013, Publishedonline: 25 July 2013 http://dx.doi.org/10.1617/s11527-013-0139-9

[18] Vona, M. Nigro, D.: (2013) Evaluation o the predictive ability o the insitu concrete strength through core drilling and its effects on the capacityo the RC columns. Materials and Structures, Vol. 46, 2013, Publishedonline: 23 November 2013 http://dx.doi.org/10.1617/s11527-013-0214-2

[19] Antonaci, P. Bocca, P. Grazzini, A.: (2006) In situ determination otoughness indices o bre rein orced concrete. Materials and Structures,Vol. 39, pp. 283290. http://dx.doi.org/10.1617/s11527-005-9003-x

[20] Masi, A. Chiauzzi, L.: (2013) An experimental study on the within-member variability o in situ concrete strength in RC building structures.Construction and Building Materials, Vol. 47, 2013, pp. 951961.http://dx.doi.org/10.1016/j.conbuildmat.2013.05.102

[21] Long, A. E. Henderson, G. D. Montgomery, F. R.: (2001) Why assess theproperties o near-sur ace concrete?Construction and Building Materials, Vol. 15, No. 2-3, 2001, pp. 6579.http://dx.doi.org/10.1016/S0950-0618(00)00056-8

[22] Villagrn Zaccardi, Y. A. Brtora, A. Di Maio, A. A.: (2013)emperature and humidity inuences on the on-site active marine

corrosion o rein orced concrete elements. Materials and Structures, Vol.46, 2013, pp. 15271535 http://dx.doi.org/10.1617/s11527-012-9994-z

[23] ay, D. C. K. am, C. .: (1996) In situ investigation o the strength odeteriorated concrete.Construction and Building Materials, Vol. 10, No. 1,1996, pp. 1726. http://dx.doi.org/10.1016/0950-0618(95)00057-7

[24] Pekr, G.: (2013) Simple basic model or concrete and its application. Part2 - Factors that inuence compressive strength and drying shrinkage.ptanyag-JSBCM , Vol. 65, No. 3, 2013, pp. 7684.http://dx.doi.org/10.14382/epitoanyag-jsbcm.2013.15

[25] Pekr, G.: (2013) Simple basic model or concrete and its application.Part 3 - Factors affecting consistency, material balance equations and mixdesign. ptanyag-JSBCM , Vol. 65, No. 4, 2013, pp. 118126.http://dx.doi.org/10.14382/epitoanyag-jsbcm.2013.22

[26] Borosnyi, A.: (2014) Variability case study based on in-situ reboundhardness testing o concrete. Part 1. Statistical analysis o inherent variability parameters. ptanyag-JSBCM, Vol. 66, No. 3, pp. 8591.http://dx.doi.org/10.14382/epitoanyag-jsbcm.2014.16

[27] Breysse, D. Marache, A.: (2011) Some Estimates on the Variability oMaterial Properties. In: Construction Reliability Sa ety, Variability andSustainability (Eds. Baroth, J., Schoe s, F., Breysse, D.). John Wiley & Sons ISBN 978-1-84821-230-5

[28] Coombes, J.: (2005) Handy Hints For Variography - Guidelines or variogram analysis.Snowden Associates Pty Ltd .

[29] Ribatet, M.: (2009) A User's Guide to the SpatialExtremes Package. cole Polytechnique Fdrale de Lausanne Switzerland

[30] Cooley, D. S.: (2005) Statistical Analysis o Extremes Motivated byWeather and Climate Studies: Applied and Teoretical Advances. PhDthesis.University o Colorado

[31] Cooley, D. S. Cisewski, J. Erhardt, R. J. Jeon, S. Mannshardt, E. Omolo, B. O. Sun, Y.: (2012) A Survey o Spatial Extremes: MeasuringSpatial Dependence and Modeling Spatial Effects.REVS A -Statistical Journal , Vol. 10, 2012, pp. 135165. ISSN 1645-6726

[32] Oliver, M. A. Webster, R.: (2014) A tutorial guide to geostatistics:Computing and modelling variograms and kriging.Catena, Vol. 113,2014, pp. 5669. http://dx.doi.org/10.1016/j.catena.2013.09.006

Appendix. FormulaeTe ollowing regional statistical measures were calculated or the spatial variability during the present analyses:

empirical semivariance o order 2:

empirical semivariance o order 1:

empirical semivariance o order 1/2:

where:u vector o spatial coordinates (with 2D componentsx and y ), (u) variable under consideration as a unction o spatial location,h lag vector representing separation between two spatial locations, (u+h) lagged value o variable under consideration,N (h) the number o data pairs separated by lagh.

Local statistical measures were dened in Part 1. o present series o papers.

Re .:Borosnyi,Adorjn:Variability case study based on in-situ rebound

hardness testing o concrete. Part 2. ptanyag Journal o Silicate Based and Composite Materials,

Vol. 66, No. 4 (2014), 9499. p. http://dx.doi.org/10.14382/epitoanyag-jsbcm.2014.17

Esettanulmny betonszerkezet helysznikemnysgmrsrl

2. rsz. A mrhelyek kztti vltozkonysgstatisztikai elemzseA g a b f d al f ll 130 2-

a a g , h l s h d - alap g d a a l a a bh l l g a a pa

l a g pa a l g la a g. B a a l, h g

g la al a a ll l h f l a g a , ad g a d g a , a l f -d ll l d ll h . A g l a apa

ala l g ha fgg a al g a f gal a a g f a l .k l a a : b ; l g l

g; apa a ; h l g; a g a
http://dx.doi.org/10.1016/S0950-0618(00)00056-8http://dx.doi.org/10.1016/0950-0618(95)00057-7http://dx.doi.org/10.1016/0950-0618(95)00057-7http://dx.doi.org/10.1016/S0950-0618(00)00056-8

8/9/2019 pitanyag 2014. 4. szm

10/48


Formation of Catalyst ModelDispersed of Pd on a thin MgO (100)

d r . f atMa BAARA Lab a f fa a d fa f l d d - (Lesims) D pa f Ph , Fa l f s ,Bad m h a u -A aba-Alg a. baa afa a@ ah .fp rof . a bdelbaki CHEMAM P pa a s h l f s a d t h l g f A aba,A aba 23000, Alg a. a. h a @ p -a aba.d : 2014. 10. 10. r d: 10. 10. 2014. h p://dx.d . g/10.14382/ p a ag- b .2014.18

Abstractth l a h f a f a a al d l d p d f h Pd / hmgo (100) a al la d b d l p g a p g a g F a f wa . th la

ba d p pa a d d b a l p (tem), la d h f q a a d h l a a d h g w h. Pallad a pa l d p d

h mgo a d h p a a g 5731073 k a d d p f 1000 .th l a a p d a d g h h f a d l a . th g al

h g f h ag a l , l a , g w h a d al . th a ad f l d a wh h b a p a a f ll w g A h law.th b ha ag w h a AFm d f Ag/mgo a d A /mgo. th ph

f al xpla d a la d g a p . i h w ha h al ap dl wh h b a p a h gh.

k w d : a al , al , ag - x d , l a , pallad

Baara Fatma a l a h D pa f Phy

Bad m h a u v y A aba a d ha wa g xp v ga f

p p f l d . D . Baa a a h ff pap wh h h fa

a d b hav f l d a d p

Abdelbaki Chemam a xp f appl d phy a d ha a w

a g xp a g py. H a h f v al

pap f ld f d d a phyh y. P f. ch a a g la pa

f - p f L llaf d (Ha d a p h r a h Ga a g

a h v d h 6.06 val .

1. IntroductionNowadays, nano-objects provide a promising research or the

identication o new undamental properties o the materialsand their potential technological applications. Much effort isdevoted to understand the physical and chemical propertieso materials, which can serve as model catalyst systems.Consequently, undamental studies have been carried out on

a range o heterogeneous catalyst, or example, metal islandsgrown on thin lms [1-11] or on single-crystals sur aces [12-18]. Palladium deposits on the MgO(1 0 0) sur ace have becomeone o the most widely used model systems, and have givenrise to many detailed experimental studies [19-21]. Althoughthe main microscopic steps governing nucleation and growtho the lms are now understood, detailed characterization othese processes has proven difficult. Earlier, empirical andtheoretical studies o Pd over single crystals MgO, investigatedde ect nucleation [22-28] when nucleation centres occupyminority o sites. On the other hand, the results o nucleationkinetics over thin lms governed by random nucleation [1,29],

each atomic site is potentially a nucleation centre. In this study,we build upon many experimental and theoretical studies [1,30-32] have been carried out to understand these processes.Te aim o this work is to investigate the microscopicmechanisms, which can calculate various parameters related tothe quantitative study ocusing on the nucleation, growth andcoalescence o Pd / thin MgO (100) using Fortran so ware.

1.1 Transmission electron microscopy experiments

o understand the rst quantitative study o nucleationand growth o Pd on thin layer o MgO(100),we exploited theexperimental work o Henryet al [1],who used transmission

electron microscopy and electron diffraction at high energy tomeasure the Pd island density as a unction o time a giventemperature and a constant ux. Firstly, the MgO (100) / LiF

(100) / NaCl (100) composite layer is achieved which serves assupport. Palladium is then deposited with a ux o 11013 atomscm-2 s-1 and exposure time o 10 to 240 s on a substrate heatedat temperatures between 573 and 673K. A er deposition, thePd islands are in situ characterized with a transmission electronmicroscopy ( EM) to determine the island density. Te resultsare interpreted according to the theory o random nucleation.Te energy o adsorption and diffusion o palladium on MgO(100) are derived rom the latter theory. It was possible to vary the average size o particles in the range 0.8 - 3.5 nm. Teobtained cluster density varies rom 0.6 to 31012 cm-2 howeverthe covered area o the substrate sur ace is o 0.4 to 15 percent.

1.2 Over view of nucleation and growth theoriesNucleation on a sur ace has been discussed in both classical

thermodynamic and in atomistic terms, and both have a longhistory. Classical nucleation theory was developed by Volmer[33]. From this theory, the critical nucleus is only one atom, whichmeans that the dimer is already stable. In that case, the classical

nucleation theory is no longer applicable. Te growth processoccurs by accretion o adatoms. It is described by the atomisticnucleation theory, which has been developed by Zinsmeister[32]. In this case, the critical nucleus composites o two atomsand the supercritical nuclei do not dissociate. Te probability oadsorption is equal to one and that only single atom can return inthe phase steam. Te requency o nucleation is then determinedby the requency o meeting o adsorbed atoms.

1.2.1 Nucleation kineticsTe rate equations given by Zinsmeister express the variation

with time o the number o clusters o size i: or i = 2,3 (1)


8/9/2019 pitanyag 2014. 4. szm

11/48



i is the attachment requency o an adatom to a clustercontaining i atoms which is expressed by:

(2)

where i is the capture number or a cluster o size i. We haveto integrate the system o differential equations in order tocalculate the number o nuclei present on the substrate. Forsake o simplicity Zinsmeister assumes that it is a constantbetween 1 and 4. From this scheme the nucleation requency is:

(3)

Supposing that the growth is negligible, the density o adatomsis equal to the stationary value:

(4)

where F is the ux o atoms impinging on the substrate and the mean li e time o an adatom be ore desorption. Ten,combining equations (2) and (4) the nucleation rate becomes:

(5)Te nucleation rate is proportional to the square o the

impinging ux or a homogeneous substrate without de ects.

1.2.2 Growth kinetics.Zinsmeister has solved the system o differential equation

assuming a constant value or the attachment requency ( i)[18]. However by this treatment several characteristics o thegrowth o clusters are not taken into explanation. Severalresearches have tried to treat more precisely the calculation othe attachment requencies [3439]. In the typical growth, the

diffusion o adatoms is limited by desorption and the diffusionlength Xs o an adatom is:

(6)

where Ds is the sur ace diffusion coefficient. It is expressed by:

(7)

Ten the mean li e time o a physisorbed molecule can beexplained by:

(8)

where Ea and Ed is the adsorption and the diffusion energy,

and a0 the jump distance, o the order o the sur ace repeatdistance, say 0.2-0.5 nm[40], d and a the requency actorsor the diffusion and the adsorption process, K the Boltzmann

constant and s the substrate temperature. In a general case, thegrowth rate o a cluster can be expressed by a power law o thedeposition time.

(9)

Te exponent or any experimental has been calculated inthe case o the lattice approximation or 3D clusters.

1.2.3 Coalescence

Te trans ormation o two touching nuclei into one nucleuscan be described by a time constant shown by Nichols andMullins [41]:

(10)

where R is the radius o the coalescing spheres, the sur aceree energy and the atomic volume o Pd.

1.2.4 Algorithm

We have been developed many programs exploiting Fortranso ware. Te ollowing list details this mapping.1. Pd deposition ux rate on MgO (100) is 1.131013

atoms cm-2 s-1.2. Pd atoms are deposited randomly onto the sur ace with

activation energy o about 0.22 eV.3. Pd nanoparticles deposited on thin MgO are tested in

the temperature range 5731073 K and deposition timeo 1000 s.

4. Pd islands are approximated to be three-dimensionalclusters.

5. Te diffusion o adatoms is limited by desorption.

Hence, the values o the sur ace diffusion are calculatedby combination o Equations (6) and (7).6. Te entry parameters are: the velocity o nucleation,

velocity o growth, the average mean li e time, thesur ace repeat distance, the diffusion length, the sur ace

ree energy, the atomic volume o Pd, the activation andthe diffusion energies.

2. ResultsFig. 1 shows the variation o cluster density as a unction

o exposure time at different substrate temperatures ranging

rom 573K to 1073K and a constant palladium ux 11013

atoms cm-2 s-1. For s = 573K and 673K, we can see that thedensity o clusters is increasing rapidly a er 10 to 70 s (seetable 1) due to the large adsorption energy or the Pd adatomconrming the nucleation stage, up to a plateau (saturationdensity) corresponding to ns = 31012 cm-2 and 1.61012 cm-2 respectively. A similar behavior is observed or the remainingsubstrate temperatures till the coalescence occurrence, werethe cluster density decreases. It is worth to note that the clusterdensity decreases when the temperature increases.

Fig. 1. Nucleation kinetics o Pd on MgO (100) at different substrate temperatures ora palladium ux 11013 atoms cm-2 s-1.1. bra A Pd nukleci kinetikja MgO (100) vkony rtegen klnbz szubsztrtum

hmrskleten, ha a palldium uxus nagysga 11013 atom/cm2s.

8/9/2019 pitanyag 2014. 4. szm

12/48



t (s) 573 K 673 K 773 K 873 K 973 K 1073 K 10 9.39e11 4.85e11 2.97e11 2.03e11 1.50e11 1.18e1170 3.02e12 1.56e12 9.58e11 6.56e11 4.86e11 3.81e11 c n n 9.80e11 6.71e11 4.98e11 3.90e111000 3.02e12 1.60e12 6.77e11 2.59e11 1.64e11 1.19e11

able 1. Calculated o the cluster density (cm-2 ) at different substrate temperatures (K)and deposition times (s).

1. tblzat A klnbz K hmrskletekre s lerakdsi idkre (sec) szmtottklaszter-srsgek (cm2 ).

In Fig. 2, the saturation density is plotted in an Arrheniusdiagram. It is represented by the equation: ns = B0 exp (E /K s). When the activation energy E is equal to 0.22 0.05eV, B0 (pre exponential actor) is 3.631010 cm-2. We show thatit increases when the substrate temperature decreases. Tisbehavior is in agreement with a recent AFM study or Ag/MgO[42], Au/MgO (100) [43] and our previous studies by EM [1]on the same system.

Fig. 2. Arrhenius plot o the saturation density o palladium clusters on MgO (100). 2. bra A palldium klaszerek Arrhenius le teltettsgi srsge az MgO (100)

vkony rtegen.

Fig. 3 and 4 show the variations o the covered area (A)and the coalescence parameter (B) as a unction o substratetemperatures obtained under the same conditions. We seethat the raction o covered area decreases when depositiontemperature rises. An opposite behavior is observed or thecoalescence parameter.

Fig. 3. Variation o the covered area (A) as a unction o substrate temperatureswith deposition times 1000s. 3. bra Az 1000 msodperc id alatt lerakdott ellet (A) nagysgnak vltozsa

a szubsztrtum hmrsklet ggvnyben

Fig. 4. Variation o the coalescence parameter (B)as a unction o substrate temperatures with deposition times 1000s.

4. bra Az 1000 sec. id alatt ltrejtt egyeslsek szma (B)a szubsztrtum hmrsklet ggvnyben.

3. DiscussionFrom our previous empirical result or Pd/MgO thin lm, the

initial nucleation curves suggest the occurrence o a randomnucleation mechanism. Te theory o this nucleation process isexpressed by the kinetic equations o Zinsmeister [32]. Robinsonand Robins [30] have given analytical solutions in two limitcases, namely at low and high temperature. In this work, we

ocus on the high substrate temperatures(regime o incompletecondensation), where the particle number density is given by:

(11)

when the random nucleation model is used ,the calculated curvesin the rst stage agree well with the experimentally measured[1] time dependencies o island density (see nucleation regimein Fig. 1). Hence, the curves show a plateau as a maximumparticle density reached at the end o the nucleation regimecharacterizing the Volmer Weber growth. Te value o saturationo the island density is an important parameter which determinethe mode o thin lm growth. Te latter has an evident inuenceon the physicochemical properties o the obtained thin lms[28]. Te linear behavior o the Arrhenius plot observed or thetemperature dependence o the saturation density o clusters hasbeen also ound in the case o Ag / Ar-cleaved MgO(1 0 0) [42]Au /MgO(1 0 0) [43] and Pd/ UHV-cleaved MgO(1 0 0) [44,45].

, with m = 7 (12)

Te coalescence curves were better tted with a Clusterdiffusion model [31] rather than Ostwald ripening model.

Te most crucial parameter in our results is the coalescencetime. It is dened as the mean time or two clusters that comeinto contact to coalesce. From Eq. (10), we can note that twoparameters are important to determine the duration o thecoalescence stage, which are R and B. One can also noticethe in uence o the deposition temperature, that modi es theclusters coalescence time. It is clearly seen that clusters coalescemore rapidly at high temperature [46]. Tis phenomenon isexplained by the process o island migration.

Te process o island migration in this calculation is essentiallydescribed by the parameter B. Te derived B values romFig. 4

are not high enough meaning that the process o island densitycoalescence is not ignored even at the initial stages o deposition[47]. Te mechanism o coalescence which can be expected at

8/9/2019 pitanyag 2014. 4. szm

13/48



such a low value o sur ace coverage is the migration o islandson the sur ace, prior to the mechanism o immobile islands inwhich the coalescence occurs at high values o sur ace coveragewhen the islands touch each other [48].

Te sur ace coverage does not depend on B parameter, butit strongly depends on the cluster density and the shape o theparticles. Te raction o the substrate covered by the clusters,

which is considered as the contact sur ace o the hal sphere isa circle can be written as:

(13)

In this range o temperatures, the diameter (D) o theclusters ollows a power law: Do tp with 0.33

8/9/2019 pitanyag 2014. 4. szm

14/48



[19] Xu, C. Oh, W. S. Liu, G. Kim, D. Y. Goodman, D. W. : (1997)Characterization o metal clusters (Pd and Au) supported on various metaloxide sur aces (MgO and iO2), Journal o Vacuum Science & echnology A, Vol. 15, No. 3, 1997, pp. 1261-1268 http://dx.doi.org/10.1116/1.580604

[20] Henry, C. R. Chapon, C. Duriez, C. Giorgio, S. : (1991) Growth andmorphology o palladium particles epitaxially deposited on a MgO(100)sur ace,Sur ace Science, Vol. 253, No. 1-3, 1991, pp. 177-189http://dx.doi.org/10.1016/0039-6028(91)90591-F

[21]Fornander, H. Hultman, L. Birch, J. Sundgren, J.-E. : (1998) Initial growtho Pd on MgO(0 0 1), Journal of Crystal Growth, Vol. 186, No. 1-2, pp. 189-202http://dx.doi.org/10.1016/S0022-0248(97)00443-0

[22] Goyhenex, C. Meunier, M. Henry, C. R.: (1996) Limitation o Augerelectron spectroscopy in the determination o the metal-on-oxide growthmode: Pd on MgO(100),Sur ace Science,Vol. 350, No. 1-3, 1996, pp. 103-112. http://dx.doi.org/10.1016/0039-6028(95)01255-9

[23] Xu, Lijun Campbell, Charles . Jnsson, Hannes Henkelmana,Graeme: (2007) Kinetic Monte Carlo simulations o Pd deposition andisland growth on MgO(1 0 0),Sur ace Science Vol. 601, No. 14, pp. 31333142 http://dx.doi.org/10.1016/j.susc.2007.05.027

[24] Venables, J. A. Harding, J. H. : (2000) Nucleation and growth o supportedmetal clusters at de ect sites on oxide and halide (0 0 1) sur aces, Journal oCrystal Growth,Vol. 211, No. 1-4, 2000, pp. 27- 33http://dx.doi.org/10.1016/S0022-0248(99)00837-4

[25] Henry, C. R. Meunier, M.: (1998) Power laws in the growth kinetics ometal clusters on oxide sur aces,Vacuum, Vol. 50, No. 1-2, pp. 157-163http://dx.doi.org/10.1016/S0042-207X(98)00034-7

[26] Giordano, Livia Di Valentin, Cristiana Pacchioni, Gian ranco Goniakowski, Jacek: (2005) Formation o Pd dimers at regular and de ectsites o the MgO(1 0 0) sur ace: cluster model calculations,ChemicalPhysics, Vol. 309, No. 1, 2005, pp. 4147http://dx.doi.org/10.1016/j.chemphys.2004.02.022

[27] Sicolo, Sabrina Pacchioni, Gian ranco: (2008) Charging and stabilizationo Pd atoms and clusters on an electron-rich MgO sur ace,Sur ace Science Vol. 602, No. 16, pp. 28012807http://dx.doi.org/10.1016/j.susc.2008.07.005

[28] Giordano, Livia Pacchioni, Gian ranco: (2005) Pd nanoclusters at theMgO(1 0 0) sur ace,Sur ace Science Vol. 575, No. 1-2, pp. 197209http://dx.doi.org/10.1016/j.susc.2004.11.024

[29] Renaud, Gilles: (2005) Real- ime Monitoring o Growing Nanoparticlesby in situ Small Angle Grazing Incidence X-Ray Scattering, AIP Con erenceProceedings,Vol. 748. No. 63, 2005

http://dx.doi.org/10.1063/1.1896476[30] Robinson, V. N. E. Robins, J. I.: (1974) Nucleation kinetics o gold

deposited onto UHV cleaved sur aces o NaCl and KBr,Tin Solid Films,Vol. 20, No. 1, pp. 155-175

http://dx.doi.org/10.1016/0040-6090(74)90043-1[31] Kashchiev, D.: (1976) Kinetics o thin lm coalescence due to crystallite

sur ace migration,Sur ace Science Vol. 55, No. 2, pp. 477493http://dx.doi.org/10.1016/0039-6028(76)90253-3

[32] Zinsmeister, G.: (1966) A contribution to Frenkels theory o condensation,Vacuum, Vol. 16, No. 10, pp. 529-535http://dx.doi.org/10.1016/0042-207X(66)90349-6

[33] Volmer, M.: Kinetic der Phasenbildung.. Steinkopff Verlag , Leipzig (1939)[34] Halpern, V.: (1969) Cluster Growth and Saturation Island Densities in

Tin -Film Growth, Journal o Applied Physics, Vol. 40, No. 11, 1969, pp.4627-4636 http://dx.doi.org/10.1063/1.1657243

[35] Sigsbee, R.: (1971) Adatom Capture and Growth Rates o Nuclei, Journalo Applied Physics, Vol. 42, No. 10, 1971, pp. 3904-3915http://dx.doi.org/10.1063/1.1659705

[36] Stowell, M. J.: (1972) Capture numbers in thin lm nucleation theories,Philosophical Magazine, Vol. 26, No. 2, 1972, pp. 349-360 http://dx.doi.org/10.1080/14786437208227433

[37] Lewis, B.: (1970) Migration and capture processes in heterogeneous nucleationand growth: I. Teory,Sur ace Science, Vol. 21, No. 2, pp. 273288

http://dx.doi.org/10.1016/0039-6028(70)90234-7[38] Venables, J. A.: (1973) Rate equation approaches to thin lm nucleation

kinetics,Philosophical Magazine, Vol. 27, No. 3, 1973, pp. 697-738http://dx.doi.org/10.1080/14786437308219242

[39] Lewis, B. Rees, G. J.: (1974) Adatom migration, capture and decayamong competing nuclei on a substrate,Philosophical Magazine, Vol. 29,No. 6, 1974, pp. 1253-1280

http://dx.doi.org/10.1080/14786437408226184[40] Venables, J. A.: (1994) Atomic processes in crystal growth,Sur ace Science,

Vol. 299-300, No. 2, 1994, pp. 798-817http://dx.doi.org/10.1016/0039-6028(94)90698-X

[41] Nichols, F. A. Mullins, W. W.: (1965) Morphological Changes o a Sur ace

o Revolution due to Capillarity -Induced Sur ace Diffusion, Journal o Applied Physics, Vol. 36, No. 6, 1965, pp. 1826-1835

http://dx.doi.org/10.1063/1.1714360[42] Menck, A.: Ph.D. thesis, Lausanne, 1998[43] Hjrup-Hansen, K. Ferrero, S. Henry, C. R.: (2004) Nucleation and

growth kinetics o gold nanoparticles on MgO(1 0 0) studied by UHV-AFM, Applied Sur ace Science, Vol. 226, No. 1-3, pp. 167172http://dx.doi.org/10.1016/j.apsusc.2003.11.017

[44] Meunier, M.: Ph.D. thesis, Marseille, 1995[45] C.R. Henry (1998) Sur ace studies o supported model catalysts,Sur ace

Science Reports, Vol. 31, No. 7-8, pp. 231-325http://dx.doi.org/10.1016/S0167-5729(98)00002-8

[46] Carrey, J. Maurice, J.-L. Petro, F. Vaurs, A. : (2002) Growth o Auclusters on amorphous Al2O3: are small clusters more mobile than atoms?,Sur ace Science, Vol. 504, 2002, pp. 7582 http://dx.doi.org/10.1016/S0039-6028(01)01796-4

[47] Carrey, J. Maurice, J.-L.: (2002) Scaling laws near percolation duringthree-dimensional cluster growth: A Monte Carlo study, Physical ReviewB, Vol. 65, No. 20, 2002, 205401

http://dx.doi.org/10.1103/PhysRevB.65.205401[48] Bartelt, M. C. Evans, J. W.: (1994) Dendritic islands in metal-on-metal

epitaxy I. Shape transitions and diffusion at island edges, Sur ace Science,Vol. 314, No. 1, 1994, pp. L829L834

http://dx.doi.org/10.1016/S0039-6028(01)01796-4[49] Harsdorff, M.: (1982) Heterogeneous nucleation and growth o thin lms,

Tin Solid Films, Vol. 90, No. 1, pp. 1-14http://dx.doi.org/10.1016/0040-6090(82)90061-X

Re .:Baara, F. Chemam, A.:Formation o Catalyst Model Dispersed o Pd

on a thin MgO (100) ptanyag Journal o Silicate Based and Composite Materials,


Pd diszperz kataliztor modell kpzdse a vkonyMgO (100) rtegenA a F a f f lha l al g l

a al d ll d lg a a d p g l Pd pall da al al a al mgo ag - x d

g . e a l d ll a a l -ppal (tem) - g a p d

d g la ada alap A l a a a l a a d . A pall d a

b pl a mgo gb a 573-1073 k-fh l - a ba l 1000 h l

d a a ala . A apa al , h g a f l a a a al bb h a a a ha : a l l ( a d

a p d ; a d ; gl a g l . A Ah gf l l a l la

g g a b h ld l ha l a ah g a gf g l

a Ag/mgo ll A /mgo AFm (a f m pa l a . A g l l g a g

g - g f l a a al ag aha ; l f l a aa l g abb, l aga abb a h l .k l a a : a al , l , ag -

a p d , pall d

8/9/2019 pitanyag 2014. 4. szm

15/48



Performance of novel type threedimensionally deformed steel fibresfor concrete

p ter KEREKES Bs (ce) d , Bme Fa l f c l e g g w bxw b@g a l. a dorjn BOROSNYI A . P f., Bme, D p . f c ma al a d e g g G l g

b .ad a @ p .b .h : 2014. 10. 28. r d: 28. 10. 2014. h p://dx.d . g/10.14382/ p a ag- b .2014.19

AbstractP l a l f a w a l g a h a d d l p p d d hpap . o l p f 2D d f d l f b (w h h p d a f P X) a d fd ff p f 3D d f d l f b (P X v45; P X v90; P X s45; P X s60) wd g d, a fa d a d d d lab a d . th lab a h w dp g l . s p p f a f a p f 3D d f d l f b wa

b d. c d abl p f h b d p f a f l f b wa al d a d a d fl x al gh f sFrc p a p d fl x wa

d a d.k w d : , l f b , 3D g , b d, fl x al g h, gh

Pter Kerekesc v l g (Bs ) d a B dap

u v y f t h l gy a d e (BmeFa l y f c v l e g g. ma f ld

: d v l p a d p f a f f b f .

Adorjn Borosnyic v l g (ms ), PhD, A a P f

a Bme D p . f c ma ala d e g g G l gy. ma f ld f

: a g a d d fl f f, appl a f - all (FrP

f f , d v g f

ppl a y g a al f h gp f a , h l

m b f h f b ta G p 4.1 s v ab lm d l , p d g b f riLem

t h al c 249-isc n d v g h a f cha a f h szte c D v .

1. IntroductionTe use o Steel Fibre Rein orced Concrete (SFRC) has been

greatly increased in the past ew decades in civil engineering[1-6]. SFRC is mostly used or industrial oors but it canprovide a good solution or shear walls and pre abricatedelements as well. Advantages o SFRC are the less labour work,the reduction o brittleness and the increase o toughness andshear resistance o concrete.

Most important per ormance parameter o steel bres madeo cold drawn steel wires is the sur ace geometry, there ore,several different types o steel bres have been developed duringthe last decades. Hooked-end, undulated, crimped, twisted,end-anchoring, at-end and several combined geometriesare commercially available or concrete construction. Certaintypes o steel bres show considerably improved bond andanchorage in concrete compared to plain bres, however, thedifferent types o steel bres are usually de ormed only in twodimensions. Physical and chemical sur ace treatments may

also be applied to enhance bond per ormance. Diameters andlengths o steel bres vary rom 0.1 to 1 mm and 10 to 60 mm,respectively. Tere is, however, a gap in experience connectedto the per ormance o steel bres de ormed in 3D.

Tree types o bond are available in SFRC: 1) elastic orcetrans er; that is the adhesion o the bres to the concrete, 2)

rictional orce trans er; that is provided by riction between thebres and the concrete giving resistance against pull-out and3) mechanical orce trans er; that is provided by mechanicalinterlock o the bres that are intentionally de ormed alongtheir lengths. Development o steel bres which are de ormed

in 3D may provide better bond per ormance by enhancingrictional orce trans er component and mechanical orcetrans er component o the bond action.

2. Steel bres with novel 3D geometry In the present research and development project (started

in October, 2013), our different types o 3D de ormed steelbres and one type o 2D de ormed steel bre were designed,manu actured and studied under laboratory conditions [7].Te 2 dimensionally de ormed bre was the basic bre that wasused or the urther 3 dimensional de ormation process duringmanu acturing. A special device was constructed to produce

the special bres. Te device was designed directly or thepresent research and development project and the prototypebres were manu actured manually with the use o the device.

Te basic 2 dimensionally de ormed steel bre received theproduct name o PetiX (Fig. 1). Te basic 2 dimensionallyde ormed steel bre has a waved shape with three waves. Temagnitude o the waves is intentionally selected to be largerthan that o the commercially available or the waved geometrysteel bres. Te larger waves were chosen due to an expectedbetter per ormance o the bres when they are de ormed in3 dimensions. With smaller waves o the basic bre the effecto the 3 dimensional de ormations was expected to be less

remarkable.Te rst 3D bre is the PetiX V45 type steel bre (Fig. 2). Tisis the shape o 3D geometry that was developed rst during theresearch and development project, and this is the type o brethat was actually the initiation that started the research, backin October 2013. Te identi ying eature o the 3 dimensionallyde ormed PetiX V45 type steel bre is the middle wave that isbent out rom the basic bres 2D plane with 45 degrees.

Te second type o the 3D bres is the PetiX V90 type steelbre (Fig. 3). Tis extremely de ormed type o bre was createdsince it was unknown how the extent o the de ormation couldinuence the behaviour o the bres in concrete. Te concept is

the same or the PetiX V90 type steel bre as it was or the PetiXV45 type steel bre, but the de ormation rate was selected to be90 degrees, rather than 45 degrees.

8/9/2019 pitanyag 2014. 4. szm

16/48



Fig. 1. 2D deformed basic steel bre, PetiX 1. bra Kt dimenziban megmunklt PetiX alapszl

Fig. 2. 3D de ormed steel bre, PetiX V45 2. bra Hrom dimenziban megmunklt PetiX V45 aclszl

Fig. 3. 3D de ormed steel bre, PetiX V90 3. bra Hrom dimenziban megmunklt PetiX V90 aclszl

Te third type o the 3D bres is the PetiX S45 type steelbre (Fig. 4). It provides a spiral-like appearance since theshape o geometry o the 2D basic PetiX bre made it possibleto create a spiral-like, twisted steel bre as well during the 3Dde ormation process. Its attribution is that every wave is bent45 degrees urther out rom plane than the previous wave.

Te ourth type o the 3D bres is the PetiX S60 type steelbre (Fig. 5). Te shape o geometry is very similar to that othe PetiX S45 type steel bre, but at this type the extent o thede ormation is 60 degrees, rather than 45 degrees.

Cold drawn steel wire raw materials were used or the presentexperimental program with a nominal tensile strength o 1400N/mm2, provided by D&D Drtru Zrt., Hungary.

Fig. 4. 3D de ormed, twisted steel bre, PetiX S45 4. bra Hrom dimenziban megmunklt PetiX S45 aclszl

Fig. 5. 3D de ormed, twisted steel bre, PetiX S60 5. bra Hrom dimenziban megmunklt PetiX S60 aclszl

8/9/2019 pitanyag 2014. 4. szm

17/48



3. Experimentswo different experimental test series were conducted with

the novel type 3D geometry steel bres. A series o three-pointbending tests were carried out on prismatic SFRC specimensand a series o hinged beam pull-out tests were carried out onspecial design prismatic mortar specimens with cast-in hinges.

For the three-point bending tests 7070250 mm prismaticSFRC specimens with 30 kg/m 3 steel bre dosage were used.Twelve individual specimens were prepared for each typeof the novel type 3D geometry steel bres to get reliableresults. Specimens were tested at a span of 200 mm under oneconcentrated load at mid-span. Specimens were monotonicallyloaded with static loading up to complete exural failure thatresulted the separation of the prismatic SFRC specimens totwo blocks. Corresponding load and de ection values werecontinuously recorded and the load-de ection responseswere prepared. Load-de ection responses were used for thecomparative performance analysis.

For the hinged beam pull-out tests 4040160 mmmortar prisms were prepared. During specimen preparation,three individual bres were placed at the middle sectionsof the prisms running through a polystyrene plate of 7 mmthickness. A steel hinge was placed over the middle sectionin each specimen as a load distributing element. Specimenswere tested at a span of 100 mm under one concentratedload at mid-span, located on the steel hinge. Specimens weremonotonically loaded with static loading up to complete bondfailure that resulted either the pull-out of the bres from theto two blocks or the splitting of the cover over the steel bres.Corresponding load and de ection values were continuouslyrecorded and the load-de ection responses were prepared.The bond performance can be studied by the load-de ectionresponses since bres can be considered to be loaded in puretension due to the special loading conditions provided by thehinged beam pull-out tests.

4. ResultsPresent paper summarizes the experimental results of

exclusively the 7070250 mm prismatic SFRC specimenstested in three-point bending. Load-de ection curves areindicated in Fig. 6 . The curves indicated represent the averagecurves of the individual 12 load-de ection responses recordedfor each type of steel bres. As a reference, hooked-end steel

bres (product of D&D, Hungary) were also tested. For the better visualization of the performance differences, Fig 6.b provides these averaged results with a vertical axis of relativeload that gives the load values of the specimens preparedwith the novel type steel bres as ratios of load recorded forthe reference specimens prepared with hooked-end bres atthe same de ection level. It can be clearly realized that thenovel type steel bres provide superior performance withincreased exural toughness. Due to the large waves of the

2 dimensionally deformed basic PetiX steel bre, higherexural resistance is available than that of the reference SFRC

specimens prepared with hooked-end steel bres. At 7.5 mm

mid-span de ection the difference exceeds 30%. Steel bresthat are deformed in 3D provide better performance thanthat of the 2 dimensionally deformed basic PetiX steel bre.Certain bre geometries (e.g. PetiX V45 and PetiX S60) cangive 1.4-times higher exural resistance than the referenceSFRC specimens prepared with hooked-end steel bres. It can

be also realized that the performance of PetiX V90 is inferiorwith decreased exural toughness. The behaviour is attributedto the middle wave of the bre that is bent out from the basic

bres 2D plane with 90 degrees. Such large deviation ingeometry without gradual change results a more pronouncedtendency of forming tangles of bres during mixing and mayresult an uneven distribution of bres within the concretematrix, additionally to the fact that the bres themselves aremore sensitive to rupturing within concrete cracks due to thelarge local deformations by their extreme shape.

Fig. 6. Load-deection responses o SFRC specimens 6. bra SFRC hasbok terheler-lehajls diagramjai

If one de nes exural toughness as the area under theload-de ection response then the performance of the brescan be compared by this integral type of material parameteras a general indicator, being independent of the values ofde ection. Relative exural toughness values are indicated inFig. 7 (taking 100% for the exural toughness of referenceSFRC specimens prepared with hooked-end steel bres). Thesuperior performance of both 2D and 3D PetiX steel bres is

clearly visible together with the inferior performance of PetiXV90. Highest exural toughness was reached by PetiX V45and PetiX S60 type 3D deformed steel bres.

8/9/2019 pitanyag 2014. 4. szm

18/48



Fig. 7. Relative toughness indices o SFRC specimens 7. bra SFRC hasbok relatv szvssgi indexei

Another suitable indicator for the enhanced bond andcorresponding increased toughness of the 3D deformed steel

bres can be the ratio of the number of the ruptured and pulled-out bres within the cross sections of cracks resulting

exural failure of the prismatic specimens. After the failure

of the specimens, the resulting two blocks of the prismswere separated from each other and the failure mode of eachindividual steel bre in the crack sections was carefullyobserved and recorded. Fig. 8 summarizes the results. The barchart indicates the ratio of the number of the ruptured bres tothe total number of bres within the cross sections of cracks(summarized for all the 12 individual prismatic specimens),given in percents. For the reference SFRC specimens preparedwith hooked-end steel bres no bre rupture was observed. All

bres were pulled-out from the concrete matrix at the crosssections of cracks. Best performance was observed for PetiXV45 and PetiX S45 type 3D deformed steel bres. It should be

highlighted here as well that the results for PetiX V90 type 3Ddeformed steel bres is misleading in this representation. Oneshould observe that the smallest number of bres was foundin the failure cross sections for V90 type 3D deformed steel

bres due to the uneven distribution of bres and the extremedeformation of this type of bre geometry resulted the highratio of ruptured bres observed and the reason is not theenhanced bond capacity of the bres.

Fig. 8. Ratio o ruptured bres in SFRC specimens 8. bra Elszakadt aclszlak arnya SFRC hasbokban

5. Conclusions

The present research and development project designed,manufactured and tested one novel type of 2D deformed steel

bre (receiving the product name of PetiX) and four different

types of 3D deformed steel bres (PetiX V45; PetiX V90;PetiX S45; PetiX S60) under laboratory conditions. It wasdemonstrated that certain types of 3D deformed steel brescan hold considerably improved the bond performance inconcrete that results increased exural toughness of SFRC

prismatic specimens tested in exure.

6. AcknowledgementsAuthors gratefully acknowledge the support of the Hungarian

Scienti c Research Fund project Durability and performancecharacteristics of concretes with novel type supplementarymaterials (OTKA K 109233). Special thanks to D&D DrtruZrt., Hungary for providing cold drawn wire raw materials andhooked-end steel bres.

References[1] Balaguru, P. N. Shah, S.P. (1992): Fiber-rein orced cement composites.

McGraw-Hill , New York, 1992, 530 p.[2] Bentur, A. Mindess, S. (2007): Fibre-rein orced cementitious composites.

2nd ed. aylor & Francis, New York, 2007, 660 p.[3] Brandt, A.M. (2008): Fibre-Rein orced Cement-based (FRC) composites

a er over 40 years o development in building and civil engineering.Composite Structures, Vol. 86, No. 1-3, pp. 39. http://dx.doi.org/10.1016/j.compstruct.2008.03.006

[4] Banthia, N. Bindiganavile, V. Jones, J. Novak, J. (2012): Fibre-rein orced concrete in precast concrete applications: research leads toinnovative products.PCI Journal , Vol. 57, No. 3, pp. 3346.

[5] Walraven, J.(2009): High per ormance ber rein orced concrete: progressin knowledge and design codes. Materials and Structures, Vol. 42,pp.12471260. http://dx.doi.org/10.1617/s11527-009-9538-3

[6] di Prisco, M. Plizzari, G. Vandewalle, L. (2009): Fibre rein orcedconcrete: new design perspectives. Materials and Structures, Vol. 42, No.9, pp. 12611281. http://dx.doi.org/10.1617/s11527-009-9529-4

[7] Kerekes, P. (2014): 2 and 3 dimensional steel bers with novel geometry:design, development and comparison o effectiveness in concrete(supervisor: Borosnyi, A.),Budapest University o echnology andEconomics (BME), Faculty o Civil Engineering, Tesis submitted or theStudents Scientic Workshop 2014, 50 p. (in Hungarian)

Re .:Kerekes, Pter Borosnyi, Adorjn:Per ormance o novel type three

dimensionally de ormed steel bres or concrete ptanyag Journal o Silicate Based and Composite Materials,


j tpus, hrom dimenziban megmunklt aclszlaktulajdonsgaieg a -f l p l d

a a b a . A a g fa a, 2 dg l a l l (P X), g fa a, 3 d bag l a l l (P X v45; P X v90; P X s45; P

s60) l f l . A lab a g la d . B a l , h g g 3 d

g a ala al l b lha a l la apad a b ba , l f ha a

l b ha l ga, a l a laa g la d ga l a .

k l a a : b , a l l, 3 d g a, pad , ha l l d g, g
http://dx.doi.org/10.1016/j.compstruct.2008.03.006http://dx.doi.org/10.1016/j.compstruct.2008.03.006http://dx.doi.org/10.1016/j.compstruct.2008.03.006http://dx.doi.org/10.1016/j.compstruct.2008.03.006

8/9/2019 pitanyag 2014. 4. szm

19/48



Utilization of repair mortarfor the loss compensation ofHungarian porous limestone

b alzS SzeMerey-kiSS Bme, Fa l f c l e g g, D pa f c ma ala d e g g G l g @g a l.

koS trk Bme, Fa l f c l e g g, D pa f c ma al a d e g gG l g a @ a l.b .h : 2014. 11. 03. r d: 03. 11. 2014. h p://dx.d . g/10.14382/ p a ag- b .2014.20

AbstractD ff ad - ad , all a a labl ( H ga ) pa a a d lab a

x d, wl d g d a w d d d ff d h lab a . tha had d ff agg ga a d b d . th a f h a h wa d a d

h fl f h d ff b d a d a a f l agg ga hp p f a . D g h lab a ha 550 p w a al da d g h en a da d . P pa a ha h gh g h a d l w ap llaa ha h H ga a p l . Add d agg ga a a h pa b l

f h all a a labl a , b a a a l a d a d a g h. n wl d g d a ha la ha al a d h d h al p p hd p l .

k w d : pa a , l , b d /agg ga a , p g h, pa b l

Balzs Szemerey-Kiss(1978) lp (mA, 2002)

P va f m sp al e g(2007), PhD (2013). G ad a d a h A ad y

f F A , D pa f r a ac va , B dap (2002). P g ad a

D g h s f P vam (2004-2007) a h B dap

u v y f t h l gy a d e , Fa lf A h . PhD a h B dap u v

f t h l gy a d e , D pa fc ma al a d e g g G

(2013). ma f ld f : ph ag lp , a al g, l

p a f p .

kos Trk(1963) g l g (ms , 1986), ms

e v al e g g a d s (1993)PhD (1995), Ds (2013), f ll p f a hB dap u v y f t h l gy, Fa l y

c v l e g g. m b f h i aA a f e g g G l g i

(P d f H ga a na al G pi a al s y f r m ha isrm

(P d f H ga a na al G pH ga a G l g al s y (P d f

e g g G l gy a d e v al G lD v ). ma f ld f : gg l gy, appl d g l gy, d g l gy,

a b a d l gy, a al g p f h ag b ld g

1. IntroductionRepair mortars or articial stones are o en used or repair

o stone monuments and artistic stone elements o acades,sculptures or statues. Tere is not a uni orm terminology orthe loss compensation o stones: the expressions o repairmortars or articial stones are used by different authors [1-4].

Previous research in this eld ocused on loss compensationmethods or stones [1,2]. Additional studies were made onthe amount o ller added to cementitious materials and theper ormance o these materials was tested [5,6]. Mortars withspecially selected llers were also tested. Both natural andarticial inorganic materials were added to mortars to improvethe physico-chemical and mechanical properties; workabilityor water retention was also analyzed [7-9]. In recent years,mechanical properties and abric analyses were also made[3,10-13]. Chemical composition o the mortars were alsoanalyzed [14-16]. Compatibility o the limestone and mortarwas studied considering strength and durability characteristics

such as mechanical resistance, water trans er properties, andphysico-chemical properties [17-20].

Fig. 1. Loss compensation with incompatible repair mortar. 1. bra Kkiegszts nem kompatibilis habarccsal.

Case studies have shown, that many o the commercially available(ready-mix) repair mortars can not be adequately used undercertain conditions and are not compatible with porous limestone[17-18] (Fig. 1). Only limited data are available rom the productproperties in catalogues, sa ety data sheets, product specications.Tere is not enough in ormation about the porosity, uid transportproperties, pore-size distribution, long-term behaviour o themortars and compatibility with porous limestones. Tere ore,

urther in ormation is needed to nd the best per ormance andcompatible mortar or the given stone. Repair mortars have beentested according to the European (EN) and American (AS M)standards and per ormance in terms o strength, porosity, hydro-technical properties were measured [12-13].

2. Materials and methodsCommercially available repair mortars are commonly usedor the restoration o stone structures made o Hungarian

oolitic limestone. Practitioners use these ready-mix mortarssince it is easy to work with them that allow a relatively astrepair. Furthermore, there are individual mortar recipes inthe restoration practice. Tese two categories o mortars (pre-mixed and site-made) were tested under laboratory conditions.Teir properties were compared with that o Hungarian porouslimestone. wo lithologies were analyzed. Te limestone wascollected rom Sskt quarry, which is located app. 30 km romthe capital o Hungary. Te two tested lithologies are:

DMF: limestone with ne-grained micro- abric (Fig. 2a),DMD: limestone with coarse-grained micro- abric (Fig. 2b).wo groups o mortars were also tested. Te rst group is

composed o different types o mortars that are commerciallyavailable repair mortars (M). Te compounds o the mortars weremade in Germany and test specimens rom ready-mix mortars

8/9/2019 pitanyag 2014. 4. szm

20/48



were prepared under laboratory conditions. 30 to50 m% o limestone sand aggregate was added tothe tested mortars to make the repair mortar morecompatible with the coarse limestone. Tis mixingo mortar with limestone sand is a commonmethod in the restoration practice in Hungary.

Te abbreviations o the tested mortars are as

ollows:M: pure repair mortars (pre-mixedmortar + water),

M30: pure mortar with 30 m% addedlimestone sand aggregate and

M50: pure mortar with 50 m% addedlimestone sand aggregate.

Second group o tested mortars containednewly mixed repair mortars with two differentcompositions. wo types o cementitiousbinders were added to the mortars:

PLM: mortar with Portland cement andhydrated lime binder,

LM: mortar with trass cement andhydrated lime binder.

Te above mentioned limestone chippings(Sskt quarry) were used or the aggregatesand. Te maximum particles size o theaggregate was 2 mm in diameter. Fabric analysiso the tested materials was per ormed accordingto EN 12407:2000 European Standard by theanalysis o thin-sections.

Fig. 2. Hugarian (Sskt), Miocene, porous limestone: (a) ne- grained, (b) coarse-grained abric.

2. bra Sskti miocn kor durva mszk: (a) nomszem, (b)durvaszem vltozata.

Te sample preparation was made underlaboratory conditions (temperature 202C,relative humidity o 5055%). Te initial curingenvironment was the same in all cases. Morethan 550 mortar specimens (303030mm insize) were prepared or uniaxial compressivetest (EN 1015-11:2000). For the microscopicanalysis thin-sections were made (thickness o30 m). Prismatic samples o 404080 mmin size were used or the capillary water uptakemeasurements, ollowing the EN 1925:2000standard. Cylindrical specimens o 10 mm indiameter were made or the pore size distributionanalyses. Te measurements were made with aCarlo Erba 2000 (GFZ Potsdam) porosimeterand the data evaluation was made by a Pascalsofware (version o 1.03). Samples were dried inoven to constant mass at 105C.

Fig. 3. Tin-section o the coarse-grained limestone (DMD) with the histogram o the pore sizedistribution o the limestone.

3. bra Durvaszem ooidos durva mszk vkonycsiszolata s pruseloszlsnak hisztogramja.

Fig. 4. Tin-section o the pure, commercial available repair mortar (M) with the histogram o the poresize distribution.

4. bra Hazai kereskedelmi orgalomban kaphat kkiegszt habarcs vkonycsiszolata s pruseloszlsnak hisztogramja.

Fig. 5. Tin-section o the commercial available repair mortar ( LM) with the histogram o the pore sizedistribution.

5. bra Kkiegszt habarcsban megjelen trasz cement s lgprusok a vkonycsiszolatban, valamint atraszcement tartalm habarcsok pruseloszlsnak hisztogramja.

Fig. 6. Tin section o the commercial available repair mortar with 50m% added limestone sand aggregate(M50) with the histogram o the pore size distribution. 6. bra Hazai kereskedelmi orgalomban kaphat kkiegszt habarcsba kevert mszk zzalk (50m%)

vkonycsiszolata s pruseloszlsnak hisztogramja.

8/9/2019 pitanyag 2014. 4. szm

21/48



3. Results3.1 Fabric

Limestone: Te micro- abric is characterized by the presenceo ooids, and the micro acies o the stone is an ooid-grainstoneto packstone. Large and visible macro pores o different size were

visualized in both selected limestones (DMD and DMF). Teporosity is characterized by the presence o intra- and inter-granularpores. According to the pore size distribution, the main pores weredetected in the range o macro-pore size (10-100 m) (Fig. 3).

Mortars(M): Te binder is mostly calcitic, and the aggregate ismostly given by small rounded quartz grains o 10500 m. Tehydraulic binder content o the samples (white portland cement)is clearly demonstrated by the unhydrated cement particles.Hydrated alite (C3S - mostly), and small amounts o belite (C2S) wasalso observed, urthermore a mixture o slaked lime was present inall o the our commercially available repair mortars (Figs. 4 to6 ).

Mortars (M, PLM, LM): Only very small visible pores weredetected under the microscope. When 30 or 50 m% o limestonesand is added to the repair mortars, the proportion o the binder-to-aggregate ratio is shi ed towards the grain dominance.

Fig. 7. Average value with the standard deviation o the compressive strength o thetested mortars. All the values measured or the limestones are shown in thezone indicated in red colour.

7. bra Habarcsok tlagos egyirny nyomszilrdsga s szrsa. A piros svbanvltak mrhetk a kt le mszkvn a nyomszilrdsgi rtkek.

Abbre- viation

UCS [MPa] Waterabsorp-tion [%]

av. realdensity(g/cm3)

day3

day7

day14

day28

day90

M 6.89 8.86 9.76 10.39 11.66 9.1 2.72M30 4.28 5.54 6.42 7.21 7.54 11.7 2.67M50 2.41 4.06 4.9 5.48 5.31 13.4 2.61PLM 3.64 5.88 6.94 7.43 6.32 5,56 2.7TLM 3.28 4.35 5.56 5.08 5.02 13.4 2.66DMF 7.60 (d 2.5) 20.5 2.69DMD 5.56 (d 2.2) 14.6 2.73

able 1. Uniaxial compressive strength (UCS), water absorption and density o thetested materials ( or abbreviations see the text).

1. tblzat A vizsglt anyagok egyirny nyomszilrdga, vz elvtele s anyagsrsge(anyagok rvidtse a szvegben).

3.2 StrengthPure repair mortars (without limestone sand aggregate) have

more than double o the strength than that o the tested porouslimestones. Added limestone sand aggregate can reduce thestrength o the mortars and it becomes closer to the strength olimestone (Fig. 7 and able 1).

Te tests have demonstrated that newly designed repairmortars 28 days a er curing (mostly LM) have similarstrength to that o porous limestone, but on a longer-term thestrength o mortar specimens has been gradually decreased.

3.3 Water absorptionTe capillary water uptake o the tested limestones is muchaster than that o the tested repair mortars (Fig. 8). Te slow

water uptake o the pure repair mortars are proportionallyincreased, when limestone aggregate was added. Te biggestincrease in capillary water uptake was measured at mortarsmade with trass cement binder. Smaller increase was detectedin water uptake at commercial available mortars with limestoneaggregate (M30, M50).

Fig. 8. Capillary water uptake in the tested materials with the pore size distribution(DMD-coarse limestone, DMF- ne limestone, M, PLM, LM testedmortars).

8. bra Kapillris vz elvtel a vizsglt anyagokban a prusmret-eloszlssal (DMD-durvaszem durva mszk, DMF- nomszem durva mszk, M, PLM, LMhabarcsok).

4. ConclusionsA close relationship between the pore structure, mechanical

properties and water absorption o repair mortars were ound.Previous experience has shown that increasing the aggregatecontent decreased the strength o the mortars and at the sametime the porosity, water absorption and capillary suction areslightly increased [19, 21]. Te lower water absorption o therepair mortars compared to natural stones is related to the very different pore size distribution. Te oolitic limestone hasexceptionally interconnected pore system with a wide rangeo pore radii, and high peaks in the pore size distributionhistograms at macro-pores with diameter o around 10-100m. Tese macro-pores are missing rom the pure mortarsand rom the mortars to which limestone aggregate is added.Among the tested repair mortars, only the mortars with trasscement binder have higher compatibility with the ooliticlimestone, but the higher amount o aggregate decreased thestrength and the workability o the mortar. It is important to

highlight that the use o trass cement increased the capillaryactivity o the mortar and at the same time decreased thestrength rom the 28th day. Repair mortars with higher strength

8/9/2019 pitanyag 2014. 4. szm

22/48



and different porosity than that o the limestone can causedamage (cracking and spalling in the original stone sur ace).Usually weak limestones have lower modulus o elasticity thanmortars made by Portland cement.

o conclude, the test results have shown that the use osignicant amount o porous limestone sand as aggregate orpre-mixed repair mortars does not guarantee the compatibilityo the repair mortar with the porous limestone, however, ithas an evidential positive effect in compatibility. Without longterm laboratory experiments, newly designed repair mortarsmight have serious compatibility problems with limestone.

5. AcknowledgmentsTis work is supported by the scientic program o the

Hungarian Scientic Research Fund (grant no. O KA PD112955).

References

[1] Griswold, J. Uricheck, S.: (1998) Loss compensation methods or stone.Journal o the American Institute o Conservation 37:89110[2] Kriston L. (2000) A k s vakolatrestaurls alapismeretei. MKE, Budapest,

113119 pp[3] Pecchioni, E. Malesani, P. Bellucci, B. Fratini, F: (2005) Articial

stones utilised in Florence historical palaces between the XIX and XXcenturies Journal o Cultural Heritage 6:227233http://dx.doi.org/10.1016/j.culher.2005.06.001

[4] Isebaert, A. Van Parys, L. Cnudde, V: (2014) Composition andcompatibility requirements o mineral repair mortars or stone A review.Construction and Building Materials 59:3950http://dx.doi.org/10.1016/j.conbuildmat.2014.02.020

[5] Benachour, Y. Davy, C. A. Skoczylas, F. Houari, H: (2008) Effect oa high calcite ller addition upon microstructural, mechanical, shrinkageand transport properties o a mortar. Cement and Concrete Research38:727736

http://dx.doi.org/10.1016/j.cemconres.2008.02.007[6] Gosselin, C. Verges-Belmin, V. Royer, A. Martinet, G.: (2009) Natural

cement and monumental restoration. Materials and Structures 42:749763 http://dx.doi.org/10.1617/s11527-008-9421-7[7] Heikal, M. El-Didamony, MH. Morsy, MS: (2000) Limestone-lled

pozzolanic cement. Cement and Concrete Research 30:18271834. http://dx.doi.org/10.1016/S0008-8846(00)00402-6[8] Blent. Y. Asim, O.: (2007) Studies on cement and mortar containing

low-calcium y ash, limestone, and dolomitic limestone. Cement &Concrete Composites 30:194201http://dx.doi.org/10.1016/j.cemconcomp.2007.07.002

[9] Ghrici, M. Kenai, S. Said-Mansour, M.: (2007) Mechanical propertiesand durability o mortar and concrete containing natural pozzolana andlimestone blended cements. Cement & Concrete Composites 29:542549

http://dx.doi.org/10.1016/j.cemconcomp.2007.04.009[10] Barsottelli, M. Cellai, G. F. Fratini, F. Manganelli, F.: (2001) Te

Hygrometric behaviour o some articial stone materials used as elementso masonry walls. Materials and Structures 34:211-216http://dx.doi.org/10.1007/BF02480590

[11] Sinan, C.: (2003) Aggregate/mortar inter ace: inuence o silica ume atthe micro- and macro-level. Cement & Concrete Composites 25:557564

http://dx.doi.org/10.1016/S0958-9465(02)00095-1[12] Pavia, S. Fitzgerald, B. reacy, E.: (2006) An assessment o lime mortars

or masonry repair. Concrete Research in Ireland Colloquium, UniversityCollege Dublin, 101108.

[13] Pavia, S. oomey, B.: (2008) Inuence o the aggregate quality on thephysical properties o natural eebly-hydraulic lime mortars. Materialsand Structures 41: 559569 http://dx.doi.org/10.1617/s11527-007-9267-4

[14] Bultrini, G. Fragala, I. Ingo, GM. Lanza, G.: (2006) Minero-petrographic, thermal and microchemical investigation o historicalmortars used in Catania (Sicily) during the XVII century A.D. AppliedPhysics. 83:529536

http://dx.doi.org/10.1007/s00339-006-3551-y[15] unoku, SS. Caner-Saltk, EN.: (2006) Opal-A rich additives used in

ancient lime mortars. Cement and Concrete Research 36:18861893

Épitőanyag 2014. 4. szám

Documents