C

O

Mw

KXa

Ab

c

ARAA

KADOCRP

1

iooccdpotpofi

Bn

w

1

ARTICLE IN PRESSG ModelULHER-3700; No. of Pages 9

Journal of Cultural Heritage xxx (2020) xxx–xxx

Available online at

ScienceDirectwww.sciencedirect.com

riginal article

echanism and effect of alkoxysilanes on the restoration of decayedood used in historic buildings

unpeng Zhou a,b, Aiqun Li a,b,c,∗, Linlin Xie a,b, Chong-Chen Wang a,b,∗, Peng Wang a,iufang Wang a

Beijing Key Laboratory of Functional Materials for Building Structure and Environment Remediation, Beijing University of Civil Engineering andrchitecture, Beijing 100044, ChinaBeijing Advanced Innovation Center for Future Urban Design, Beijing University of Civil Engineering and Architecture, Beijing 100044, ChinaSchool of Civil Engineering, Southeast University, Nanjing 210096, China

a r t i c l e i n f o

rticle history:eceived 25 August 2019ccepted 25 November 2019vailable online xxx

eywords:ncient buildingsecayed wood

a b s t r a c t

It was necessary to protect and restore the decayed wood in the ancient buildings, in which reinforcingand restoring the damaged components with in situ and non-destructed strategy was preferred, followingthe heritage protection principles. For this purposes, series experiments were designed and performed toinvestigate the reaction mechanism and effects of ethyl orthosilicate and methyl triethoxysilane towardthe decayed wood components sampled from Shuiyu Village, a historical tribe in Beijing. The resultsdemonstrated that the organosilicon layer was fabricated onto the wood cell walls by condensationreaction, in which the Si O Si bonds were formed between alkoxy groups, and the Si O C bonds were

rganosilicon compoundsonservationeaction mechanismhysico-mechanical properties

generated between cellulosic fibers and organosilicon. The organosilicon layer contributed to furtherimprove physico-mechanical performances like hydrophobicity, mechanical properties, thermal stabilityof treated wooden components, except a slight color deviation. In all, the matrix of the ethyl orthosilicateand methyl triethoxysilane is a reliable agent for retaining the decayed part of wood components, whichaccomplished the satisfied performance improvements.

© 2019 Elsevier Masson SAS. All rights reserved.

. Introduction

Wood is widely used as the main structural material of build-ngs in China with a long history, which results in a large numberf splendid and valuable architectural heritages [1]. Wood, a kindf natural biodegradable substance, will be decayed to shrinkage,rack by some organisms under suitable temperature and humidityonditions [2]. Therefore, maintenance and reinforcement for theecayed wood components of traditional wood structure are indis-ensable. The present conservation technologies for deteriorationf wood components could be divided into two categories. One israditional technology, which is to resect the decayed part of com-

Please cite this article in press as: K. Zhou, et al., Mechanism and effhistoric buildings, Journal of Cultural Heritage (2020), https://doi.org/

onents and add new wood billets on the scathed area [3,4]. Thether one is to replace the decayed parts with sawdust and carbonber, and bond with epoxy resin [5–7]. But these are controversial,

∗ Corresponding authors at: Beijing Key Laboratory of Functional Materials foruilding Structure and Environment Remediation, Beijing University of Civil Engi-eering and Architecture, Beijing 100044, China.

E-mail addresses: aiqunli@seu.edu.cn (A. Li), chongchenwang@126.com,angchongchen@bucea.edu.cn (C.-C. Wang).

https://doi.org/10.1016/j.culher.2019.11.012296-2074/© 2019 Elsevier Masson SAS. All rights reserved.

especially when the components (even they are decayed), are abun-dant in historical value, art value and skill value, and removing themdetracts from the building’s historical identity and integrity [8,9].Furthermore, the unstable new additions would also affect stabilityof building structures [10]. Therefore, it is especially important toexplore new ways that could retain the original wood components[11].

The wooden structures exposed to moisture are predominantlyeroded by fungi [12]. In particular white-rot fungi and brown-rotfungi are responsible for substantial damages of wood, which leadsto decomposition of polysaccharides within lignin and cellulosetexture of the cell walls from the lumen face, through the sec-ondary wall, until the highly lignified middle lamella [13,14]. Inthis process, the cell walls, deprived of the cellulose and lignin,irreversibly collapse, resulting in chemical and physical changes oftheir components [15], and the wood will shrinks and cracks onceunprotected. As a result, the method of consolidating decayed partof wood through the wood cell textures reinforced, which retain

ect of alkoxysilanes on the restoration of decayed wood used in10.1016/j.culher.2019.11.012

the original wood components, could be the best way to solve theproblems of wood deterioration, compared with traditional tech-nology. The method is mainly applied with infiltration consist ofpenetrating a special liquid material into wood and widely utilized

ING ModelC

2 tural H

io

itbdcot[htimabtcaw

mocccrtbmdAms

stpuicscismsihomto

2

2

opiBo

ARTICLEULHER-3700; No. of Pages 9

K. Zhou et al. / Journal of Cul

n conservation of decayed wood for their excellent performancesf improving physical and mechanical properties [16,17].

However, for the decayed wood components of ancient build-ngs, some concerns should be taken into account when applyinghe method: hydrophobicity, mechanical properties, thermal sta-ility and esthetics. The reasons are listed as the following: (1) Theegradation of wood is mainly caused by excessive moisture whichould affect wood dimensional stability and result in shrinkager swelling. Therefore, the improvement of hydrophobicity seemso be the optimum solution to prevent wood from being decayed13]. (2) As the underlying problem of moisture in decayed woodas been solved, it is essential to restore mechanical strength, sohat the wood components could continue to perform their abil-ty in decorative and structural functions. (3) In addition to restore

echanical properties of decayed parts, the method applied shouldlso protect the wood from fire threat, which is crucial to the dura-ility of components and the long-term retention. (4) According tohe authenticity principles of heritage protection, a purpose of theonservation is, first and foremost, to preserve the original appear-nces of precious wood heritages [13]. Thus, the esthetics of treatedood should also be concerned.

However, up to now, the attentions in related research wereuch paid to dimensional stabilization, mechanical performance

r overall appearances of treated wood [11,18–20], but less to theomprehensive study on the reaction mechanism and the con-erns above that are especially important for conserving decayedomponents of ancient buildings. Herein, this research met theequirements in the field of restoration with a systematic inves-igation on the consolidation effects of decayed wood of ancientuildings in terms of hydrophobicity, mechanical properties, ther-al stability and esthetics, based on the idea of reserving the

ecayed wood components and consolidating decayed parts in situ.s well, the reaction mechanism of restoration was clarified viaeans of various characterizations like scanning electron micro-

cope, Fourier transform infrared spectroscopy and so on.Generally, some organosilicon compounds were adopted to con-

olidate the decayed wood parts due to their mechanical properties,hermal stability, along with the ability to stabilize wood mor-hology by intensifying the cell walls [18,21]. In addition, theysually consist of several organic functional groups and read-

ly hydrolysable alkoxy groups [22,23]. Their polymerizations areomposed of hydrolysis and condensation reactions called theol–gel process [24,25]. In the process, hydrolyzed alkoxy groupsan combine with other hydrolyzed alkoxy groups by establish-ng stable Si O Si bonds, producing a three-dimensional networktructure of gel, after alkoxysilanes molecules are activated byoisture [26]. As the water molecules evaporate, the gel will

hrivel to different types of silicon flakes with both organic andnorganic properties like excellent scirrhosity, climate resistance,ydrophobicity and thermal stability [27]. In this work, the mixturef two typical alkoxysilanes like the ethyl orthosilicate (TEOS) andethyl triethoxysilane (MTES) was selected as consolidation agent

o investigate their ability and mechanism to restore decayed woodf ancient buildings.

. Materials and methods

.1. Wood samples

The objects in the research were small cuboid samples cutut from the decayed parts of wood rafters that were made of

Please cite this article in press as: K. Zhou, et al., Mechanism and effhistoric buildings, Journal of Cultural Heritage (2020), https://doi.org/

oplar wood and extracted from some ancient buildings locatedn the Shuiyu Village. The village, situated in Fangshan District,eijing, is a famous historical tribe with the history more thanne hundred years, which was selected into the list of first batch

PRESSeritage xxx (2020) xxx–xxx

traditional villages of Beijing in March 2018. The decayed woodsamples, with the measured density of only 37% of sound woodand water content of 8.1%, were divided into two treatment groups.The samples applied for mechanism characterizations, thermal sta-bility and esthetic tests with dimension of 5 × 20 × 20 mm (radial(R) × tangential (T) × longitudinal (L) direction) were labeled asgroup A. The samples used for hydrophobicity and mechanicaltests with dimension of 15 × 15 × 15 mm (radial (R) × tangential(T) × longitudinal (L) direction) were labeled as group B (Fig. 1).

2.2. Chemical and reagents

Both TEOS and MTES were utilized in this experiment. In addi-tion to the general properties of silicon materials, TEOS and MTESdisplay the good performance on their own distinct characteristics.TEOS was selected due to its cross-linking and reinforce propertiesthat results from hydrolysable ethoxy groups. MTES was selectedfor its potential hydrophobic characteristic (methyl groups) andtackify properties that originate from its active groups and chemi-cal activities (Fig. S1). As well, the performance of the combinationof two agents is very outstanding, which could meet the criti-cal requirements for heritage restoration [28,29]. Furthermore, thetwo agent are most likely to be applied in practice in a relativelyshort time, for their controllable reaction procedure and massiverelevant research data [30–32].

2.3. Characterizations

The weights of pre- and post-treated samples were measuredto calculate weight percent gain (WPG) of wood samples, followingEq. (1):

WPG = Wt − W0

W0× 100 (1)

Some characterizations techniques were utilized to investigatethe reaction mechanism of restoration for decayed wooden compo-nents. The microcosmic appearances of wood samples were imagedwith the scanning electron microscope (Hitachi TM3030 Plus) cou-pled with the EDS detector of Model 550i (to obtain the X-raymaps for silicon) at magnification of 1200–1800 times. Changesin chemical compositions of wood samples were probed by Fouriertransform infrared spectroscopy (Nicolet 6700 Spectrometer) in thewave-number range of 4000–400 cm−1. Changes in bond energyof wood samples were detected with X-ray photoelectron spec-troscopy (Thermo ESCALAB 250XI).

2.4. Property tests

Some instruments were used to assess the changes of physico-mechanical properties of treated wood. Hydrophobicity of woodsamples was measured according to contact angle of water dropleton the surfaces of wood by the optical contact angle & interfacetension meter (KINO SL2OOKS). Mechanical properties of woodsamples were evaluated under pressure of 8 N by performancetester (DZS-III) of hard and brittle material. Thermal stability ofwood samples was measured with thermal gravimetric analyzer(Bo Yuan DTU-3) in the range of 25–650◦ at a heating rate of 10◦

/min with air streaming. Esthetic variations of wood samples werecalculated by precise color reader (FRU WR-18) with �E value.

2.5. Restoration experiment

ect of alkoxysilanes on the restoration of decayed wood used in10.1016/j.culher.2019.11.012

Ethyl orthosilicate, methyl triethoxysilane, alcohol and waterwere mixed in a molar ratio (0.66:1:4.49:1.66), with pH beingadjusted to 2.0 by 0.1 mol/L HCl solution. After the mixed solu-tion was stirred for 10 min at 20 ◦, pH was adjusted to 7.0 with

ARTICLE IN PRESSG ModelCULHER-3700; No. of Pages 9

K. Zhou et al. / Journal of Cultural Heritage xxx (2020) xxx–xxx 3

F ollecting rafters; (c) overall appearance of a decayed rafter; (d) cross-section of a decayedr

0t2slisat

3

3

tMt

S

(

(

(

ifiwnhm

Table 1WPG of treated wood samples.

Sample no. 1 2 3 4 5 Average

ig. 1. Decayed poplar wood: (a) overlooking view of Shuiyu Village; (b) location of cafter; (e) cuboid wood samples cut out from decayed rafters.

.5 mol/L NaOH solution to form consolidation solution. Then,he samples were immersed in the consolidation solution for0 min. After that, the treated wood samples were taken out of theolution and air-dried at room temperature for two days, the fol-owing characterizations and physico-mechanical tests were madendividually. For comparison, equal amounts of untreated woodamples (decayed without impregnating the treatment solution)nd sound wood samples (non-decayed without impregnating thereatment solution) were characterized and tested.

. Results and discussion

.1. Explorations of reaction mechanism

The sol–gel chemistry of TEOS and MTES have been well inves-igated [33]. After being activated by water molecules, TEOS and

TES will experience a series of hydrolysis and condensation reac-ions as Equations (1.1), (1.2), (2.1)–(2.3) [34,35].

(1) Hydrolysis reaction

i(OEt)4 + 4H2O = Si(OH)4 + 4EtOH (1.1)

Si|

CH3

(OEt)3 + 3H2O = Si|

CH3

(OH)3 + 3EtOH (1.2)

(2) Condensation reaction

HO)3SiOH + HOSi(OH)3 = (HO)3Si–O–Si(OH)3 + H2O (2.1)

OH)2 Si|

CH3

OH + HO Si|

CH3

(OH)2 = (HO)2 Si|

CH3

− O − Si|

CH3

(OH)2 + H2O (2.2)

OH)2 Si|

CH3

OH + HOSi(OH)3 = (HO)2 Si|

CH3

− O − Si(OH)3 + H2O

(2.3)

In fact, hydrolysis and condensation reactions were carried outn a complicated manner. The condensation will start once therst ethoxy groups are hydrolyzed and the water condensation

Please cite this article in press as: K. Zhou, et al., Mechanism and effhistoric buildings, Journal of Cultural Heritage (2020), https://doi.org/

ill be accompanied by alcoholic condensation [36,37]. For conve-ience in discussion, only ideal hydrolysis (all the alkoxy groups areydrolyzed) and water condensation are considered in the experi-ent.

WPG (%) 83.67 75.92 89.35 81.13 86.44 83.30

Furthermore, the hydrolysis rate of siloxane have been provedto be faster than condensation rate in acidic condition, while therates are opposite under basic condition [38]. The hydrolysis ofMTES was faster than that of TEOS in acidic condition, however,the condensation of MTES becomes slower under basic cataly-sis [39]. Therefore, the sol–gel progress in this experiment couldbe explained as follows: TEOS and MTES would be hydrolyzed toform Si(OH)4 and CH3Si(OH)3 separately with acid catalysis afterthey were mixed with water. When the pH value of solution wasadjusted to 7.0, Si(OH)4 groups would firstly condensate into smalloligomers which were contributed by water condensations in theform of Si O Si bonds [29]. (CH3)Si(OH)3 was late added on thesurfaces of oligomers via the polycondensation between hydroxylgroups, considering that it condensates more slowly than that ofSi(OH)4 [39]. Lots of oligomers containing exposed CH3 groups werethen produced, as illustrated in Fig. 2. After treatment solution ofoligomers was infiltrated into decayed wood, oligomers would fur-ther cross link with other oligomers via Si O Si bonds, and withhydroxyl groups of wood cellulose via Si O C bonds and hydrogenbonds [40,41]. As a result, three-dimensional network structure oforganosilicon gel were formed in wood structure. The gel layer fab-ricated onto wood cell walls would finally be established after theevaporations of water and alcohol [42] (Fig. 2).

WPG was generally applied to assess changes in weight of sam-ples [43,44]. As listed in Table 1, it was observed that the WPGraises with the average value exceeding 83%, suggesting that a largenumber of products were generated in the wood structure.

Scanning electron microscopy was applied to explore the mor-phology and structure changes of treated wood [45]. As shown inFig. 3(a)–(c), the cell walls of untreated wood attacked by fungi arethin and flabby in the terms of pit area, cross section and longi-tudinal section. However, great changes have taken place in thestructures after the wood were treated. Fig. 3(d)–(f) shows thatsome flaky gels could be observed in the cell gaps and some are vis-ible on wood cell surfaces, as was assumed in reaction mechanism.The attached gel layer made the cell walls of treated wood thicker

ect of alkoxysilanes on the restoration of decayed wood used in10.1016/j.culher.2019.11.012

than those of the untreated wood and potentially more stable andtougher, which would improve the physico-mechanical propertiesof wood components in the ancient buildings. The EDS elemental

ARTICLE IN PRESSG ModelCULHER-3700; No. of Pages 9

4 K. Zhou et al. / Journal of Cultural Heritage xxx (2020) xxx–xxx

Fig. 2. The proposed possible mechanisms of organosilicon gel layer formed in wood.

Fig. 3. SEM images of wood samples: (a)–(c) pit area, longitudinal section and cross sectioof treated wood samples.

Table 2Elemental analysis of the scanned area.

Elements. Intensity(c/s) Atomic % Atomic ratio Concentration

C 7.25 40.155 0.9996 28.750O 17.39 40.172 1.0000 38.314Si 55.90 19.672 0.4897 32.935

dawom5c

vibrations of Si O Si bonds unidentified in spectra of sound anduntreated wood [53,54]. This confirms the formation of silicon net-

Total 100.000 100.000

etection revealed that the element of silicon originating from TEOSnd MTES is observed on the surface of flaky gels (Fig. 4(a) and (b)),hich further confirms that the flaky gels are mainly composed

f silicon material. Elemental analysis of the scanned area by EDS

Please cite this article in press as: K. Zhou, et al., Mechanism and effhistoric buildings, Journal of Cultural Heritage (2020), https://doi.org/

apping (Table 2) shows that the intensity and atomic of silicon are5.90% and 19.672% respectively, proving the high content of sili-on element. Beside, silica was evenly formed in cell wall of wood

n of untreated wood samples; (d)–(f) pit area, longitudinal section and cross section

according to the EDS elemental detection, which was affirmed bythe detected element silicon in the cross section (Fig. 4(c) and (d)).

FTIR was used to testify changes in chemical compositions ofwood [46,47]. Fig. 5 shows FT-IR spectra of sound, treated anduntreated wood samples. In FTIR spectrum of untreated wood, theband at 3423 cm−1 was assigned to the O H of cellulose and lignin[48,49]. The spectrum of treated wood shows a band at 1100 cm−1

corresponding to the asymmetric stretching vibration of Si O Sibridges [50,51], which overlaps with C O and C O C stretchingbands in the spectra of sound and untreated wood [38,52]. More-over, in spectrum of treated wood, bands occurring at about 792 and449 cm−1 were ascribable to the symmetric stretching and bending

ect of alkoxysilanes on the restoration of decayed wood used in10.1016/j.culher.2019.11.012

work in the treated wood. Besides, the spectrum of treated woodshows weak intensity bands at 1275 and 2978 cm−1 which were

ARTICLE IN PRESSG ModelCULHER-3700; No. of Pages 9

K. Zhou et al. / Journal of Cultural Heritage xxx (2020) xxx–xxx 5

Fig. 4. Mapping of element Si, O, C of treated wood samples: (a) the scanned area of longitu(c) the scanned area of cross section; (d) element distribution in the scanned areas of cro

aSoCmssutbcsas

hydrophobicity is measured by contact angle of water droplet on

Fig. 5. FTIR spectra of wood samples.

ssigned to absorption vibrations of Si C bonds and CH3 units ini CH3 [52,55]. The CH3 units could be valid for the improvementf materials’ hydrophobicity because of its non-polar trait [56,57].ompared with the hybrid wood (wood powder mechanically beingixed with gel), Si O Si band in the spectrum of treated wood

lightly shifts to 1062 cm−1. As shown in Figs. S2 and 5, the bluehifts of Si O Si bond peak from 1050 cm−1 in sound wood andntreated wood to ca. 1062 cm−1 (Treated 1–Treated 5) indicatedhat organosilicon layer was likely linked with surfaces of cell wallsy the intermolecular hydrogen bonding interactions [58,59]. Itould be concluded the combination of organosilicon layer with the

Please cite this article in press as: K. Zhou, et al., Mechanism and effhistoric buildings, Journal of Cultural Heritage (2020), https://doi.org/

urfaces of cell walls. Furthermore, treated wood spectrum show characteristic band at 956 cm−1, which can be assigned to thetretching vibrations of Si OH [60–62].

dinal section; (b) element distribution in the scanned areas of longitudinal section;ss section.

In the assumed reaction mechanism, the combination oforganosilicon with the surface of cell walls via Si O C bondsin a condensation reaction were suggested yet [41,63]. The X-ray photoelectron spectroscopy spectra of mechanical hybrid andtreated wood were further analyzed to prove this. As depictedin Fig. 6(a), the surface atomic intensity and building energy ofSi2p, C1s and O1s of treated wood have changed, confirming thatthe organosilicon treatment led to significant chemical changesin wood structure. There is also a shift from 103.03 to 103.35 eVin the binding energy of Si2p (Fig. 6(b)), indicating the change ofexternal binding energy of Si element. The binding energy value(103.35 eV) of Si2s of treated wood is coincident with the existenceof a silicon atom in Si O C bonds [59]. The additional evidencefor the existence of Si O C bonds was given by the XPS analysisof C1s spectra. Fig. 6(c) and (d), representing XPS C1s spectrum ofmechanical hybrid and treated wood respectively, show three typesof carbon bonds C1 ( C H or C C), C2 ( C O), C3 ( O C O ),C4 ( C Si) [63,64]. The shift of the binding energy of the C2 compo-nent, from 285.291 to 286.015 eV for mechanical hybrid and treatedwood, was considered to be evidence of the presence of C O Sistructures [59].

3.2. Physico-mechanical properties test

The changes in microstructures and chemical compositions ofwood cells would normally exert impact on properties of wood,thus the hydrophobicity, mechanical properties, thermal stabilityand esthetics were evaluated.

Hydrophobicity of solid surfaces is influenced by both thesurface morphology and chemical composition. Generally, therougher surface structure or the lower surface energy contributedto the better hydrophobicity of surface [65–67]. In addition, the

ect of alkoxysilanes on the restoration of decayed wood used in10.1016/j.culher.2019.11.012

solid surfaces [68,69]. The solid surfaces is classified as hydrophobicwhen the contact angle is greater than 90◦, and hydrophilic whenthe contact angle is less than 90◦ [49,70]. Fig. 7 shows the contact

ARTICLE IN PRESSG ModelCULHER-3700; No. of Pages 9

6 K. Zhou et al. / Journal of Cultural Heritage xxx (2020) xxx–xxx

Fig. 6. XPS spectra of mechanical hybrid and treated sample: (a) survey spectra; (b) migration of Si2s binding energy; (c) C1s spectrum of hybrid samples; (d) C1s spectrumof treated wood samples.

Ft

atttditunp

contrary, the water contact value of treated wood remains sta-

ig. 7. Contact angle trends of treated, untreated and sound wood samples withime.

ngle trends of treated, untreated and sound wood samples withime. Initially, the contact angles of three types samples are greaterhan 105◦, displaying high hydrophobicity of the surfaces. However,he contact angles of untreated and sound wood samples graduallyecrease with time prolonging and reach almost 60◦ at 70 s, which

ndicates the transformation of wood surfaces from hydrophobico hydrophilic. A reasonable explain is that the hydrophobicity of

Please cite this article in press as: K. Zhou, et al., Mechanism and effhistoric buildings, Journal of Cultural Heritage (2020), https://doi.org/

ntreated and sound wood was provided by the geometric rough-esses of wood surfaces which were stemmed from manufacturingrocess of samples, as shown in Fig. 8. Such hydrophobicity is unsta-

Fig. 8. Typical surface morphologies of treated, untreated and sound wood samples.

ble for the instability and capillarity of the rough surfaces. On the

ect of alkoxysilanes on the restoration of decayed wood used in10.1016/j.culher.2019.11.012

ble, because the hydrophobicity of treated wood was provided byorganosilicon layer fabricated onto wood cell walls. The organosili-con layer sealed the pores of wood surfaces and had a large number

IN PRESSG ModelC

tural Heritage xxx (2020) xxx–xxx 7

owawivwt

wanacituoitftott

vsbdpt

sdTitf

Fig. 9. The thermogravimetric (TG) and differential thermal (DT) analyses on

TR

ARTICLEULHER-3700; No. of Pages 9

K. Zhou et al. / Journal of Cul

f methyl groups fabricated onto its surfaces (Fig. 8), providing theood stable hydrophobicity that was also affirmed by the water

dsorption test determination. As illustrated in Fig. S3, the treatedood exhibited a very low rate of weight percentage gain dur-

ng 8 d of immersion in distilled water, achieving almost constantalue after 160 h, which is consistent with sound wood. In contrast,eight percentage gain of untreated wood is 2 times higher than

hat of treated wood because of its loose cellular structure.The mechanical properties of treated, untreated and sound

ood samples was investigated by indentation tests and the resultsre shown in Table 4. The parameters reveal that indentation hard-ess, recovery resistance and elastic modulus of untreated woodre separately about 10%, 1.26% and 3.56% of sound wood, indi-ating great losses of strength because of the degradation. Thendentation hardness, recovery resistance and elastic modulus ofhe treated wood were 15, 27 and 20 times higher than those ofntreated ones, which can maintain or even enhance the functionsf the wood components for the ancient buildings. The significant

ncrease in the mechanical properties of treated wood could be dueo the excellent mechanical performances of organosilicon layerabricated onto the wood cell walls. As Table 3 shows, the indenta-ion hardness, recovery resistance, elastic modulus and toughnessf organosilicon particles are 0.42, 11.99, 1.39 and 15.77 respec-ively, and the corresponding mechanical performances are higherhan that of sound wood.

It should be noticed that the energy dissipation is an oppositeariable to toughness in indentation test, and when energy dis-ipation gets higher, the toughness gets worse. Therefore, it cane concluded from Table 3 that treated wood demonstrates greatecrease in toughness with an increase of 30% in energy dissi-ation rate, though reinforced by organosilicon particles of highoughness.

The thermal stability of treated, untreated and sound woodamples was tested via the thermogravimetric analyses (TGA) andifferential thermal analyses (DTA), as depicted in Fig. 9. On the

Please cite this article in press as: K. Zhou, et al., Mechanism and effhistoric buildings, Journal of Cultural Heritage (2020), https://doi.org/

GA curves of untreated and sound wood, there is a rapid decreasen their weight by flaming between 250 and 500 ◦. By contrast,he treated wood displays a much lower rate of weight decreaseor the existence of organosilicon. From 500 to 650 ◦, more resid-

able 3esults of mechanical tests of treated, untreated and sound wood samples.

Samples groups Indentation hardness (GPa) Recovery resist

Sound wood 0.0034 3.8265

Untreated wood 0.00035 0.0483

Treated wood 0.00575 1.35755

Organosilicon particles 0.42 11.99

Fig. 10. Overall appearances of untrea

treated, untreated and sound wood samples.

uals of treated wood are remain than that of untreated and soundwood whose residuals gradually decrease to ca. 10% and 2% of theiroriginal weight. In addition, rates of weight decreases of untreatedand sound wood are roughly similar, but the differences are gener-ated by the evaporation of crystal water and the flaming of intactorganic components in sound wood at 25–100 and 350–400 ◦. Onthe DT curves of untreated and sound wood, the striking exothermicpeaks for flaming are observed separately at 250–325, 360–440 and450–550 ◦. However, on the DT curve of treated wood, these strik-ing peaks are much broadened and weakened, shifting its peaks tothe higher temperature.

The results suggest that the fire-resistance property of treatedwood was enhanced. The enhancements in fire-resistance canbe attributed to the organosilicon layer observed in Fig. 3, fororganosilicon layer formed in wood cell structures would meltedand then covered the surfaces of wood cells with the temperatureincreasing. This organosilicon layer could stop oxidation and heattransfer from proceeding into the inner wood cell walls, and thus

ect of alkoxysilanes on the restoration of decayed wood used in10.1016/j.culher.2019.11.012

achieved a significant enhancement in fire-resistance properties[43,71].

ance (GPa) Elastic modulus (GPa) Energy dissipation rate (%)

0.0773 77.460.00275 48.240.0592 57.301.39 15.77

ted and treated wood samples.

ARTICLE ING ModelCULHER-3700; No. of Pages 9

8 K. Zhou et al. / Journal of Cultural H

Table 4The changes in the color values for the samples after treatment.

Sample no. 1 2 3 4 5 Average

toahhcwwTcupitto

4

asbTtcmwiom

dgwitf

pblacieaBm

wedwifrt

[

[

[

[

[

[

[

[[

[

[

[

[ment with silanes, Holzforschung 60 (2006) 40–46.

�E 6.81 5.61 4.11 5.16 4.53 5.24

The aims of organosilicon treatment are not only to improvehe hydrophobicity, mechanical properties and thermal stabilityf decayed wood components, but also to preserve the originalppearances that is critical in the terms of authenticity principles oferitage protection. Because few changes in dimensional of woodave occurred in this experiment, color changes were exclusivelyonsidered in esthetic measurement. Compared with untreatedood samples, several treated wood samples of the same batchere used to exhibit color changes via variable values of �E in

able 4. The values of �E ranging from 4.11 to 6.81 suggests that theolor changes of treated wood are slight different from that of thentreated wood, and acceptable from the perspective of heritagerotection. It can be observed that the treated wood retains orig-

nal natural coloration of wood with minor color deviations fromhe untreated wood (Fig. 10). The color changes could be attributedo the transparency and refraction of organosilicon layer fabricatednto the wood cell walls.

. Conclusions

In this study, the effects of organosilicon compounds of MTESnd TEOS on the hydrophobicity, mechanical properties, thermaltability and esthetics of decayed wood components in ancientuildings were investigated with the reaction mechanism explored.he experimental results provided a great deal of valuable informa-ion for applying the organosilicon compounds in decayed woodomponent conservation. As expected, the organosilicon treat-ent produced tremendous changes to morphological structures ofood cells, which led to significant improvements of performances

n different properties of decayed wood and met the requirementf retaining the original wood components with value and infor-ation reserved.

The exploration of mechanism by SEM-EDS, FTIR and XPSemonstrated that a large numbers of organosilicon flakes wereenerated in the cell gaps, cell surfaces as well as inside the cellalls, which led to a protective organosilicon layer. The organosil-

con layer was condensated by the hydrolyzed alkoxy groupshrough forming siloxy bonds of Si O Si, and fabricated on the sur-aces of wood cells by hydrogen bonds and stable bonds of Si O C.

The formation of organosilicon layer further improved thehysico-mechanical properties of wood. Firstly, a stable hydropho-ic surfaces on wood cell walls was contributed from organosilicon

ayer which could sealed the pores of wood surfaces and provided large number of hydrophobic methyl groups. Secondly, the woodell structures were greatly reinforced by the organosilicon layer forts excellent mechanical properties. Thirdly, fire-resistant was alsonhanced by the organosilicon layer which could prevent oxidationnd heat transfer from proceeding into the inner cell structures.esides, the treated wood retained its natural coloration, despiteinor color deviations from the untreated wood.

It should be underlined that performances of decayed woodere improved obviously, differences in physico-mechanical prop-

rties significantly existed between treated and sound wood. Theifferences could make the treat-decayed parts of wood discordantith other sound parts when the external temperature and humid-

Please cite this article in press as: K. Zhou, et al., Mechanism and effhistoric buildings, Journal of Cultural Heritage (2020), https://doi.org/

ty change. Therefore, the organosilicon treatment could be suitableor wooden component, only after the differences in properties areigorously investigated. Further, it was visualized in the mechanicalests that the organosilicon layer was not available for toughness

[

PRESSeritage xxx (2020) xxx–xxx

improvement and needs to be ameliorated by the introduction oftoughening materials in the future studies.

Acknowledgements

This work was supported by the National Natural Science Foun-dation of China (51978033), Project of Construction of InnovationTeams and Teacher Career Development for Universities and Col-leges Under Beijing Municipality(IDHT20170508), Beijing ScholarsProject, Funded Project of the Beijing Advanced Innovation Centerfor Future Urban Design (UDC2016030200).

Appendix A. Supplementary data

Supplementary data associated with this article can be found, inthe online version, at https://doi.org/10.1016/j.culher.2019.11.012.

References

[1] Q. Chun, Z. Yue, J. Pan, Experimental study on seismic characteristics of typicalmortise-tenon joints of Chinese southern traditional timber frame buildings,Sci. China Technol. Sci. 54 (2011) 2404.

[2] S.V. Glass, S.L. Zelinka, Moisture Relations and Physical Properties of Wood,Wood Handbook: Wood as an Engineering Material: Chapter 4. Centennial ed.General Technical Report FPL; GTR-190, vol. 190, US Dept. of Agriculture, ForestService, Forest Products Laboratory, Madison, WI, 2010, pp. 4.1–4.19.

[3] D. D’Ayala, H. Wang, Conservation practice of chinese timber structures: ‘nooriginality to be changed’or ‘conserve as found’, J. Architect. Conserv. 12 (2006)7–26.

[4] K.E. Larsen, N. Marstein, Conservation of Historic Timber Structures: An Eco-logical Approach, Butterworth-Heinemann Oxford, 2000.

[5] A.S. Wheeler, A. Hutchinson, Resin repairs to timber structures, Int. J. Adhes.Adhes. 18 (1998) 1–13.

[6] X. Zhang, Y. Sun, M. Liu, R. Yang, Research progress on repair and reinforcementof beams in timber building, in: MATEC Web of Conferences, EDP Sciences,2019, p. 01019.

[7] R.A. Young, Historic Preservation Technology: A Primer, Wiley, 2008.[8] V.C. ICOMOS, International Charter for the conservation and restoration of

monuments and sites, in: First Developed at the 2nd International Congressof Architects and Technicians of Historic Monuments, 1964.

[9] P.B. Lourenc o, Recommendations for restoration of ancient buildings and thesurvival of a masonry chimney, Constr. Build. Mater. 20 (2006) 239–251.

10] P. Lavisci, S. Berti, B. Pizzo, P. Triboulot, R. Zanuttini, A shear test for structuraladhesives used in the consolidation of old timber, Eur. J. Wood Wood Prod. 59(2001) 145–152.

11] D. Henriques, J. de Brito, S. Duarte, L. Nunes, Consolidating preservative-treatedwood: combined mechanical performance of boron and polymeric products inwood degraded by Coniophora puteana, J. Cult. Herit. 15 (2014) 10–17.

12] C.A. Clausen, Biodeterioration of Wood, Wood Handbook: Wood as an Engi-neering Material: Chapter 14. Centennial ed. General Technical Report FPL;GTR-190, vol. 190, US Dept. of Agriculture, Forest Service, Forest Products Lab-oratory, Madison, WI, 2010, pp. 14.11–14.16.

13] M. Broda, B. Mazela, A. Dutkiewicz, Organosilicon compounds with variousactive groups as consolidants for the preservation of waterlogged archaeolog-ical wood, J. Cult. Herit. 35 (2019) 123–128.

14] Á.T. Martínez, M. Speranza, F.J. Ruiz-Duenas, P. Ferreira, S. Camarero, F. Guillén,M.J. Martínez, A. Gutiérrez Suárez, J.C.d. Río Andrade, Biodegradation of Ligno-cellulosics: Microbial, Chemical, and Enzymatic Aspects of the fungal Attack ofLignin, 2005.

15] C.G. Björdal, Microbial degradation of waterlogged archaeological wood, J. Cult.Herit. 13 (2012) S118–S122.

16] W.J. Homan, A.J. Jorissen, Wood modification developments, Heron 49 (2004)360–369.

17] R. Goettsche, H.-N. Marx, Wood preservatives, Google Patents (1988).18] C. Mai, H. Militz, Modification of wood with silicon compounds. Inorganic sil-

icon compounds and sol–gel systems: a review, Wood Sci. Technol. 37 (2004)339–348.

19] D.W. Grattan, R.W. Clarke, Conservation of waterlogged wood, in: Conservationof Marine Archaeological Objects, Elsevier, 1987, pp. 164–206.

20] F. Lionetto, M. Frigione, Effect of novel consolidants on mechanical and absorp-tion properties of deteriorated wood by insect attack, J. Cult. Herit. 13 (2012)195–203.

21] T. Hübert, M. Shabir Mahr, Sol–gel wood preservation, Handb. Sol–Gel Sci.Technol. (2016) 1–52.

22] S. Donath, H. Militz, C. Mai, Creating water-repellent effects on wood by treat-

ect of alkoxysilanes on the restoration of decayed wood used in10.1016/j.culher.2019.11.012

23] K. Mojsiewicz-Pienkowska, M. Jamrógiewicz, K. Szymkowska, D.Krenczkowska, Direct human contact with siloxanes (silicones)–safetyor risk part 1. Characteristics of siloxanes (silicones), Front. Pharmacol. 7(2016) 132.

ING ModelC

tural H

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[silica nanoparticles modified with long-chain alkylsilane, Holzforschung 67

ARTICLEULHER-3700; No. of Pages 9

K. Zhou et al. / Journal of Cul

24] D. Levy, M. Zayat, The Sol–Gel Handbook, 3 Volume Set: Synthesis, Characteri-zation, and Applications, John Wiley & Sons, 2015.

25] C. Mai, H. Militz, Modification of wood with silicon compounds. Treatmentsystems based on organic silicon compounds—a review, Wood Sci. Technol. 37(2004) 453–461.

26] S. Donath, H. Militz, C. Mai, Wood modification with alkoxysilanes, Wood Sci.Technol. 38 (2004) 555–566.

27] S. Sutha, R. Yuvakkumar, V. Rajendran, R. Palanivelu, Effect of thermal treat-ment on hydrophobicity of methyl-functionalised hybrid nano-silica particles,Mater. Lett. 90 (2013) 68–71.

28] A. Jitianu, G. Amatucci, L.C. Klein, Organic–inorganic sol–gel thick films forhumidity barriers, J. Mater. Res. 23 (2008) 2084–2090.

29] H. Yang, P. Pi, Z.-Q. Cai, X. Wen, X. Wang, J. Cheng, Z.-r. Yang, Facile preparationof super-hydrophobic and super-oleophilic silica film on stainless steel meshvia sol–gel process, Appl. Surf. Sci. 256 (2010) 4095–4102.

30] L. Klein, G. Garvey, Tetraethylorthosilicate (Teos), MRS Online Proc. LibraryArch. 32 (1984).

31] J.C. Ro, I.J. Chung, Sol–gel kinetics of tetraethylorthosilicate (TEOS) in acid cat-alyst, J. Non-Cryst. Solids 110 (1989) 26–32.

32] S. Sakka, K. Kamiya, The sol–gel transition in the hydrolysis of metal alkoxidesin relation to the formation of glass fibers and films, J. Non-Cryst. Solids 48(1982) 31–46.

33] B. Unger, M. Bücker, S. Reinsch, T. Hübert, Chemical aspects of wood modifica-tion by sol–gel-derived silica, Wood Sci. Technol. 47 (2013) 83–104.

34] A. Buckley, M. Greenblatt, The sol–gel preparation of silica gels, J. Chem. Educ.71 (1994) 599.

35] K. Kamiya, T. Yoko, K. Tanaka, M. Takeuchi, Thermal evolution of gels derivedfrom CH3Si (OC2H5)3 by the sol–gel method, J. Non-Cryst. Solids 121 (1990)182–187.

36] H. Schmidt, H. Scholze, A. Kaiser, Principles of hydrolysis and condensationreaction of alkoxysilanes, J. Non-Cryst. Solids 63 (1984) 1–11.

37] M.-C.B. Salon, M. Abdelmouleh, S. Boufi, M.N. Belgacem, A. Gandini, Silaneadsorption onto cellulose fibers: hydrolysis and condensation reactions, J. Col-loid Interface Sci. 289 (2005) 249–261.

38] M. Musolino, F. Aricò, P. Tundo, An innovative and sustainable approach to spraypaint graffiti removal from Istrian stone through the silica sol-gel chemistry: apreliminary assessment, J. Cult. Herit. 36 (2019) 268–274.

39] M.J.V. Bommel, T.N.M. Bernards, A.H. Boonstra, The influence of the additionof alkyl-substituted ethoxysilane on the hydrolysis—condensation process ofTEOS, J. Non-Cryst. Solids 128 (1991) 231–242.

40] M. Broda, B. Mazela, K. Radka, Methyltrimethoxysilane as a stabilising agentfor archaeological waterlogged wood differing in the degree of degradation, J.Cult. Herit. 35 (2019) 129–139.

41] M. Bücker, C. Jäger, D. Pfeifer, B. Unger, Evidence of Si–O–C bonds in cellu-losic materials modified by sol–gel-derived silica, Wood Sci. Technol. 48 (2014)1033–1047.

42] Y. Xie, C.A. Hill, D. Sun, Z. Jalaludin, Q. Wang, C. Mai, Effects of dynamic aging(hydrolysis and condensation) behaviour of organofunctional silanes in theaqueous solution on their penetrability into the cell walls of wood, BioRe-sources 6 (2011) 2323–2339.

43] S. Saka, T. Ueno, Several SiO2 wood-inorganic composites and their fire-resisting properties, Wood Sci. Technol. 31 (1997) 457–466.

44] M.H. Schneider, J.G. Phillips, High weight percent gain (WPG) furfural-ureamodification of wood, Google Patents (2009).

45] H. Zhao, J.H. Kwak, Z.C. Zhang, H.M. Brown, B.W. Arey, J.E. Holladay, Studyingcellulose fiber structure by SEM, XRD, NMR and acid hydrolysis, Carbohydr.Polym. 68 (2007) 235–241.

46] B. Tjeerdsma, H. Militz, Chemical changes in hydrothermal treated wood: FTIRanalysis of combined hydrothermal and dry heat-treated wood, Holz als roh-und Werkstoff 63 (2005) 102–111.

Please cite this article in press as: K. Zhou, et al., Mechanism and effhistoric buildings, Journal of Cultural Heritage (2020), https://doi.org/

47] K.K. Pandey, A. Pitman, FTIR studies of the changes in wood chemistry follow-ing decay by brown-rot and white-rot fungi, Int. Biodeter. Biodegr. 52 (2003)151–160.

48] S.A.A.K.M. Hamed, M.L. Hassan, A new mixture of hydroxypropyl cellulose andnanocellulose for wood consolidation, J. Cult. Herit. 35 (2019) 140–144.

[

PRESSeritage xxx (2020) xxx–xxx 9

49] S. Wang, J. Shi, C. Liu, C. Xie, C. Wang, Fabrication of a superhydrophobic surfaceon a wood substrate, Appl. Surf. Sci. 257 (2011) 9362–9365.

50] D. Lucca, R. Ghisleni, J.-K. Lee, Y. Wang, M. Nastasi, J. Dong, A. Mehner, Effects ofion irradiation on the structural transformation of sol-gel derived TEOS/MTESthin films, Nucl. Instrum. Methods Phys. Res. Sect. B: Beam Interact. Mater.Atoms 266 (2008) 2457–2460.

51] A. Fidalgo, R. Ciriminna, L.M. Ilharco, M. Pagliaro, Role of the alkyl–alkoxideprecursor on the structure and catalytic properties of hybrid sol–gel catalysts,Chem. Mater. 17 (2005) 6686–6694.

52] N. Rangelova, L. Radev, S. Nenkova, I.M. Salvado, M.V. Fernandes, M. Her-zog, Methylcellulose/SiO2 hybrids: sol-gel preparation and characterization byXRD, FTIR and AFM, Open Chem. 9 (2011) 112–118.

53] E.K. Kim, J. Won, J.-y. Do, S.D. Kim, Y.S. Kang, Effects of silica nanoparticle andGPTMS addition on TEOS-based stone consolidants, J. Cult. Herit. 10 (2009)214–221.

54] Z. Olejniczak, M. Łeczka, K. Cholewa-Kowalska, K. Wojtach, M. Rokita, W. Moz-gawa, 29Si MAS NMR and FTIR study of inorganic–organic hybrid gels, J. Mol.Struct. 744 (2005) 465–471.

55] T.-J. Ha, S.G. Choi, S.-B. Jung, B.-G. Yu, H.-H. Park, The improvement ofmechanical and dielectric properties of ordered mesoporous silica film usingTEOS–MTES mixed silica precursor, Ceram. Int. 34 (2008) 947–951.

56] Y. Ma, H.R. Lee, T. Tsuru, Study on preparation and hydrophobicity of MTESderived silica sol and gel, in: Advanced Materials Research, Trans Tech Publ,2012, pp. 2563–2566.

57] Y. Ko, B. Ratner, A. Hoffman, Characterization of hydrophilic—hydrophobicpolymeric surfaces by contact angle measurements, J. Colloid Interface Sci. 82(1981) 25–37.

58] Y. Li, X. Dong, Y. Liu, J. Li, F. Wang, Improvement of decay resistance of woodvia combination treatment on wood cell wall: Swell-bonding with maleicanhydride and graft copolymerization with glycidyl methacrylate and methylmethacrylate, Int. Biodeter. Biodegr. 65 (2011) 1087–1094.

59] M.A. Tshabalala, P. Kingshott, M.R. VanLandingham, D. Plackett, Surface chem-istry and moisture sorption properties of wood coated with multifunctionalalkoxysilanes by sol–gel process, J. Appl. Polym. Sci. 88 (2003) 2828–2841.

60] S.A. Kulkarni, S.B. Ogale, K.P. Vijayamohanan, Tuning the hydrophobic prop-erties of silica particles by surface silanization using mixed self-assembledmonolayers, J. Colloid Interface Sci. 318 (2008) 372–379.

61] R. Al-Oweini, H. El-Rassy, Synthesis and characterization by FTIR spectroscopyof silica aerogels prepared using several Si(OR)4 and R′ ′ Si(OR′)3 precursors, J.Mol. Struct. 919 (2009) 140–145.

62] N. Gopal, K. Narasimhulu, J. Rao, EPR, optical, infrared and Raman spectral stud-ies of actinolite mineral, Spectrochim. Acta Part A: Mol. Biomol. Spectrosc. 60(2004) 2441–2448.

63] J. Lu, P. Askeland, L.T. Drzal, Surface modification of microfibrillated cellulosefor epoxy composite applications, Polymer 49 (2008) 1285–1296.

64] X. Hua, S. Kaliaguine, B. Kokta, A. Adnot, Surface analysis of explosion pulps byESCA, Wood Sci. Technol. 28 (1993) 1–8.

65] R.N. Wenzel, Resistance of solid surfaces to wetting by water, Ind. Eng. Chem.28 (1936) 988–994.

66] R.N. Wenzel, Surface roughness and contact angle, J. Phys. Chem. 53 (1949)1466–1467.

67] X. Fu, X. He, Fabrication of super-hydrophobic surfaces on aluminum alloysubstrates, Appl. Surf. Sci. 255 (2008) 1776–1781.

68] A. Nakajima, K. Hashimoto, T. Watanabe, Recent studies on super-hydrophobicfilms, in: Molecular Materials and Functional Polymers, Springer, 2001, pp.31–41.

69] J. Andrade, S. Ma, R. King, D. Gregonis, Contact angles at the solid—water inter-face, J. Colloid Interface Sci. 72 (1979) 488–494.

70] X. Wang, Y. Chai, J. Liu, Formation of highly hydrophobic wood surfaces using

ect of alkoxysilanes on the restoration of decayed wood used in10.1016/j.culher.2019.11.012

(2013) 667–672.71] H. Miyafuji, S. Saka, Na2O–SiO2 wood-inorganic composites prepared by the

sol–gel process and their fire-resistant properties, J. Wood Sci. 47 (2001)483–489.