ARTICLE IN PRESS

Solar Energy Materials & Solar Cells 92 (2008) 1149– 1161

Contents lists available at ScienceDirect

Solar Energy Materials & Solar Cells

0927-02

doi:10.1

� Corr

E-m

journal homepage: www.elsevier.com/locate/solmat

Preparation of a TiMEMO nanocomposite by the sol– gel methodand its application in coloured thickness insensitive spectrallyselective (TISS) coatings

Bostjan Japelj a, Angela Surca Vuk a, Boris Orel a,�, Lidija Slemenik Perse a, Ivan Jerman a, Janez Kovac b

a National Institute of Chemistry, Hajdrihova 19, SI-1000 Ljubljana, Sloveniab Jozef Stefan Institute, Jamova 39, SI-1000 Ljubljana, Slovenia

a r t i c l e i n f o

Article history:

Received 21 February 2008

Accepted 5 April 2008Available online 27 May 2008

Keywords:

Sol–gel

Nanocomposite

Selective paints

Solar absorber

48/$ - see front matter & 2008 Elsevier B.V. A

016/j.solmat.2008.04.003

esponding author. Tel.: +386 14760276; fax:

ail address: boris.orel@ki.si (B. Orel).

a b s t r a c t

An organic–inorganic nanocomposite was prepared via sol–gel processing from 3-(trimethoxysilyl)-

propyl methacrylate (MAPTMS) and titanium(IV) isopropoxide (TIP) precursors (TiMEMO) in the form of

a viscous resin, and used as a binder for the preparation of coloured thickness insensitive spectrally

selective (TISS) paints and corresponding solar absorber coatings. The spectral selectivity of TiMEMO-

based TISS paints was optimized by varying the concentrations of binder and different pigments: black,

coloured (red, green and blue) and aluminium flakes, the latter imparting low thermal emittance, which

was correlated to the presence of titanium in the TiMEMO sol–gel host. The formation and the ensuing

structure of the sol–gel TiMEMO hybrid was studied in detail and the nanocomposite structure of the

TiMEMO binder formed was assessed from infrared and 29Si NMR measurements, which confirmed the

formation of Ti–O–Si linkages established after the hydrolysed precursors condensed into a compliant

resinous material. XRD measurements provided additional information about the existence of small

coherent domains of silsesquioxane units in the sol–gel host. The abrasion resistance of the non-

pigmented TiMEMO binder deposited in thin film form on a PMMA substrate was assessed by the Taber

test, and its hardness compared with other resin binders which have been used for making TISS paint

coatings. The surface properties of the non-pigmented TiMEMO binder and the ensuing TISS paint

coatings were determined from contact angle measurements. The results showed that the water contact

angles of non-pigmented TiMEMO binder increased from 701 to 125–1351 for the corresponding

pigmented TISS paint coatings, inferring the influence of surface roughness on surface energy in the

presence of pigments. SEM measurements revealed a striking similarity in the surface morphology of

the TISS paint coatings with some other surfaces exhibiting the Lotus effect.

& 2008 Elsevier B.V. All rights reserved.

1. Introduction

Despite fast growing needs for spectrally selective coatingsstemming from the fact that in Europe nearly 50% of fossil fuels arenowadays spent for heating buildings (www.esttp.org), surpris-ingly, relatively small numbers of new selective coatings have beendeveloped for these specific applications. Since the early 1950s,when Tabor [1] proposed and demonstrated the usefulness ofselective surfaces for increasing the photothermal efficiency ofsolar collectors, many types of absorber have been reported andproduced [2,3] but only a few became available commercially.

The European solar thermal market is today dominated byblack highly selective (as ¼ 0.93–0.94, eT ¼ 0.04–0.07) but ex-pensive thickness sensitive spectrally selective (TSSS) sputtered

ll rights reserved.

+386 14760300.

cermet coatings, such as TiNOx, Alanod Sunselect and BlueTech. Acheaper black TSSS paint coating [4,5] prepared via the coil-coating technique is produced by Chromagen (Israel) but themarket is relatively small. However, none of the mentionedcoatings is suitable for solar fac-ade absorbers. TSSS coatings areneither mechanically nor corrosion resistant enough to assure thelong service life of unglazed solar fac-ade collectors. An additionaldrawback to the widespread use of solar fac-ades lies in the blackappearance of solar absorbers. According to Weiss [6], more than85% of architects would prefer colours other than black, even if alower efficiency would have to be accepted. Accordingly, the mainfocus of this study was to develop coloured selective paintcoatings with various hues suitable for unglazed solar absorbers.

Thickness insensitive spectrally selective (TISS) paint coatingsrepresent the material of choice because they combine theadvantages of paints (longevity, chemical resistance achieved bya high thickness of the applied layer, variety of colours and simpleapplication) with spectral selectivity, which does not depend on

ARTICLE IN PRESS

B. Japelj et al. / Solar Energy Materials & Solar Cells 92 (2008) 1149–11611150

the thermal emittance of the substrate. The latter is achieved by theaddition of metallic flakes (Al, Cu and Ni) to the paint dispersion,significantly decreasing their thermal emittance. The idea of makingblack TISS paint coatings was first reported by McKinley andZimmer [7], Telkes [8] and Moore [9,10] and since then has beensuccessfully realized by Hoeflaak [11]. His black TISS paint coatingsmade of black pigment, an organic resin binder and low-emittingaluminium flakes exhibited as ¼ 0.84–0.90 and eT ¼ 0.41–0.47.

Coloured TISS paint coatings (various shades of blue, green andred) prepared in our laboratory [12–15], were made by embeddingappropriate amounts of aluminium flakes, and coloured and blackpigments in silicone [14] or polyurethane [15] binders. However,in order to prepare even better coloured TISS coatings, new resinbinders with infrared absorption lower than those of the alreadyused polyurethane and silicone binders, higher abrasion resis-tance and possibly antisoiling (i.e. hydrophobic and oleophobic)properties are required. Sol–gel organic–inorganic hybrids [16–18]seem to be suitable because the sol–gel processing can be done atlow temperature and because of the variety of commerciallyavailable sol–gel precursors.

Sol–gel based materials have not been frequently used as bindersfor selective coatings. Rincon et al. [19] reported TSSS coatingsprepared from tetraethoxysilane (TEOS) in combination withtitanium(IV) isopropoxide (TIP) with added carbon soot and carbonnanotubes as pigments. However, high curing temperatures(300–600 1C) were needed to fully cure TSSS selective coatings.Posset et al. [20,21] reported TISS coatings made by mixing Al flakeswith single end-capped 3-isocyanatopropyltriethoxysilane or 3-(triethoxysilyl)propyl succinic anhydride, which is combined with apoly(vinylbutyral) copolymer bearing –OH groups. The ensuing TISScoatings have good mechanical properties due to the added sol–gelprecursor, which enhanced the amount of silica (inorganic) phase,but this led to the increase of eT values above 0.45. It is obvious thatthermal emittance and mechanical hardness are interrelated proper-ties; increasing the amount of the sol–gel inorganic phase in thebinder mixture deteriorates the selectivity. Therefore, a sol–gelhybrid having an inherently low thermal emittance is needed.

In principle, there are two possible ways to avoid an excessivethermal emittance of sol–gel hybrids, which mainly originates fromthe massive infrared absorption of the Si–O–Si bonding, typical oflinear (or branched) silsesquioxanes (R8(SiO3/2)8, R ¼ organicgroup), characteristic condensation products of silanes. Onepossibility to overcome this is to use polyhedral oligomericsilsesquioxanes (POSS) [22], such as octahydrido POSS, having cyclic(cage-like) H8(SiO3/2)8 structure, known as low-k materials (e0o4.0)in microelectronics. Another possibility, which we applied in thisstudy, is to use binders which have an inherently low infraredabsorption due to the presence of heavier atoms (or their clusters)randomly substituted within the Si–O–Si chains. Heavier atoms(titanium, for example) decouple the internal electric field estab-lished due to the strong dipole–dipole interactions of polar groupsexisting in linear or branched Si–O–Si chains. Moreover, such atomswhen bonded to oxygens show vibrational bands below 1000 cm�1,and thus do not contribute as strongly as Si–O–Si modes to thethermal emittance in the spectral range 1200–1000 cm�1. Accord-ingly, the binder for TISS paints was prepared from acryloxypropyl-triethoxysilane (APTMES or MEMO, for short [23] and TIP, with anorganic–inorganic hybrid structure (TiMEMO, for short).

The TiMEMO sol–gel hybrid is not new and has been used as‘‘hard’’ coating to increase the abrasion resistance of acrylic(PMMA) and polycarbonate transparent panes [24–28]. However,practically no information could be found about the use of theTiMEMO sol–gel hybrid as a binder for pigmented coatings such asTISS paints, or about the interactions of TiMEMO with pigments,important in achieving non-agglomerated pigment dispersionsand spectral selectivity.

In this paper we first report the preparation of TiMEMOorganic–inorganic hybrids, accentuating the need for controllingthe extent of hydrolysis and condensation reactions of theMAPTMS and TIP precursors, and their effect on the ensuingstructure of TiMEMO binders. In order to demonstrate thesuitability of TiMEMO binder for solar fac-ade absorber coatings,the thermal infrared absorbance and abrasion resistance ofTiMEMO were next measured and compared to those of othercommonly used organic binders. Finally, some examples ofselective coloured TISS paint coatings are given, demonstratingthe advantages of the TiMEMO binder for achieving high solarabsorptance and low thermal emittance in selective coatings.Interestingly, the presence of pigments dramatically decreases thesurface energy values of the TISS paint coatings, as demonstratedby water contact angle measurements.

2. Experimental

2.1. Instrumental

IR absorption spectra of the sols and films deposited on siliconwafers by dip-coating (10 cm/min) were recorded on a Bruker IFS66/S instrument with a resolution of 4 cm�1 using 32 scans foreach sample. Solar absorptance (as) of coatings in the visible andnear infrared spectral regions was determined from the reflec-tance spectra measurements performed on a Perkin ElmerLambda 950 UV/Vis/NIR with an integration sphere (module150 mm). Thermal emittance values (eT) were obtained from thereflection spectra measured on a Bruker IFS 66/S spectrometer,equipped with an integrating sphere (OPTOSOL) using a gold plateas a standard for diffuse reflectance. A standard procedure wasused for the evaluation of as and eT values [29]. XRD spectra wereobtained on a PANalytical X’pert PRO X-ray diffractometer (CuKa,l ¼ 1.5406 A), while 29Si NMR spectra were recorded on a VarianUnity Plus 300 MHz with a Dotty CPMAS head. XPS analyses werecarried out on a PHI-TFA XPS spectrometer (Physical ElectronicsInc.). The sample surfaces were excited by X-ray radiation from anAl source and the area analysed was 0.4 mm in diameter. The O 1s,Si 2p and Ti 2p spectra were acquired with an energy resolution ofabout 1.0 eV. Sample charging during XPS analysis was compen-sated by a low-energy electron gun-neutralizer.

SEM micrographs were obtained on a FE-SEM Supra 35 VPelectron scanning microscope. Since the paint coatings containedconductive aluminium flakes, no additional preparation of thesamples was needed to make micrographs. TEM micrographswere obtained on a JEOL 2000 FX transmission electron micro-scope, operating at 200 kV.

The thickness of the coatings was measured on a Taylor-Hobson Talysurf 2. The abrasion test was done on a Teledyne Taberabraser, Model 503, by rubbing the films with a pair of CS-10wheels, with a 500 g load per wheel. The abrasion resistance ofthe films was graded using the difference of haze between thewear track and non-abraded regions after 500 cycles of abrasion.The contact angles were measured on a contact angle goniometer(Kreuss). Total surface energy values gtot were determined byintroducing the contact angle values of water, methylene iodideand formamide test liquids into the van Oss relation [30]. Thedetails regarding contact angle measurements were reportedelsewhere [31].

2.2. Preparation of MEMO and TiMEMO binders

A typical procedure used for the preparation of the MEMO andTiMEMO nanocomposite binders is presented in Fig. 1. First,

ARTICLE IN PRESS

Fig. 1. Preparation of TiMEMO binders.

B. Japelj et al. / Solar Energy Materials & Solar Cells 92 (2008) 1149–1161 1151

MAPTMS (Aldrich) was diluted in tetrahydrofuran (THF), thenacidified water (0.1 M HCl) was added to perform hydrolysis. After5 min mixing, TIP (Merck) diluted in THF was added dropwise for30 min during stirring. The TiMEMO binder was obtained after theremoval of the THF under reduced pressure, resulting in a liquidwith an appropriate viscosity for making pigment dispersions. Nochelating agent such as ethylacetoacetate or acetic acid wasneeded to slow down the fast condensation of TIP [26,28]. Itseemed that the use of THF instead of use of chelating ageing wascrucial for the preparation of TiMEMO binder with long-termstability and appropriate viscosity for paint preparation.

Three different TiMEMO binders were prepared with molarratios of the reactants MAPTMS:TIP:THF ¼ 23:9:100 (A), 20:18:100 (B) and 17:27:100 (C). The content of TIP was 28, 47 and61 wt% (WTIP ¼ mTIP/(mTIP+mMAPTMS)), respectively, for the bindersA, B and C, corresponding to 10, 20 and 30 wt% of TiO2 in theTiMEMO binder ðWTiO2

¼ mTiO2=ðmTiO2

þmMAPTMSÞÞ. The preparednanocomposites were denoted as 10TiMEMO, 20TiMEMO,30TiMEMO. It should be noted that the corresponding concentra-tions of titanium in the binders were much higher than reportedin the literature [24–28]. In addition, the reference MEMO hybridwas prepared from MAPTMS (MAPTMS:THF ¼ 25:100) withoutthe addition of TIP, following the preparation scheme shownin Fig. 1.

2.3. Preparation of TISS paints

Coloured TISS paints were prepared using a standard proce-dure [14,15]. Black (CuCr2O4 spinel black (Ferro, D)) and red (iron

oxide pigment (BASF, D)) pigment dispersions were prepared first,by mixing the corresponding pigments with the 20TiMEMObinder in certain proportions and grinding it in a ball mill. Theconcentration (i.e. partial pigment-to-volume concentration ratio,PVC) of pigments and low-emitting Al flakes in dispersions waskept the same as for the TISS paints reported previously [14,15].TISS paints were sprayed on a metallic substrate (decapped andheat rolled sheet metal) with a thickness of up to 50 mm and thenthermally cured in an air-circulating oven for 30 min at 150 1C. Thecoatings made of commercially available binders were curedaccording to the supplier’s instructions. Among them we shouldmention Nanocryls C 140 product (Nanoresins), a colloidaldispersion of up to 50 wt% amorphous silica in hexandioldiacry-late as a binding agent. The dispersed phase consists of surface-modified, spherically shaped SiO2 nanoparticles with a diameterof 20 nm with a very narrow particle size distribution. In this waythe specific advantages of organic and inorganic materials werecombined almost perfectly.

3. Results and discussion

3.1. Structural properties of MEMO and TiMEMO

3.1.1. IR spectroscopy

The evolution of MEMO and TiMEMO gels from the corre-sponding sols was followed by measuring their infrared transmis-sion and 29Si NMR spectra. Infrared spectra of sols deposited onsilicon wafers showed that the hydrolysis of MAPTMS either inMEMO (Fig. 2A) or TiMEMO (Fig. 2B) sols was fast, which was

ARTICLE IN PRESS

2000

0

1

2

3

4

5

6

1170

1635

1700

1719

1106

1042

816

982

Abs

orba

nce

5 min 30 min 1 hour 2 hours 4 hours 1 day 7 days

909

Si-O-Si

MeOH

Si-OH

C = C

C = O

0

1

2

3

4

Abs

orba

nce

5 min 30 min 1 hour 2 hours 1 day 7 days

1170

1635

1699

1719

1096

1041

816

927

910

Si-O-Si

MeOH

Si-OH

C = C

C = O

Ti-O-Si

Wavenumber (cm-1)1500 1000 500

2000Wavenumber (cm-1)

1500 1000 500

Fig. 2. IR transmission spectra of: (A) MEMO and (B) 10TiMEMO binders deposited

on silicon wafers during ageing of the sol.

2000

0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

1.6

1.8

2.0

Si-OH

Si-O-Si

1038

1719

1635

927

976

1095

1167

1064

Abs

orba

nce

MEMO 10TiMEMO 20TiMEMO 30TiMEMO

1106

Si-O

-Ti

C = C

C = O

2000-1.0

-0.8

-0.6

-0.4

-0.2

0.0

0.2

0.4

0.6

0.8

1.0Si-O-Ti

Abs

orba

nce

10TiMEMO 20TiMEMO 30TiMEMO

89392

7967

Si-O-Si

Wavenumber (cm-1)1500 1000 500

Wawenumber (cm-1)1500 1000 500

Fig. 3. IR transmission spectra of MEMO and various TiMEMO binders deposited

on silicon wafers: (A) sols aged 7 days and (B) their subtracted spectra (TiMEMO-

MEMO).

B. Japelj et al. / Solar Energy Materials & Solar Cells 92 (2008) 1149–11611152

evident from the loss of the bands attributed to the methoxy(Si–OMe) groups at 2844 and 722 cm�1 (Fig. 2). Already5 min after the addition of acidified water, a strong silanolband (Si–OH) at 909 cm�1 appeared. During the course of ageing(7 days) the spectra changed gradually, showing a decreaseof the Si–OH bands and a gradual intensity increase of the bandsin the 1000–1100 cm�1 region attributed to siloxane (Si–O–Si)linkages and eventual formation of POSS units. It should bestressed that condensation became slower after 1 day, asshown by the small intensity change of the Si–O–Si band in thefollowing 7 days. This indicated that the sols were relativelystable, which is a precondition for their application as binders forpaints. The IR spectra also revealed that the CQO band of theacrylic group moved from 1719 to 1700 cm�1 during thecondensation process due to the establishment of hydrogenbonding between the CQO and unreacted Si–OH groups[32,33]. After the consumption of silanol groups for the formationof siloxane network, the CQO vibrational mode returned to1719 cm�1. Expectedly, the sol–gel reactions did not affect theCQC groups suggesting the absence of polymerization of acrylicgroup (see below).

Inspection of the IR spectra of various TiMEMO (Fig. 3A)revealed an increase in intensity in the spectral range of 850–990 cm�1 with regard to the IR spectrum of MEMO. Expectedly,the subtraction spectra (TiMEMO–MEMO) in Fig. 3B confirmedthat the intensity in this spectral region increased with increasingamount of TIP in the corresponding TiMEMO samples. Thisintensity increase can undoubtedly be attributed to the formationof Si–O–Ti linkages, as found in organic–inorganic polymethyl-methacrylate PMMA/SiO2/TiO2 composite thin films at 930 cm�1

[26,28].It should be stressed that thermal and UV curing also

influenced the Si–O–Si vibrational modes. Namely, they wereslightly shifted from 1095 to 1110 cm�1 (Fig. 4), indicating theestablishment of branched siloxane structures or, eventually,formation of cyclic or open cube-like silsesquioxane species.Polymerization of acrylic groups during thermal or UV curing ledto a complete disappearance of the CQC vibrational mode at1634 cm�1. As a consequence, the CQO vibrational mode shiftedfrom 1719 to 1734 cm�1, which is typical of the acrylic group [34]due to the loss of conjugated OQC–CQC bonds [32,33]. Thisindicated the complete polymerization of acrylic groups.

ARTICLE IN PRESS

2000-0.1

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

C = C

Si-O-Ti

1637

1719

1734

1110

1167

1095

927

Abs

orba

nce

0 minutes 15 minutes 30 minutes

C = O

Si-OH

2000-0.1

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

Abs

orba

nce

0 minutes 30 minutes 60 minutes

C = C

Si-O-Ti

1637

1719

1732 1110

1167

1095

927

C = O

Si-OH

1500 1000 500Wavenumber (cm-1)

Wavenumber (cm-1)1500 1000 500

Fig. 4. IR transmission spectra of 10TiMEMO binder deposited on Si wafer: (A)

uncured (0 min) and thermally cured for 15 and 30 min at 150 1C, and (B) uncured

(0 min) and UV cured for 30 and 60 min (Irgacure 184, UVA 100 W/m2).

4000

0.0

0.5

1.0

1.5

2.0

2.5

3.0

3.5

4.0

Si-O-Si

1096

430

Ti-O

-Ti

1051

Abs

orba

nce

10TiMEMO initial 10TiMEMO thermally treated

927

Si-O

-Ti

3500 3000 2500 2000 1500 1000 500Wavenumber (cm-1)

Fig. 5. IR transmission spectra of initial and thermally treated (400 1C, 30 min)

10TiMEMO binders deposited on silicon wafers obtained from sols aged 7 days.

B. Japelj et al. / Solar Energy Materials & Solar Cells 92 (2008) 1149–1161 1153

The establishment of Si–O–Ti linkages in 10TiMEMO sampleswas also noted when curing of a polymerizable acrylic group wasperformed either by thermal treatment (150 1C, 30 min) (Fig. 4A)or UV curing (Irgacure 184 (Ciba), UVA 100 W/m2, 60 min) (Fig.4B). For either type of curing a band at 927 cm�1 characteristic ofthe Si–O–Ti vibrational mode [26,28] developed in the IR spectra(Fig. 4). Its intensity significantly increased when thermaltreatment was performed at 400 1C (Fig. 5), indicating that strongSi–O–Ti linkages developed at higher annealing temperatures.

3.1.2. 29Si NMR spectroscopy

The important finding of IR spectroscopy, i.e. the formation ofTi–O–Si linkages, could also be independently confirmed by 29SiNMR measurements. Trifunctional alkoxysilanes like MAPTMSshow chemical shifts in the 29Si NMR range from �35 to �75 ppm[35] and the signals appear in four group of lines described byTn(i,j), where n represents the number of siloxane bonds on acertain silicon atom, while i and j represent the number of –OHand –OR groups, respectively. 29Si NMR spectra obtained frombulk MEMO gels (Fig. 6A) revealed a strong signal at �57.3 ppm

attributed to the T2 condensation unit, indicating the presence of acyclic tetramer T2(1,0) (T4(OH)) [36]. The asymmetricity on the‘‘right-hand tail’’ of this signal in addition showed that linearsiloxane species are also present. Namely, due to a reduction inSi–O–Si angles, the cyclic oligomers are located at downfieldpositions compared to the corresponding linear species [35,36]. Itwas also found that sol–gel condensation of trialkoxysilanes isslowed down with increase in organic R group, which also causessteric hindrance and favours cyclization [37].

However, after addition of titanium atoms local disturbances ofthe Si–O environments are expected [38], and we noted that TIP inTiMEMO samples changed the 29Si NMR spectra. The T2 signalsplit into �57.3 ppm (T2(1,0)) and �58.2 ppm (T2(0,0)), and a T3

signal in the range from �66.1 to �67.2 ppm characteristic ofeither higher polyhedral structures (T7(OH)3, T8 cube-like species,etc.), branched (beads-on-string) silsesquioxane network orladder type condensation products appeared [35,36,39]. Theappearance of the T2 signal at �58.2 ppm in the TiMEMO spectraindicated the presence of a cyclic tetramer T2(0,0), which reflectedthe possible formation of Ti–O–Si linkages [38]. The existence ofTi–O–Si bonds, already found in the IR spectra, was also inferredfrom the 29Si NMR spectra of thermally cured TiMEMO samples(Fig. 6B). Curing at 150 1C (30 min) led to formation of distinct T3

signals in all samples, but their intensity decreased with theincrease in TiO2 content on account of the intensity of the T2 signalsuggesting the presence of Ti–O–Si linkages in TiMEMO samples.

3.1.3. X-ray diffraction

The XRD spectra of MEMO and TiMEMO samples thermallycured at 150 1C for 30 min are depicted in Fig. 7. The diffractionpatterns of transparent samples crushed to powders showed abroad peak at 2Y ¼ 20–211 and, in addition, a distinct reflection atlow Bragg angles (4–101) appeared, with an intensity cruciallydepending on the concentration of TiO2 in TiMEMO. Using Bragg’srelation, the positions of peaks were attributed to the cor-relation distance (d), which existed in real space between thestructural units.

The broad diffraction peak at �211, the intensity of whichdiminished with the increasing concentration of TiO2 in TiMEMO,was associated with the presence of amorphous silica domains[40], indicating overall disorder in the siliceous skeleton. Mostgels prepared from alkoxysilanes exhibit a similar broad peak

ARTICLE IN PRESS

0

-58.

2

T3

-67.

2

-57.

3

x 4

x 4Inte

nsity

δ (ppm)

MEMO 10TiMEMO 20TiMEMO 30TiMEMO

x 4

T2

0

-66.

1 30

TiM

EMO

-58.

2

-66.

1 M

EMO

-66.

7 20

TiM

EMO

T3

-67.

2 10

TiM

EMO

MEMO 10TiMEMO 20TiMEMO 30TiMEMO

-57.

3

T2

-10 -20 -30 -40 -50 -60 -70 -80

δ (ppm)-10 -20 -30 -40 -50 -60 -70 -80

Fig. 6. 29Si NMR spectra of MEMO and various TiMEMO binders of: (A) uncured

and (B) thermally cured nanocomposites (solid state).

0

30TiMEMO20TiMEMO

10TiMEMO

21.5

°

5.4°

6.9°

46.2

°

20.5

°

Inte

nsity

4.4°

4.5°

MEMO

200

2Θ (°)10 20 30 40 50 60 70 80

Fig. 7. XRD spectra of thermally cured (30 min at 150 1C) MEMO and various

TiMEMO powders.

B. Japelj et al. / Solar Energy Materials & Solar Cells 92 (2008) 1149–11611154

between 201 and 251 that could be ascribed to correlations among–CSiO3 units [41–43]. The calculated correlation distances d forour samples were 4.4 A for pure MEMO and around 4.1 A forTiMEMO, coinciding well with structural unit distance of 4.2 A invitreous silica [44].

The origin of the peaks that appeared at low Bragg angles, from6.91 (MEMO) to 4.41 (30TiMEMO), exhibiting correlation distancesfrom 12.8 to 20.1 A, is more problematic. In the XRD spectra (Fig. 7)the peaks that appear at small Bragg angles are important forelucidation of the mesoporous or nanocomposite structure of thematerials. The XRD patterns of mesoporous TiO2 [45,46], SiO2 [47]and mixed titania-silica [48,49] materials usually reveal a singlelow-angle diffraction peak around 2Y �21 characteristic of theaverage pore-pore correlation distance in local pore ordering. Inmesoporous materials, the lack of high angle diffractions indicatesthe absence of long-range pore ordering and therefore disordering inthe atomic arrangement. On the other hand, the low-anglereflections up to 2Y �101 could also stem from silica clusters thatcan form through condensation of terminal ethoxy groups to variouslinear, cyclic or cube-like silsesquioxane polyhedra [41,42]. Forinstance, a correlation distance of�1 A was observed for transparentMTEOS gels [41,42], where it was attributed to the distance betweenthe adjacent (R-SiO3/2)n clusters of T8 silsesquioxane cubes. Someother crystalline alkyltrialkoxysilanes (methyl, isobutyl) after con-densation consist of T8 and T10 cubes, revealing sharp diffractionpeaks in this region of the Bragg angles. For example, the crystallineoctaisobutylsilsesquioxanes with well-defined cube-like structureexhibit this peak at 2y �7.81. However, it was reported that theposition of these peaks shift to lower angles, i.e. longer distances, in(R-SiO3/2)8 with larger organic groups [50,51].

The low-angle diffraction peak of pure MEMO was noted at6.91, but was of low intensity (Fig. 7). The position of thisreflection agreed with the establishment of small coherentdomains (L ¼ 83.8 A) of polyhedral silsesquioxane units in thepredominant overall disorder of the siliceous skeleton. This is inaccordance with the results of 29Si NMR spectroscopy of thermallytreated MEMO samples (Fig. 6B), revealing a predominant T2

signal at �57.3 ppm, while the T3 signal at �66.1 ppm, character-ising silicons triply bonded to oxygens, was much less expressed.The presence of titanium in TiMEMO samples shifted thecorresponding low-angle XRD diffraction peak to lower angles,reaching an angle of 4.41 for the highest concentration of TiO2

added. This value is approaching that characteristic of mesopor-ous TiO2 (i.e. �21) [45–49] but remained too high to reflect theeventual mesoporosity of our samples. Simultaneously with theshift of this diffraction peak to lower angles (Fig. 7), its intensityincreased significantly, suggesting the increase in the correlationdistance due to the presence of higher amount of titanium atomsin the MEMO host.

3.1.4. XPS analysis

In order to identify the structure of the TiMEMO nanocompo-site, all samples were investigated by XPS, a method capable ofrevealing interaction between TiO2 and SiO2 mixed oxides.Namely, the binding energies of O 1s, Ti 2p and Si 2p reflect theeventual formation of Ti–O–Si linkages [52,53] and some distinctpeak shifts were also detected for our TiMEMO samples. Of all theTiMEMO samples in Fig. 8 the changes in binding energies werethe smallest for Ti 2p and Si 2p photoelectrons. The formershowed a peak at 459.3 eV for all compositions and the latter at102.5 eV, but a minor shift of Si 2p binding energy in the case of30TiMEMO toward 102.0 eV was noticed. It is generally admittedthat the shift in binding energy is correlated with the degree ofinteraction between TiO2 and SiO2 in the titanium silicate phase.The fact that the Ti 2p binding energy does not appear either at

ARTICLE IN PRESS

4700

1000

2000

3000

4000

5000c/

s

Binding energy (eV)

Ti 2p

1150

400

800

1200

1600

2000

c/s

Binding energy (eV)

Si2p

5450

5000

10000

15000

c/s

Binding energy (eV)

MEMO10TiMEMO20TiMEMO30TiMEMO

MEMO10TiMEMO20TiMEMO30TiMEMO

MEMO10TiMEMO20TiMEMO30TiMEMO

O 1s

465 460 455 450 110 105 100 95 90

540 535 530 525 520

Fig. 8. (A) Ti 2p fraction, (B) Si 2p fraction and (C) O 1 s fraction of XPS spectra of MEMO and various TiMEMO powders.

B. Japelj et al. / Solar Energy Materials & Solar Cells 92 (2008) 1149–1161 1155

460.4 eV (anatas) [54] or at 458.6 eV (rutile) [55] but at 459.3 eV(Fig. 8C), and the fact that the Si 2p binding energy does notappear at 104.8 eV (pure silica) [52] but at 102.5 eV, supports thefact that neither TiO2 nor SiO2 existed in TiMEMO samples ascrystalline phases, but only short-range ordering of structuralunits existed in an overall amorphous silicious and titaneousskeleton [52]. The persistence of Ti 2p and Si 2p binding energiesat 459.3 and 102.5 eV evidently showed that there is no or only aminor change in the delocalization of electrons around Ti and Siatoms (Fig. 8A and B). This can be explained by the fact that themajority of the titanium and silicon units exist independently,while a part of the siloxanes inserts the Ti–O–Si linkages.

The binding energy of O 1s photoelectrons of MEMO was foundat 532.5 eV (Fig. 8C) and shifted toward lower values with theincreasing of Ti content in TiMEMO. In the case of 30TiMEMO theO 1s binding energy reached 531.5 eV, while for pure TiO2 it was529.9 eV [52]. Orignac et al. [56] reported that the presence ofminor Ti–O–Si linkages in the silica network contributed to theappearance of an asymmetrical O 1s peak in the XPS spectra. Inour case the asymmetrical shape of O 1s peak of the TiMEMO wasevident, but their deconvolution was not performed. Namely, theO 1s of Si–O–Si linkages neighbouring the Ti–O–Si bonds areaffected by delocalization of electrons due to the difference inelectronegativity between Ti and Si atoms [52].

We can conclude that the silicious and titaneous skeletons areorganized as a random net-like distribution of atoms withincorporated silsesquioxane structural units (�80 A), and con-nected with some degree of mixed Ti–O–Si linkages. Thecompletely amorphous structure of MEMO and TiMEMO nano-

composites was also confirmed by the TEM micrographs and SAEDimages (Fig. 9).

3.2. Properties of binders and TISS paint coatings

3.2.1. Abrasion resistance

In order to compare the surface mechanical properties ofcommercial and TiMEMO binders, the abrasion resistance of theresins sprayed on PMMA substrate was measured. The Taberabrasion test revealed that the abrasion resistance of thecommercial most commonly used binders was lower than theabrasion resistance of 20TiMEMO. An exception was Nanocryls

(Nanoresins AG), which is used as an additive for abrasionresistant lacquers (Figs. 10 and 11). It can be noticed that PUA,chlorinated rubber and acrylic binders have the lowest abrasionresistance; their haze is about 0.30. The vinyl and PU binders’abrasion resistance is higher, with haze values of about 0.16, whilethe abrasion resistance of 20TiMEMO is excellent and in the samerange as that of Nanocryls with a haze of about 0.02–0.04.

3.2.2. Average absorption coefficients of binders and thermal

emittance of TiMEMO TISS paint coatings

Prior to the preparation of TISS paint coatings on the basis ofTiMEMO binder, the influence of TiMEMO on thermal emittance(eT) was investigated. Namely, a lowering of the average absorp-tion coefficient of TiMEMO binder was expected due to thepresence of heavy Ti atoms in the silica host, which—whenbonded to oxygen—show IR vibrational bands below 600 cm�1, i.e.

ARTICLE IN PRESS

Fig. 9. (A) TEM micrograph and (B) SAED image of a 20TiMEMO powder.

400

0.0

0.1

0.2

0.3

0.4

Nanocryl*20TiMEMO

PUPMMA substrateVinyl

Acrylic

Chlorinated rubber

0.020.04

0.150.160.17

0.28

0.32

Abs

orba

nce

Wavelength (nm)

0.33PUA

0.00

0.05

0.10

0.15

0.20

0.25

0.30

0.35

Haz

e

Nan

ocry

l*

20Ti

ME

MO

PUViny

l

Acr

ylic

Chl

orin

ated

rubb

er

PU

A

500 600 700 800 900 1000

Fig. 10. Taber abrasion results: (A) absorbance vs. wavelength, and (B) haze of

20TiMEMO in comparison with various commercially available binders. *Nano-

cryls (Nanoresins AG).

B. Japelj et al. / Solar Energy Materials & Solar Cells 92 (2008) 1149–11611156

in the spectral region where they do not contribute as strongly tothe thermal emittance as the siloxane bands in the spectral region1000–1200 cm�1. The average absorption coefficient (ka

aver) ofvarious commercially available binders and TiMEMO was esti-mated from their IR absorbance spectra. The coatings ofcommercial binders and various TiMEMO were prepared withoutaddition of black pigment dispersions or Al flakes. Binders wereapplied on IR transmitting silicon wafers and their thickness wasdetermined. The ka

aver values of all the binders considered weredetermined from the slope of the average absorbance (Aaver) vs.

thickness (d) of the binder coating by applying the equation [57]:

kavera ¼

Aaver

d¼ 4pken (1)

where n is the wavenumber and ke is the extinction coefficient.The Aaver values were obtained by averaging the intensity of thevibrational bands observed in the spectral region from 4000 to400 cm�1. Averaging over the black body emission curve, as usedfor the determination of eT values, represents an alternativemethod.

The results (Fig. 12) revealed that the commercially availableNanocryls binder, which is used for making hard lacquers,exhibited the highest IR absorption (kaver ¼ 0.057) due to thepresence of silica particles. Expectedly, MEMO followed it closelywith a ka

aver of 0.026, because MEMO also consists of silica unitswith siloxane (Si–O–Si) bonding, which exhibit rather intensive IRabsorption bands in the 1100–1000 cm�1 region. The lowestIR absorbance was obtained for binders based on vinyl andchlorinated rubber, lacking a stiff and strongly polar polyetherpolymeric backbone. Pure polyurethane binder (ka

aver¼ 0.018)

exhibited higher absorption than the polyurethane binderfunctionalized with hydroxyacrylate groups (PUA) (ka

aver¼ 0.013).

A similar inherently low IR absorption (kaaver¼ 0.012) was also

found for the acrylic binder which therefore represents aprospective binder for TISS paints. Expectedly, the presence of

ARTICLE IN PRESS

Fig. 11. Photographs of the results of Taber abrasion tests of: (a) PUA, (b) chlorinated rubber, (c) acrylic, (d) vinyl, (e) PU, (f) 20TiMEMO and (g) Nanocryls (Nanoresins AG)

binders.

B. Japelj et al. / Solar Energy Materials & Solar Cells 92 (2008) 1149–1161 1157

titanium in TiMEMO coatings decreased the kaaver values of pure

MEMO, reaching the values of pure PU and PUA in the case of30TiMEMO. The latter results confirmed the assumption that thepresence of heavier titanium atoms in the hybrid sol–gel hostdecoupled the polar–polar interactions of Si–O–Si modes in theMEMO sol–gel host by establishing Ti–O–Si linkages (Fig. 10), asalready confirmed by IR, 29Si NMR and XPS measurements.

In order to confirm the advantageous influence of theinherently low infrared absorption of the TiMEMO binder, TISSpaints made of either TiMEMO or PUA binder containing the sameamount of Al flake pigment were made and the correspondingeT values were determined from their IR reflectivity spectra

(Fig. 13A). Expectedly, the as values of TiMEMO and PUA coatingswere low (as ¼ 0.23) and the same for both, but the correspondingeT values were 0.17 for the former and 0.32 for the latter TISS paintcoating; TiMEMO binder practically did not absorb in the IRspectral region. In the next step, the amount of TiMEMO and PUAbinder was gradually increased and eT values for each of thecorresponding TISS paint coatings were determined (Fig. 13B). Theresults exhibited the considerably higher eT values of the PUA-Alflake-containing coatings compared to the same coatings made ofTiMEMO binder. Finally, the influence of the black pigment on eT

values of Al flake-containing TIMEMO-based TISS paint coatingswas assessed by making paints having various amounts of black

ARTICLE IN PRESS

0.00

0.01

0.02

0.03

0.04

0.05

0.06

Chl

orin

ated

rubb

er

Viny

l

Acr

ylic

Pol

yure

than

e/A

cryl

ic

30Ti

ME

MO

20Ti

ME

MO

Pol

yure

than

e

10Ti

ME

MO

ME

MO

k ave

r

Nan

ocry

l*

Fig. 12. Comparison of average absorption coefficients kaaver of commercial,

MEMO and various TiMEMO binders. kaaver was obtained from IR absorption

spectra averaged in the spectral region from 4000 to 400 cm�1; * Nanocryls

(Nanoresins AG).

20.00

0.25

0.50

0.75

1.00

20TiMEMO based TISS paint coating (eT=0.17; aS=0.23) PUA based TISS paint coating (eT=0.32; aS=0.23)

Ref

lect

ance

Wavelength (μm)

difference in reflection

200.00

0.25

0.50

0.75

1.00

e T

Binder concentration (%)

Al flakes-20TiMEMO coatings Al flakes-PUA coatings black TISS coatings based on 20TiMEMO

4 6 8 10 12 14 16

40 60 80 100

Fig. 13. Properties of coatings prepared from either 20TiMEMO or PUA binder and

Al-flakes: (A) reflectance spectra, and (B) the influence of binder concentration on

eT values. The behaviour of black TISS paint coatings based on 20TiMEMO is also

shown.

B. Japelj et al. / Solar Energy Materials & Solar Cells 92 (2008) 1149–11611158

pigment (Fig. 13B, triangles). Surprisingly, eT values increasedmuch faster than noted for the TiMEMO-Al flakes coatings,enabling the conclusion that the absorption of the pigment andnot the binder represent the main obstacle to making low thermalemitting TISS paint coatings. A high infrared transmitting blackpigment has to be sought for in future in order to achieve lowthermal emittance TISS paint coatings.

3.3. Colours vs. spectral selectivity of TISS paint coatings

Fig. 14 shows the variation of spectral selectivity (eT vs. as),lightness (L*), colour coordinates (a*, b*), metric chroma (C*) andchromatic performance criteria C*PC0.34 [14,15] values of TiME-MO-based red, green and blue TISS paint coatings having variousamounts of added black pigment. C*PC0.34 was constructed fromthe performance criteria (CP0.34) [14,15,58,59] corresponding tothe following relation:

PCðunglazedÞ ¼ as � 0:95þ 0:34ð0:95� eTÞ (2)

For unglazed solar collectors eT values are weighted by thefactor 0.34, which increases to 0.50 for glazed systems. Eventhough a complete analysis of the as and eT interrelations forcoloured selective coatings is still lacking, the PC values enabledthe estimation of the gains of the coloured TISS paint coatingscompared to those of the black spectrally non-selective coatings.Typically, for coatings with moderate selectivity (as�0.85 andeT�0.35), the PC value is 0.104, i.e. higher than the PC of blacknon-selective paint coatings (PC ¼ 0), but lower than the PC ofhighly selective (Sunselect, for example) coatings (as ¼ 0.94,eT ¼ 0.06, PC ¼ 0.293). The positive PC values demonstrate thespectral selectivity of the applied paint coatings.

The results in Fig. 14 show that inevitably, selectivity andcolour deteriorated with the concentration of black pigment,corroborating the need to find a new black pigment which wouldhave a higher IR transmission than the pigment used in this study.It should be noted that C*PC0.34 values were in the range from 0.4to 2 and were about 20% higher than the previously reported ones[14,15]. The most promising TISS paint coatings were the blueones due to their high C* values for chosen L*, as and eT values.

3.4. Contact angle phenomena

It is desirable that paint coatings for solar fac-ade applicationpossess not only abrasion resistance but a self-cleaning (i.e.antisoiling) effect as well. The antisoiling properties are usuallyattributed to the so-called Lotus effects also expressed by thehigh contact angle for water (y4140–1501) and small (ao5–101)sliding angle of water drops, properties which are responsible forkeeping the coating surfaces clean (‘‘self-cleaning’’ properties).The ‘‘Lotus effect’’ is a combination of surface roughness and lowsurface energy obtained due to the presence of –CH2 and –CH3

groups incorporated in the paints by the binder itself or suitableadditives. This effect is achieved by the addition of suitablepowders (AEROXIDEs types, Evonik), waxes (Michemsguardtypes, Mitchelman) and nanosilanes (Sivo types, Evonik). Theseadditives remain in the coatings permanently, temporarily or theadditives could be considered sacrificial (waxes) since they couldbe easily removed by washing. A typical example of highlyhydrophobic (water contact angle y ¼ 136–1381) and permanentfinish (20 repetitive washings) for cotton fabrics has been ob-tained by impregnating fibres with sol–gel based bis end-cappedalkoxysilane linked to polydimethylsiloxane chain (PDMSU, forshort) [31], which we also used in this study as a hydrophobicsol–gel additive for a coloured TISS paint coating sample. After theaddition of hydrolysed (0.1 M HCl) PDMSU to the TISS paint, the

ARTICLE IN PRESS

0.00.00.10.20.30.40.50.60.70.80.91.0

TISS paint coatings based on 20TiMEMO or PUA

1black PUA

a s

eT

1red 20TiMEMO

green20TiMEMOblue 20TiMEMO

11

1

black 20TiMEMO

dark red 20TiMEMO

black

00

10

20

30

40

50

60

70

TISS paint coatings based on 20TiMEMO

bluegreen

dark red

red 11

1

1

L*

C*

1black

black

-40-50

-40

-30

-20

-10

0

10

20

TISS paint coatings based on 20TiMEMO

green

blue

reddark red

1

1

1

1

b*

1

black

black black

-3

-2

-1

0

1

2

3

C*P

C0,

34

TISS paint coatings based on 20TiMEMO

1redgreen

blue

1

11

black 1

dark redblack

0.1 0.2 0.3 0.4 0.5 0.6 10 20 30 40 50 60

a*-30 -20 -10 0 10 20 30 40

1

Fig. 14. (A) aS vs. eT values, (B) lightness (L*) vs. chroma (C*) values, (C) colour coordinates b* vs. a* and (D) iPC0.34 values of black, red, blue and green TISS paint coatings

prepared with either 20TiMEMO or PUA binder. The increasing aS values were attained by the addition of the same amounts of black pigment dispersions to TISS paints

having identical composition (horizontal arrows in figures indicate the direction of increasing concentration of black pigment).

Table 1Contact angles for various solvents and surface energy values calculated according

to van Oss equation [30]

Binder Water MeJ2 FA s+ s� sLW sAB stot

Acryl 78 45 61 0.01 8.97 37.01 0.52 37.53

Vinyl 75 65 62 0.82 12.12 25.70 6.30 32.01

PU 84 32 66 0.44 6.42 43.37 3.35 46.72

PUA 73 36 59 0.08 13.29 41.56 2.12 43.68

20TiMEMO 64 43 45 0.75 14.72 37.54 6.64 44.19

Nanocryl 64 57 58 0.18 23.58 30.30 4.10 34.40

TiMEMO-PDMSU-MEMO 107 96 92 1.01 1.31 10.18 2.30 12.49

MEMO 73 51 50 1.28 7.92 33.71 6.38 40.09

TISS paint Water MeJ2 FA s+ s� sLW sAB stot

Black 127 66 89 0.03 3.79 25.13 0.70 25.83

Blue 119 57 92 0.81 0.23 30.3 0.86 31.16

Green 126 69 99 0.49 0.82 23.43 1.27 24.70

Red1 121 67 94 0.24 0.36 24.56 0.58 25.15

MeI2—methylene iodide (CH2I2); FA—formamide; stot—total surface energy;

sLW—apolar interactions; sAB—polar interactions; s+—electron–acceptor interac-

tions; s�—electron–donor interactions.

B. Japelj et al. / Solar Energy Materials & Solar Cells 92 (2008) 1149–1161 1159

corresponding coating exhibited a water contact angle up to138–1401, exceeding the water contact angles of pure non-pigmented TiMEMO binder by nearly 801 (Table 1), even thoughthe addition of PDMSU to the TiMEMO binder (no pigment added)increased the water contact angle just to 1071.

This drastic increase (from 1071 (TiMEMO binder) to 138–1401(coloured TiMEMO TISS paint)) of the water contact angles wasattributed to the roughness effect provided by the pigmentparticles, which were displaced to the coating surface from theinterior of the paint. Inspection of the water contact angles forblack, green, red and blue TISS paint coatings (Table 1) showedthat even without the added hydrophobic PDMSU sol–gel hybridthe corresponding water contact angles of TiMEMO TISS paintcoatings became 119–1271, i.e. up to 601 higher compared to thewater contact angles of the coatings made of non-pigmentedTiMEMO binder (641). Consequently, the surface energy values ofthe pigmented TISS TiMEMO coloured coatings decreased drasti-cally, particularly the corresponding polar portion (sAB) (Table 1),leading to low-energy surface energy coatings having stot valuesin the range from 24 to 31 mJ/m2, i.e. typical of stronglyhydrophobic surfaces (Fig. 15).

The effect of surface morphology on the low surface energyvalues was verified from SEM micrographs (Fig. 16), which clearlyshowed the highly agitated surface of the pigment particles. Onthe contrary, SEM micrographs of the analogous TISS paintcoatings prepared from PUA binder showed a relatively smoothtype surface with the pigment particles well embedded in the

paint. The origin of this effect is not clear, but it was possible tospeculate that it stemmed from the specific interaction betweenthe surface of the pigment particles with the sol–gel TiMEMObinder, having many unreacted silanol groups, which are able tointeract with the hydroxylated surface of the pigment particles.

ARTICLE IN PRESS

0102030405060708090

100110120130140

blac

k TI

SS

- Ti

ME

MO

Al f

lake

s - P

UA

Con

tact

ang

le (°

)

blac

k TI

SS

- P

UA

PU

A

Al f

lake

s - T

iME

MO

dark

red

TIS

S -

TiM

EM

O

TiM

EM

O

red

TIS

S -

TiM

EM

O

blue

TIS

S -

TiM

EM

O

gree

n TI

SS

- Ti

ME

MO

20TiMEMO 20TiMEMO0

10

20

30

40

50

blue TISS paint

γtot

γLW

γ-

γAB

Sur

face

Ene

rgy

(mJ/

m2 )

γ+

γtot

γLW

γ-

γAB

γ+

black TISS paint

Fig. 15. (A) The contact angles of PUA and 20TiMEMO binder TISS paint coatings in

comparison with PUA and 20TiMEMO binders, and (B) their surface energy values.

Fig. 16. SEM micrographs of surface of the: (A) black PUA, and (B) black

20TiMEMO TISS paint coatings.

B. Japelj et al. / Solar Energy Materials & Solar Cells 92 (2008) 1149–11611160

Independent evidence that the interactions between the pigment(black, green, red and blue) particles and not with Al flakes play adecisive role in forming the rough and therefore low-energysurface was obtained from the water contact angles measured forTISS paint coating containing only Al flakes. The correspondingwater contact angles reached only �891, and were nearly 30–401smaller than for TISS paint coaings with incorporated coloured orblack pigments. It seems that large Al flakes (60–80 mm) could notcontribute to the required surface roughness typical of the Lotuseffect [60–64]. This problem requires special study which shouldbe made in future.

4. Conclusions

The results obtained in this study confirmed that the TiMEMObinder with a nanocomposite structure enabled the preparation ofTISS paint coatings having as values up to 90% and eT close to 30%.The TiMEMO binder consisted of silica units (T2, T3) and titaniaatoms covalently interlinked to Si units via Si–O–Ti bonding, asshown by IR, 29Si NMR and XPS spectroscopy. The main advantageof TiMEMO lies in the fact that the presence of the titania phaseconsiderably decreases the IR absorption of MEMO and othersilica-based nanocomposites, and improves the mechanicalproperties compared to purely organic polymeric binders. Morework is needed to establish the long-term stability of these novelhybrid materials for application in TISS paint coatings for solar fac-ade collector systems.

Acknowledgement

B. Japelj thanks the Ministry of Higher Education, Science andTechnology for a Ph.D. grant.

References

[1] H. Tabor, Transactions of the Conference on the Use of Solar Energy, vol. 11,Part I, Section A1–23, University of Arizona Press, Tucson, 1956.

[2] C.G. Granqvist (Ed.), Materials Science for Solar Energy Conversion Systems,Pergamon Press, Oxford, 1991.

[3] C.E. Kennedy, Review of mid-to-high temperature solar selective absorbermaterials, Technical Report, NREL/TP-520-31267, /http://www.osti.gov/bridgeS.

[4] B. Orel, Z. Crnjak Orel, N. Leskovsek, M. Klanjsek Gunde, Surf. Coat. Int. 75(1992) 488.

[5] B. Orel, Z. Crnjak, I. Radoczy, Sol. Energy Mater. 18 (1988) 97.[6] W. Weiss, I. Stadler, in: Proceedings of the Industry Workshop of the IEA Solar

Heating and Cooling Programme, Task 26, IEA, Delft, 2002, pp. 59–63.[7] W.D. McKinley, P.B. Zimmer, in: Proceedings of the 11th National SAMPE

Technical Conference, Boston, 1979, pp. 435–442.[8] M. Telkes, US Patent 4.011.190 (1997).[9] S.W. Moore, Sol. Energy Mater. 12 (1985) 435.

[10] S.W. Moore, Proc. SPIE 502 (1984) 68.[11] M. Hoeflaak, J. Gerrit, Patent DE3105936 (1981).[12] B. Orel, A. Surca Vuk, H. Spreizer, D. Merlini, M. Vodlan, Slovene Patent

S121510 (2005).[13] B. Orel, H. Spreizer, A. Surca Vuk, M. Fir, D. Merlini, M. vodlan, M. Kohl,

Optimisation of selective solar absorber paints: performance, durability andcolours, in: Gleisdorf solar 2006—Internationales Symposium fur Sonnene-nergienutzung, AEE INTEC, Gleisdorf, 2006, pp. 123–129.

[14] B. Orel, H. Spreizer, L. Slemenik Perse, M. Fir, A. Surca Vuk, D. Merlini,M. Vodlan, M. Kohl, Sol. Energy Mater. Sol. Cells 91 (2007) 93.

[15] B. Orel, H. Spreizer, M. Fir, A. Surca Vuk, D. Merlini, M. Vodlan, M. Kohl, Sol.Energy Mater. Sol. Cells 91 (2007) 108.

ARTICLE IN PRESS

B. Japelj et al. / Solar Energy Materials & Solar Cells 92 (2008) 1149–1161 1161

[16] U. Schubert, N. Husing, A. Lorenz, Chem. Mater. 7 (1995) 2010.[17] P. Judeinstein, C. Sanchez, J. Mater. Chem. 6 (1996) 511.[18] D.A. Loy, K. Rahimian, Building hybrid organic–inorganic materials using

silsesquioxanes, in: H.S. Nalwa (Ed.), Handbook of Organic–Inorganic HybridMaterials and Nanocomposites, vol. 1, American Scientific Publishers,Valencia, Canada, 2003, pp. 126–163.

[19] M.E. Rincon, J.D. Molina, M. Sanchez, C. Arancibia, E. Garcia, SEMSC 91 (2007)1421.

[20] U. Posset, K. Rose, Innovative hybrid coatings for fac-ades, in: N. Auner, J. Weis(Eds.), Organosilicon Chemistry V, From Molecules to Materials, Wiley-VCH,Weinheim, Germany, 2003, pp. 465–649.

[21] K. Rose, U. Posset, K.-H. Haas, M. Kohl, Farbe Lack 108 (2002) 29.[22] M.A. Grant, B. Craig, J. Electrochem. Soc. 145 (1998) 4019.[23] K.-H. Haas, Crosslinked heteropolysiloxanes as inorganic–organic polymers:

precursors, synthesis, properties and applications, in: H.S. Nalwa (Ed.),Handbook of Organic–Inorganic Hybrid Materials and Nanocomposites, vol. 1,American Scientific Publishers, Valencia, Canada, 2003, pp. 207–240.

[24] T.C. Sandvig, US Patent 4073967 (1978).[25] T.C. Sandvig, US Patent 4042749 (1977).[26] W. Que, Z. Sun, Y. Zhou, Y.L. Lam, S.D. Cheng, Y.C. Chan, C.H. Kam, Mater. Lett.

42 (2000) 326.[27] S. Sepeur, N. Kunze, B. Werner, H. Schmidt, Thin Solid Films 351 (1999) 216.[28] G. Gu, Z. Zhang, H. Dang, Appl. Surf. Sci. 221 (2004) 129.[29] M.G. Hutchins, Spectrally selective materials for efficient visible, solar and

thermal radiation control, in: M. Santamouris (Ed.), Solar Thermal Technol-ogies for Buildings, James & James, London, England, 2003, pp. 37–64.

[30] C.J. van Oss, R.J. Good, M.K. Chaudhury, Langmuir 4 (1988) 884.[31] J. Vince, B. Orel, A. Vilcnik, M. Fir, A. Surca Vuk, V. Jovanovski, B. Simoncic,

Langmuir 22 (2006) 6489.[32] J. Martinez Urreaga, M.C. Matias, V. Lorenzo, M.U. de la Orden, Mater. Lett. 45

(2000) 293.[33] S.K. Medda, D. Kundu, G. De, J. Non-Cryst. Solids 318 (2003) 149.[34] Y. Yunhua, D. Yi, Colloid. Polym. Sci. 281 (2003) 794.[35] F. Brunet, J. Non-Cryst. Solids 231 (1998) 58.[36] M. Feuillade, C. Croutxe-Barghorn, C. Carre, J. Non-Cryst. Solids 352 (2006)

334.[37] L. Matejka, O. Dukh, D. Hlavata, B. Meissner, J. Brus, Macromolecules 34

(2001) 6904.[38] V.I. Zarko, Yu.I. Gorlov, V.V. Brei, Theor. Ex. Chem. 22 (1986) 224.

[39] M. Voronkov, V.I. Lavrentyev, Top. Curr. Chem. 102 (1982) 199.[40] L.D. Carlos, V. de Zea Bermudez, R.A. Sa Ferreira, L. Marques, M. Assunc- ao,

Chem. Mater. 11 (1999) 581.[41] Z. Zhang, H. Wakabayashi, T. Akai, J. Sol–Gel Sci. Technol. 12 (1998) 153.[42] B. Orel, R. Jese, A. Vilcnik, U. Lavrencic Stangar, J. Sol–Gel Sci. Technol. 34

(2005) 251.[43] B. Orel, R. Jese, U. Lavrencic Stangar, J. Grdadolnik, M. Puchberger, J. Non-

Cryst. Solids 351 (2005) 530.[44] S.C. Moss, in: D. Adler, H. Fritzche, S. Ovshinsky (Eds.), Physics of Disordered

Materials, Plenum Press, New York, USA, 1985.[45] Q. Dai, L.Y. Shi, Y.G. Luo, J.L. Blin, D.J. Li, C.W. Yuan, B.L. Su, J. Photochem.

Photobiol. A 148 (2002) 295.[46] Q. Sheng, S. Yuan, J. Zhang, F. Chen, Micropor. Mesopor. Mater. 87 (2006) 177.[47] T.R. Pauly, R.J. Pinnavaia, Chem. Mater. 13 (2001) 987.[48] A. Bhaumik, S. Samanta, N.K. Mal, Micropor. Mesopor. Mater. 68 (2004) 29.[49] W. Wang, M. Song, Mater. Res. Bull. 41 (2006) 436.[50] A.J. Barry, W.H. Daudt, J.J. Domicone, J.W. Gilkey, J. Am. Chem. Soc. 77 (1955)

4248.[51] K. Larsson, Arkiv Kemi 16 (1960) 209.[52] S.M. Jung, O. Dupont, P. Grange, Appl. Catal. A: Gen. 208 (2001) 393.[53] Z. Li, B. Hou, Y. Xu, D. Wu, Y. Sun, W. Hu, F. Deng, J. Solid State Chem. 178

(2005) 1395.[54] Z. Li, B. Hou, Y. Xu, D. Wu Y. Sun, J. Coll. Inter. Sci. 288 (2005) 149.[55] Y.-L. Lin, T.-J. Wang, Y. Jin, Powder Technol. 123 (2002) 194.[56] X. Orignac, H.C. Vasconcelos, R.M. Almeida, J. Non-Cryst. Solids 217 (1997)

155.[57] M. Buback, Special techniques and applications, in: B. Schrader (Ed.), Infrared

and Raman Spectroscopy, VCH, Weinheim, Germany, 1995.[58] COLOURFACE, Coloured collectors facades for solar heating systems and

building insulations, EU project, Contract No. ENK6-CT-2001-30009 (2001).[59] SOLABS, Development of unglazed solar absorbers (resorting to coloured

selective coatings on steel material) for building facades, and integration intoheating systems, EU project, Contract No. ENK6-CT-2002-00679 (2002).

[60] E. Burkarter, C.K. Saul, F. Thomazi, N.C. Cruz, L.S. Roman, W.H. Schreiner, Surf.Coat. Technol. 202 (2007) 194.

[61] O.-U. Nimittrakoolchai, S. Supothina, J. Eur. Cer. Soc. 28 (2008) 947.[62] L. Guo, W. Yuan, J. Li, Z. Zhang, Z. Xie, Appl. Surf. Sci. 254 (2008) 2158.[63] A. Nakajima, K. Hashimoto, T. Watanabe, Monatsh. Chem. 132 (2001) 31.[64] D. Quere, Rep. Prog. Phys. 68 (2005) 2495.