d e n t a l m a t e r i a l s 2 7 ( 2 0 1 1 ) 972–982

avai lab le at www.sc iencedi rec t .com

journa l homepage: www. int l .e lsev ierhea l th .com/ journa ls /dema

Improving performance of dental resins by adding titaniumdioxide nanoparticles

Jirun Suna,∗, Aaron M. Forsterb, Peter M. Johnsonc, Naomi Eidelmana, George Quinna,Gary Schumachera, Xinran Zhangc, Wen-li Wuc

a American Dental Association Foundation, Paffenbarger Research Center, Gaithersburg, MD 20899-8546, USAb Materials and Construction Research Division, National Institute of Standards and Technology, Gaithersburg, MD 20899-8615, USAc Polymers Division, National Institute of Standards and Technology, Gaithersburg, MD 20899-8541, USA

a r t i c l e i n f o

Article history:

Received 7 January 2011

Received in revised form

24 May 2011

Accepted 21 June 2011

Keywords:

Dental resin

Dental adhesives

Shear bond strength

Titanium dioxide

Nanoparticles

a b s t r a c t

Objective. The objective of this study is to improve the performance of dental resins by

adding a small amount of titanium dioxide nanoparticles (TiO2 NPs), which have outstanding

mechanical properties and unique photoactivities.

Methods. Acrylic acid modified TiO2 NPs (AP25) were prepared and added to a mixture

of bis-phenol-A-dimethacrylate and triethylene glycol dimethacrylate (mass ratio 1:1) at

seven mass fractions. Disks made of these resins were subjected to FTIR microspectroscopy,

nanoindentation, microindentation, and 3-point bending to determine the degree of vinyl

conversion (DC) modulus and hardness. The shear bond strengths (SBS) of dentin adhesives

containing various amount of AP25 were also examined.

Results. The DC increased as a function of mass fraction of AP25 and reached a plateau

at 0.1%. The DC of the resin mixture was improved by ≈7% up to 91.7 ± 0.8%. The elastic

modulus and hardness of the composites increased initially as more AP25 were added, and

decreased after reached the maximum value at approximately 0.06% mass fraction of AP25.

The maximum elastic modulus was ≈48% higher than that of the NP-free resin, and the

maximum hardness was more than twice higher than that of the NP-free resin. Using these

resin composites as dental adhesives, the mean SBS using resins with 0.1% mass fraction

of AP25 was ≈30% higher than those using NP-free resin.

Significance. By adding a small amount of AP25 to the resin, the DC and the mechanical prop-

erties of resins were improved dramatically. These findings could lead to better performing

dental adhesives.

emy

(2) chemically (photocatalytic activities including killing bac-

© 2011 Acad

1. Introduction

Titanium dioxide (TiO2) nanoparticles (NPs) have been demon-

strated to be an effective multifunctional material [1,2]. Asthe particle size decreases, especially <50 nm, they exhibitphoto-induced activities that originate from the semicon-

∗ Corresponding author at: 100 Bureau Drive, MS 8546, Gaithersburg, MDE-mail address: jsun@nist.gov (J. Sun).

0109-5641/$ – see front matter © 2011 Academy of Dental Materials. Pudoi:10.1016/j.dental.2011.06.003

of Dental Materials. Published by Elsevier Ltd. All rights reserved.

ductor band gap of TiO2 NPs [2–5]. With their energy higherthan the band gap, photons can generate electron–hole pairswhose energy may be applied (1) electrically (solar cells) [4];

20899-8546, USA. Tel.: +1 301 975 5439; fax: +1 301 975 9143.

teria and viruses) [3]; or (3) by changing the hydrophilicity ofthe particle surface and resulting in superhydrophilicity [6].This energy is expressed chemically via free radicals (HO•)

blished by Elsevier Ltd. All rights reserved.

2 7

ftbhtim[mt

fttaTlhpn[

rTotas[tw

iacbdTaaattNl(riohcs

2

2

Rp

microscopy (TEM) and dynamic light scattering (DLS) wereused to determine particle size and morphology, while Fouriertransform infrared (FTIR) spectroscopy and thermal gravi-

1 Certain equipment, instruments or materials are identifiedin this paper in order to adequately specify the experimental

d e n t a l m a t e r i a l s

ollowing a redox reaction mechanism [7]. The photoactivi-ies mentioned above can be efficiently activated over timey easily accessible resources (such as light) [2]. TiO2 NPs alsoave excellent mechanical properties, for example, the elas-ic modulus of TiO2 NPs is approximately 230 GPa, and it isnexpensive with titanium being the fourth most abundant

etal on earth, following aluminum, iron and magnesium2]. The unique photoactivities of TiO2 NPs and their superior

echanical properties make them one of the ideal additiveso enhance the performance of polymeric materials [8].

In dentistry, strong and durable dental adhesives are vitalor the success of long-term resin composites used to restoreooth cavities [9,10]. Dental adhesives bond resin composite toooth structure and must withstand a harsh oral environmentnd endure occlusal loads from biting, grinding and chewing.he major components in dental adhesives are dimethacry-

ate derivatives (dental resins) with/without fillers. Besidesigh bond strength to tooth and composites, good mechanicalroperties of the dental resins including modulus and hard-ess are also essential for success of the dental restoration

11,12].TiO2 NPs have been used as an additive in dental mate-

ials to match the opaque properties of teeth [13], but usingiO2 NPs as additives to enhance the mechanical propertiesf dental resins, has not been successful due to the inconsis-ent agglomeration of TiO2 NPs [14]. Many methods have beenttempted to prevent the aggregation by modifying the TiO2

urface with different reagents including surfactants/ligands15,16], proteins [17] and acids [18]. However, adding TiO2 NPso substantially improve mechanical properties of the resinsas still a concept rather than a fact.

In this study, we developed a convenient protocol tomprove the performance of dental resins by adding a smallmount (less than 0.1% mass fraction) of TiO2 NPs withontrolled agglomeration. A mixture (mass fraction 1:1) ofis-phenol-A-dimethacrylate (BisGMA) and triethylene glycolimethacrylate (TEGDMA) was used as base resin, and theiO2 NPs were treated with acrylic acid. The agglomerationnd dispersion of particles was controlled by centrifugationnd solvent evaporation. The successful attachment of acryliccid to the surface of TiO2 NPs was confirmed using Fourierransform infrared (FTIR), X-ray powder diffraction (XRD) andhermal gravimetric analysis (TGA). The dispersion of TiO2

Ps agglomerates in ethanol was evaluated using dynamicight scattering (DLS) and transmission electron microscopyTEM). Significant improvement of the performance of theesins were achieved: first, the addition of the TiO2 NPsncreased degree of vinyl conversion (DC) of the resin; sec-nd, the mechanical properties including elastic modulus andardness were improved dramatically; third, using the TiO2-ontaining resin as bonding agents of dentin adhesives, thehear bond strength was enhanced.

. Materials and methods

.1. Materials and sample preparation

esins, including 2,2-bis(4-(2-hydroxy-3-methacryloxypro-oxy)phenyl)propane (BisGMA), triethyleneglycol dimethacry-

( 2 0 1 1 ) 972–982 973

late (TEGDMA),1 and pyromellitic glycerol dimethacrylate(PMGDM), were obtained from Esstech Inc. Mg(N-tolylglycineglycidyl methacrylate)2 was also from Esstech Inc.The initiators, camphorquinone (CQ), and ethyl 4-N,N-dimethylaminobenzoate (4E) and benzoyl peroxide, werepurchased from Aldrich Corp. Acrylic acid was obtained fromSigma Corp. Titanium dioxide nanoparticles (P25, AEROXIDETiO2), a known photo-catalytically active material composedwith both anatase and rutile phases, were provided by Evonik.All reagents were used as received.

The P25 TiO2 NPs were modified with acrylic acid, andthe product was labeled as AP25. A mixture of acrylic acid(7.2 mL) and water (0.8 mL) was first stirred in a 25 mL vial.Then 8 mL ethyl acetate solution of P25 (0.2 g P25) was addeddropwise in 30 min, and then the mixture was agitated at 37 ◦Cfor 24 h. A milky mixture was formed and transferred into a50 mL centrifuge tube. The AP25 precipitated in the bottomand was collected after centrifuging at 3000 rpm for 6 min.This precipitate was then redistributed in 25 mL of ethanoland centrifuged at 3000 rpm for 3 min to remove the remain-ing acrylic acid. The same step was also used to prepare AP25organosols in ethanol. The AP25 organosols (≈0.12% by massor ≈0.02% by volume) in ethanol did not form precipitate forseveral days. The mass fraction of AP25 in these organosolswas determined as the solid percentage after the ethanol wasremoved via vacuum (1–2 mm Hg) at 22 ◦C for 24 h.

BisGMA and TEGDMA were mixed at mass ratios 1:1. Theresin mixture was activated for blue light photopolymeriza-tion by adding 0.2% CQ and 0.8% 4E (by mass) and stored inthe dark until use. The AP25 ethanol organosols were mixedinto the activated resins by 10 min sonication. The ethanolwas then removed via vacuum (1–2 mm Hg) at room tempera-ture for 24 h. Resin monomers with seven mass fractions (0%,0.02%, 0.03%, 0.05%, 0.06%, 0.08% and 0.5%) of AP25 were pre-pared.

Preparation of resin disks. A mixture of monomers and AP25was pipetted into a TEFLON cylinder mold (8 mm in diameterand 1 mm thick) that was placed on top of a piece of Mylar filmand a glass slide. Another piece of Mylar film covered the topof the filled mold. The monomer was cured using a DentsplyTriad 2000 visible light curing unit with a tungsten halogenlight bulb (250 W and 120 V) for 1 min each from both opensides of the mold. Five disks of each AP25 mass fraction wereprepared.

2.2. Characterization of AP25 TiO2 NPs

The AP25 powder was characterized via X-ray diffraction(XRD) for crystalline forms of TiO2. Transmission electron

details. Such identification does not imply recommendation by theNational Institute of Standards and Technology or the ADAF, nordoes it imply the materials are necessarily the best available forthe purpose.

l s 2

974 d e n t a l m a t e r i a

metric analysis (TGA) were used to evaluate the attachmentof acrylic acid on the surface of TiO2 NPs. XRD patterns ofP25 and AP25 powders were recorded with Cu K� radiation(� = 0.154 nm) using a Rigaku 2200 D-Max X-ray diffractometer(Rigaku/USA Inc., Danvers, MA, USA) operating at 40 kV and40 mA at 10–80◦ 2� range with intervals of 0.010◦ 2�. The samedivergence and anti-scatter slits (1◦) and receiving slit (0.6 mm)were used for all samples. The TEM images were obtainedusing a Philips EM400T operating at 200 kV. Samples for TEMmeasurements were prepared by dropping a 0.5 �L AP25 solu-tion onto a carbon-coated copper grid followed by evaporationat room temperature over night. The DLS was performed on aninstrument using classical geometric optics i.e., the detectedvolume was defined by a series of lens–aperture–pinhole com-binations. A Coherent Innova 90 laser with a wavelength of532 nm was used. The DLS measurements were carried outat 11 different angles from 30◦ to 130◦ with 10◦ intervalsat 25 ◦C. The hydrodynamic radius (Rh) of particles was cal-culated using the Brookhaven Instruments-DLS software bythe method of cumulants using one- or two-exponential fits[19,20]. The Rh distribution was evaluated by both cumulantsand CONTIN [21,22]. The FTIR measurements of AP25 NPs orP25 NPs were carried out in the Nexus 670 FTIR spectropho-tometer (Thermo Scientific, Madison, WI). AP25 NPs or P25 NPswere mixed with KBr powder (1.5 mg AP25 in 150 mg KBr) andpressed into pellets. A total of 64 scans were collected from650 cm−1 to 4000 cm−1 at 4 cm−1 resolution. FTIR of acrylicacid was measured using the same methods except that theacrylic acid was sandwiched between two KBr pellets. TGAwas carried out using a TGA Q500 (TA Instruments). Samples(≈5 mg) were placed on a platinum weighing pan and heatedfrom room temperature to 800 ◦C at a rate of 20 ◦C/min underflowing N2.

2.3. Degree of vinyl conversion (DC)

The degree of vinyl conversion for the resins in the disks afterphotopolymerization was determined using FTIR reflectancemicrospectroscopy (FTIR-RM) [23]. The Nicolet ContinuumFT-IR microscope (Thermo Scientific, Madison, WI) oper-ated in reflectance mode and interfaced with a Nicolet6700 FT-IR spectrophotometer is equipped with two liquidnitrogen-cooled mercury cadmium telluride detectors (MCT-A: 11,700–650 cm−1; MCT-B: 11,700–400 cm−1), a video camera,and a computer-controlled x–y translation stage. Spectra werecollected with 64 scans from 650 cm−1 to 4000 cm−1 at 8 cm−1

spectral resolution with a beam spot size of 90 �m × 90 �m.Ten spectra each of three disks (8 mm in diameter and 1 mmin thickness) of every combination of resin and AP25 wereobtained from the flat top and bottom of the disks. Each spotwas manually focused before data collection. The reflectancespectra were proportioned against a background of a goldcoated slide and transformed to absorbance spectra using theKramers–Kronig transform algorithm for dispersion correc-tion, which converts the reflectance spectra to absorbance-likespectra. The degree of vinyl conversion (DC) was calculated

as the reduction in the methacrylate peak (1634 cm−1) heightusing the phenyl absorbance peak (1610 cm−1) as an inter-nal standard [24,25]. The peak heights were determined usingthe ISys software (Spectral Dimensions, Olney, MD, USA) [26].

7 ( 2 0 1 1 ) 972–982

The DC was the average of 30 spectra of three disks ofeach mass fraction of AP25. The DC of all of the resin disksthat were subjected to nanoindentation, Knoop hardness andmicroindentation were determined. The standard uncertaintyassociated with the FTIR-RM measurements is <1%.

2.4. Nanoindentation

Nanoindentation measurements were performed using anAgilent NanoXP instrument equipped with a 10 �m radius, 90◦

diamond cone indenter. Samples were indented to a max-imum depth of 500 nm or 1500 nm using a single loadingand the continuous stiffness method. The contact stiffnessbetween the sample and tip was measured by superposinga small oscillation (45 Hz, 5 nm) over the load profile. Theloading time was approximately 120 s with a 30 s hold at themaximum load before unloading. This stiffness was used tocalculate the elastic modulus of the sample [27] assuming aconstant Poisson’s ratio of 0.45, a representative value for den-tal composites [28]. The elastic modulus and hardness weredetermined as the average value obtained over a depth rang-ing from 250 nm to 450 nm at a maximum depth of 500 nmand 950 nm to 1450 nm at a maximum depth of 1500 nm foreach indent and the average of 15 measurements are reported.All indentation experiments were conducted using a constantindentation strain rate of 0.05 s−1. The standard uncertaintyassociated with the nanoindentation measurement is 5%.

2.5. Knoop hardness

The Knoop hardness in this study was measured in accor-dance with ASTM standard E 384. A Leitz Miniload 2microhardness machine was used with indentation loads of0.25–5 N. Indentation sizes were measured with the samemachine using a 10× or 50× objective depending on the lengthof the indentation on the samples. The loading time for anindentation was of the order of 15 s with a dwell at peak loadof 15 s. Knoop hardness was the result of test force divided bythe indentation projected surface area: HK = 14.229P/d2, whereP is the indenter force and d is the long diagonal length [29].The Knoop hardness test was performed on resin disks thathave been evaluated by FTIR-RM first. The hardness was anaverage of five measurements of one resin disk of each massfraction of AP25. The standard uncertainty associated with themicroindentation measurement is 5%.

2.6. Microindentation to determine elastic modulus

A microindentation technique using static load indenterswas also used to measure the elastic modulus of the AP25nanocomposites [30]. The gravity load from stainless steelspheres (radius = 6.35 mm) indent the substrate, forming acontact area dictated by the substrate modulus and inden-ter geometry. Contact areas were measured using an invertedoptical microscope images (Leica DMIRE II), and image analy-sis was performed to find contact radii. From Hertzian contact

mechanics, the Young’s modulus, E, of the substrate could becalculated from indenter geometry and the indentation load,using the previously set Poisson’s ratio of 0.45. Indentationswere performed at five different positions, and images were

2 7

imttswa

2

FSm(fiwu2smMco2dc

2

Toccwp(pUtpgmpaamaeuiDrdatwweic

d e n t a l m a t e r i a l s

mmediately taken after the indenter was placed on the poly-er substrate. For the 0.5% by mass nanoparticle sample, a

hinner sample was required to visualize the indentation dueo particle light scattering obscuring the contact area. For thisample, a modified version of the Hertzian indentation modelas used to correct for the non-infinite substrate thickness

nd calculate the elastic modulus [31].

.7. Flexural modulus

lexural modulus was determined according to ISO4049: 2009.ix rectangular specimens of each material for each test wereade by inserting the composite material into a brass mold

25 mm × 2 mm × 2 mm), covering the surfaces with a Mylarlm to prevent air-inhibited layers. In this test, all of the barsere cured using a Dentsply Triad 2000 visible light curingnit with a tungsten halogen light bulb (250 W and 120 V) formin each from both open sides of the mold. After curing, thepecimens were stored at room temperature for 24 h. Flexuralodulus of the resins was determined using Universal Testingachine (Instron 5500R, Instron Corp., Canton, MA, USA) at a

ross-head speed of 1 mm/min. The specimens were placedn a 3-point bending test device, which was constructed with0 mm distance between supports and ensuring an equallyistributed load. The flexural modulus of each resin was cal-ulated according to ISO4049: 2009.

.8. Shear bond strength (SBS) test

he SBS test followed a previously established protocol inur center [32,33]. Briefly, teeth were embedded with Fastrayomposite (Harry J. Bosworth Company, Skokie, IL, USA) inylindrical holders and ground perpendicular to their long axisith 320-grit SiC paper until the occlusal enamel was com-letely removed. A three-step adhesive procedure was used:

1) the dentin surface was etched with a 37% (by mass) phos-horic acid gel (Etch-Rite; Pulpdent Corp., Watertown, MA,SA) for 15 s and rinsed with distilled water. After rinsing,

he dentin surface was kept hydrated with a moist blottingaper; (2) a mixture of 20 �L 5% (by mass) Mg(N-tolylglycinelycidyl methacrylate)2 acetone solution and 40 �L of 20% (byass) PMGDM acetone solution that contained 2% benzoyl

eroxide (by mass based on resin) was applied as a primernd brushed on the dentin surface accumulating 5 layers,ir drying between layers to evaporate the solvent; and (3)ixtures of resin and AP25 were applied as bonding agents

nd were brushed once on the coated dentin surface. Thentire dentin surface was then light cured for 10 s with these of an 8 mm tip on a quartz halogen light source hav-

ng 450 mW/cm2 intensity (Max 100, Caulk/Dentsply, Milford,E, USA). A poly(tetrafluoroethylene)-covered stainless steel

ing with an opening, 4 mm in diameter and 1.5 mm in depth,efined the bonding area through which the composite waspplied on to the coated dentin. The ring was held down withhe assistance of a polycarbonate holder and the iris was filledith TPH composite (Caulk/Dentsply, Milford, DE, USA), which

as then irradiated for 1 min with the same light source. The

ntire assembly was placed in distilled water 5 min after lightrradiation and stored for 24 h at approximately 22 ◦C beforeonducting a bond test in the shear mode. A commercially

( 2 0 1 1 ) 972–982 975

available bonding solution (Scotchbond, 3M ESPE, St. Paul, MN,USA) was used as the control. Six mass fractions (0%, 0.02%,0.08%, 0.10% 0.12%, and 0.50%) of AP25 were examined, andfive measurements of each mass fraction were carried out.

A holding device was used to evaluate the SBS. The brassring holding the dentin-bonded composite was placed againsta vertical surface of a nylon block. The ring and the compositewere sheared off, at a crosshead speed of 0.5 mm/min, witha flat chisel pressing against the edge of the steel iris. Theflat chisel was connected to the platen of a Universal Test-ing Machine (Instron 5500R, Instron Corp., Canton, MA, USA).The maximum debond load was converted into the SBS of thespecimen. The mean values of SBS were the average of fivemeasurements for each adhesive.

2.9. Statistical analysis

The DC, modulus and hardness were analyzed using one-wayanalysis of variance (ANOVA) with a 95% confidence intervalto indicate significant differences.

3. Results and discussion

3.1. Characterization of P25 and AP25 TiO2 NPs

FTIR spectroscopy and TGA examination confirmed theattachment of the acrylic acid onto the surface of the TiO2

NPs. The –C C– signal at 1636 cm−1 in the FTIR spectrum ofAP25 (Fig. 1A) verified the existence of the double bond afterthe surface treatment of TiO2 NPs. This peak also exists inthe subtraction spectrum of the spectra of AP25 and P25 (seeinset of Fig. 1A), which confirmed that the double bonds wereadded to P25. For comparison, the FTIR measurements of AP25and P25 were run under the same conditions. The vibration offree carbonyl group at 1703 cm−1 disappeared, which indicatedno residual acrylic acid. The asymmetric (�as) and symmet-ric (�s) stretching vibrations of the COO− groups were foundat 1539 and 1437 cm−1. The wave-number difference between�as(COO−) and �s(COO−) that attribute to the chelating biden-tates between acrylic acid and TiO2 surface was 102 cm−1.The characteristic band of the Ti–O–Ti network was found at833 cm−1.

In addition to the FTIR results, TGA results showed2.08 ± 0.17% of weight loss of AP25 from 140 ◦C to 500 ◦C while0.95 ± 0.21% of weight loss of P25 was found in this tempera-ture (Fig. 1B), thus, the attachment of acrylic acid to the surfaceof TiO2 NPs was approximately 1% by mass. The temperaturerange that weight vs. temperature curves of AP25 and P25 arenot parallel to each other was applied to calculate the massfraction of acrylic acid attachment on the P25.

The above results indicated that the double bonds fromthe acrylic acid were successfully attached to the surface ofTiO2 NPs. These double bonds have the potential to bind theTiO2 NPs into the resin network upon photopolymerization. In

return, the AP25 NPs could be creating macro-crosslinked hardpoints in the network due to the excellent mechanical proper-ties of TiO2 (E ≈ 230 GPa). Even if two acrylates on a single AP25particle react, additional crosslinks will be achieved. Above all,

976 d e n t a l m a t e r i a l s 2

0

20

40

60

80

100

0

20

40

60

80

100

800 100012001400 16001800

Wavenumbers (cm-1)

%T

ran

sm

itta

nce

16

36

14

35

Acrylic Acid

AP25

A

476.06°C

97.88%

140.33°C

100.1%

473.75°C

99.00%

96.8

97.8

98.8

99.8

We

igh

t [%

]

0 200 400 600 800

Temperature [°C]

P25AP25

B

30 40 50 60 70 80

2-Theta(°)

x10^3

2.0

4.0

6.0

8.0

10.0

Inte

nsity(C

ounts

)

P25AP25

Anatase (101)

Rutile (110)

C

Fig. 1 – (A) FTIR spectra of acrylic acid and AP25 and thesubtraction spectrum of the spectra of AP25 and P25 (inset);(B) TGA curves of P25 and AP25; and (C) XRD patterns of P25and AP25.

of AP25 NPs were added into the resin the DC was improved

the attachment of double bonds on TiO2 NPs is essential toachieve strong TiO2-containing resins.

The AP25 TiO2 NPs used in this study were mixtures ofanatase (1 0 1) and rutile (1 1 0) phases and dominated byanatase phase. The fraction of the rutile phase (Fr) in the mix-tures was evaluated according to XRD peak intensity using thefollowing formula: Fr = (1 + 0.8(Ia/Ir))−1, where Ia and Ir repre-sent the integrated intensities of anatase and rutile diffractionpeak [34]. The 2-theta of the peaks used for the intensity calcu-lation of anatase and rutile are at 25.3◦ and 27.4◦, respectively(Fig. 1C). The Fr value of AP25 was 22.1%. The Fr value of P25in this study was 19.4%.

Both DLS and TEM indicated that the majority of the AP25particles in the ethanol organosols are agglomerates instead

of individual nanoparticles. The size of the agglomerates wasoptimized by using centrifugation to eliminate big agglom-erates. TEM images (Fig. 2) shows (1) space among these

7 ( 2 0 1 1 ) 972–982

aggregates is within 1 �m, which is a piece of important infor-mation to valid the using of nanoindentation for modulusand hardness evaluation; and (2) the size of a single NP is inthe range of 15–30 nm. The hydrodynamic radius (Rh) of theseagglomerates in an ethanol organosol (Rh = 267 nm ± 32 nm)was determined by DLS in 11 angles from 30◦ to 130◦ at every10◦.

The Rh distribution of the AP25 agglomerates in ethanolsolution was evaluated by both cumulant and CONTIN meth-ods. The normalized variance (also called the polydispersityparameter), �2/� 2 was less than 0.2, which indicated that Rh

was not broadly distributed. CONTIN was also used to esti-mate the size distribution. The fit was divided into 30 gridpoints, and the “dust” term was selected (to handle very large,slow aggregates). The chosen CONTIN fits for all runs revealonly one major peak which is not less than 90% mass fractionof the amplitude and in a narrow size range. All informationsupported the notion that the Rh of the AP25 agglomeratesin ethanol organosols after centrifugation was not distributedover a broad range. The transparency of the organosols wasalso increased after centrifugation, which was a good indica-tion of evenly distributed small aggregates.

The narrow size distribution of AP25 aggregates and theimproved transparency of ethanol organsols were firstly dueto the good compatibility of AP25 to ethanol, and secondlyto the centrifugation process that eliminated the large aggre-gates. Large aggregates were precipitated to the bottom of thecentrifuge tube, and the suspended AP25 NPs did not formprecipitation for several days. These stable organosols werevery important in order to distribute the AP25 NPs evenly intothe resin monomers and vital to achieve a strong resin withAP25 NPs. Poor distribution of AP25 NPs would diminish theadvantage of adding AP25 NPs. Experimental results showedthat AP25 NPs were not as effective when dispersed in resinsby sonication as those distributed using ethanol organosols.When the AP25 aggregates were well distributed into themonomer using AP25 ethanol organosol, the performance ofthe resins was improved dramatically. Ethanol acted as bothsolvent and mixing enhancer for even distribution of AP25 inthe monomers. In this study, DC, elastic modulus, hardnessand SBS were used as examples to illustrate the performanceimprovement of the dental resins after adding AP25 TiO2 NPs.The equal mass fraction mixture of BisGMA and TEGDMA waschosen as the base resin because they are commonly used inthe dental field, and their properties have been well studied[35].

3.2. Degree of vinyl conversion (DC)

The DC of the resins increased as a function of mass fraction ofAP25 NPs and reached a plateau at approximately 0.1% massfraction of AP25 NPs. Table 1 lists the DC of resins with differ-ent mass fractions of AP25 NPs. The DC was plotted againstmass fraction of AP25 NPs (Fig. 3). Without AP25, the DC ofresin was 84.8% ± 1.6%, which is high for BisGMA and TEGDMAresin with CQ and 4E as the photo-initiator [35,36]. When 0.02%

by 3%. One-way ANOVA indicated a significant difference ofDC between AP25-containing and AP25-free resins. The DC atthe plateau was ≈7% higher than that of the pure resin.

d e n t a l m a t e r i a l s 2 7 ( 2 0 1 1 ) 972–982 977

Fig. 2 – TEM images of AP25 NPs

Table 1 – Degree of vinyl conversion resins with differentmass fractions of AP25.

AP25% by mass DC STDEV (%)

0 84.80 1.590.02 87.78 1.020.03 89.58 0.780.05 88.08 1.550.06 88.53 1.12

wfDTr

FBNduA

0.08 89.57 1.880.5 91.65 0.79

The 3–7% improvement of DC was a small change but itas an important improvement. First, the DC was increased

rom an already high value (≈85%); second, the increased

C demonstrated the unique photo-induced activities ofiO2 NPs; third, the increased DC occurred under blue lightadiation which was interesting for these TiO2 NPs, whose

ig. 3 – Degree of vinyl conversion (DC) of the mixture ofisGMA and TEGDMA with different mass fraction of AP25Ps. The inset image indicates the transparency of theisks. The letter “A” that was written on a paper and putnder the disks is clear even though the mass fraction ofP25 in the resin is 0.08%.

at different magnifications.

photo-activities were normally activated under UV radiation;finally, the increased DC and the macro-crosslinked hardpoints in the network created by AP25 could make a hugedifference in terms of the number of crosslinks and polymerchain mobility.

The photoactivity of TiO2 NPs plays an important role forAP25 as a coinitiator [37]. The AP25 NPs themselves also ini-tiated the polymerization, but at a very slow rate. When noCQ and 4E were added, a resin monomer that contained 0.5%(by mass) AP25 turned into a soft gel after 2 min of blue lightradiation. The DC of this gel was 15.8% ± 6.4%. These resultsindicate that there exist free radicals on the surface of AP25,and these radicals are more likely generated in the ideal loca-tion for the bound acrylic acid on the NPs to crosslink into thebulk polymer network.

The picture in the inset of Fig. 3 indicates the transparencyof these resins. The letter “A” that was printed on a paperand put under the resin disks is still clear even when 0.08%(by mass) of AP25 NPs is added into the resin. Lower massfractions of AP25 NPs and thinner samples have good trans-parency. Cautions should be taken when resins with massfraction of AP25 greater than 0.1% and thicker than 1 mm areused, because the increased opacity could lead to insufficientcuring.

3.3. Modulus and hardness determined bynanoindentation

Nanoindentation is often used to measure the modulus andhardness of small volumes near the surface of materials. Inorder to determine whether the mechanical properties of thecomposites varied with depth and reduced the viscoelasticeffects on the indentation measurement, two different maxi-mum depths were investigated using the continuous stiffnessmeasurement (CSM) method. This method collects stiffness

data as a function of depth by superposing a sinusoidal oscil-lation over the loading profile. The stiffness is used in theHertz contact equations to calculate the elastic modulus of thematerials and the hardness is the stress beneath the indenter.

978 d e n t a l m a t e r i a l s 2

Fig. 4 – Nanoindentation (CSM) results at 500 nm maximumdepth and 1500 nm maximum depth; elastic modulus (E)and hardness (H) of resins containing different mass

sification, displacement, and fracture in the material beneath

fractions of AP25 NPs.

The contact area is determined as a function of depth usingindentations into a standard fused silica material [38].

Fig. 4 shows the elastic modulus (Fig. 4A) and hardness(Fig. 4B) of disks with an increasing mass fraction of AP25.According to the results from nanoindentation, the elasticmodulus and hardness of the resin are improved with theaddition of AP25 TiO2 NPs, but there is a maximum increasein properties. The elastic modulus exhibits a sharp increase,48% with increasing mass fraction of AP25 up to 0.06% massfraction at a maximum indentation depth of 500 nm. At a max-imum indentation depth of 1500 nm, a similar increase, 41%, inelastic modulus is observed at the 0.06% mass fraction. At thehighest mass fraction, the elastic modulus decreases which isexpected for a material that appears cloudy (low transparency)due to agglomerated particles. Overall, the elastic modulus forthe AP25-resins measured at two different indentation depthsagree fairly well.

Similar behavior is observed for the hardness measure-

ment, although the hardness approximately doubled for thetwo indentation depths. The maximum hardness was reachedwhen 0.06% mass fraction of AP25 was added, and it was

7 ( 2 0 1 1 ) 972–982

2.4 times higher than that of the pure resin. In general, thehardness for the 1500 nm maximum indentation depth wasgreater than that for the 500 nm maximum indentation depth.Since the elastic modulus indicated no significant differencesin mechanical properties with depth, this could be an indica-tion of slight errors (≈10%) in the contact radius at the largerdepth.

An optimal mass fraction of AP25 exists that permits theresin to reach its maximum elastic modulus and hardness,which depends on how evenly the AP25 NPs were distributedwithin the resin and the quantity of AP25 NPs added into theresin. TiO2 has a very high hardness and modulus, which con-tributes to the mechanical properties of the resin. On the otherhand, adding unevenly distributed AP25 TiO2 NPs will increasethe opacity of resins and may cause weak points in the resins,thus diminishes the advantage gained from the addition ofAP25 TiO2 NPs. As a result, the elastic modulus and hardnessdid not increase proportionally to the mass fraction of AP25,but instead those properties began to decrease after reach-ing the maximum value. Both elastic modulus and hardnesstesting indicate that the optimal mass fraction of AP25 TiO2

NPs is in the range of 0.06–0.1%. The density of titanium diox-ide is 4.23 g/cm3, and the corresponding volume fraction ofAP25 TiO2 NPs is 0.014% to 0.021%, which is very close to thevolume fraction of ethanol solution of AP25 TiO2 NPs aftercentrifugation.

3.4. Knoop hardness

The hardness determined through nanoindentation wascompared with those determined through conventionalmicroindentation machines at different loads. A Knoopmicroindenter was used because it makes easy-to-measureindentations [39]. For many materials, hardness decreaseswith increasing indentation load and eventually becomesapproximately constant at large loads.

The hardness of AP25 NP-resins decreases as the loadincreases. It became constant when the load was 2 N andhigher (Fig. 5A). This trend is typical in composites, ceram-ics and metals [40]. The hardness of NP-free resin decreasedfirst and then returned to reach constant at the load of 2 Nand higher, which was a trend that had been found on poly(methyl-methacrylate) [40]. As shown in Fig. 5, the hard-ness difference among resins with different mass fractionsof AP25 is constant when the indentation loads are 2 N andhigher. Fig. 5B shows the hardness of resins with variousmass fractions (0%, 0.03% and 0.08%) that are determined viananoindentation (load is 0.003 N) and Knoop hardness testingat indentation loads of 0.5 N and 4 N. The hardness deter-mined via nanoindentation and Knoop hardness at differentindentation loads all indicate that there exists a maximumhardness as a function of AP25 mass fraction, and the optimalmass fraction of AP25 NPs is in the range from approximately0.06% to 0.1%. Caution should be taken to interpret these data:normally, different hardness instruments produce differenttrends due to various amounts and types of deformation, den-

the indenters. Nanoindentation is a quasi-static method usinga spherical indenter tip while the Knoop hardness test is astatic method using an elongated pyramidal tip. In this study,

d e n t a l m a t e r i a l s 2 7 ( 2 0 1 1 ) 972–982 979

0 2 4 60.10

0.15

0.20

0.25

0.30

0.35

0% AP25

0.03% AP25

0.08% AP25

H [

GP

a]

Load [N]

A

0.0 0.1 0.2 0.3 0.4 0.5

0.2

0.3

TiO2 [wt%]

Load = 4 N

Load = 0.5 N

Load = 0.003 N

H [

GP

a]

B

Fig. 5 – Hardness of resins at different indentation loads.(A) Same resin (the concentration of AP25 NPs was 0%,0.03% or 0.08% by mass) under different indentation loadsand (B) same load on resins with different amount of AP25NPs. The load of 0.003 N was carried out usingnanoindentation. All of the other loads were done usingt

btroitmtev

3

Fnmsmw

2

4

0.50.080.060.050.03

Mo

du

lus [G

Pa

]

TiO2 [ wt%]

Microindentation

Nanoindentation

3-point bending

0

*

Fig. 6 – Moduli of resin with different mass fractions ofAP25 determined using nanoindentation,microindentation, and 3-point bending. Disks (8 mm indiameter and 1 mm in thickness) were used to determineelastic modulus by nanoindentation and microindentation.The (*) sample required a thinner film to visualize thecontact area due to low transparency of the sample. Resinbars were used to determine the flexural modulus by

he Knoop hardness method.

oth methods showed a similar trend of hardness as a func-ion of AP25 mass fraction. This optimal mass fraction mighteflect the distribution of AP25 NPs in the resins. Above theptimal mass fraction, large aggregates may form, resulting

n resin composites that have decreased mechanical proper-ies. Resins with evenly distributed AP25 NPs at the optimal

ass fraction tended to have greater hardness than those con-aining poorly distributed AP25 NPs. The decrease of both thelastic modulus and hardness after reaching the maximumalues may also caused by increasing opacity.

.5. Modulus determined by three methods

ig. 6 lists moduli determined using three methods:anoindentation, microindentation, and 3-point bending. Asentioned in Section 3.3, the increase in elastic modulus mea-

ured by nanoindentation is greater than that predicted byicromechanics [41]. For comparison, the elastic modulusas also measured by microindentation (thin film indenta-

3-point bending.

tion), which used a much larger contact area and samplevolume to measure the elastic modulus. Microindentationused a larger indenter (12.7 mm in diameter), which increasedthe contact area (the diameter of contacting area is up to100 �m) and measured the average elastic modulus of alarger substrate volume than for nanoindentation. While bothmeasure elastic modulus, nanoindentation is limited sam-pling polymer near the surface, while the microindentationmeasurement penetrates deeper into the bulk substrate. Inaddition, 3-point bending tests were used to measure theflexural modulus of resins. The two indentation methods mea-sured the moduli using contacting surface area from 1 �m upto 100 �m in diameter, and they agree well with each other interms of the trend of the elastic modulus as a function of massfraction of AP25, but microindentation shows a lower modu-lus than nanoindentation. This difference could be due to thesampling volumes, if the surface has a slightly higher conver-sion than the overall bulk. The trend of elastic modulus as afunction of mass fraction of TiO2 NPs is also in good agreementwith the trend of hardness determined by both nanoindenta-tion and Knoop indentation. Larger than expected increasesin mechanical properties have been demonstrated when den-tal fillers contain functionalities which can crosslink withthe resin material [42]. In this system, acrylic acid doublebonds attached to AP25 NPs would provide a similar crosslinkbetween the nanoparticles and resin.

The moduli of TiO2-free resins determined by these threemethods are in good agreement. But as the mass fraction ofAP25 increase, the trend of elastic modulus is different from

the flexural modulus determined using 3-point bending whichevaluated the bulk properties of the materials. The flexuralmodulus of resins reaches maximum at a lower mass fraction

l s 2 7 ( 2 0 1 1 ) 972–982

0

5

10

15

20

25

30

0.50.120.10.080.020

Shear

bond s

trength

[M

Pa]

Control

TiO2 [wt%]

Fig. 7 – Shear bond strength (SBS) when resins at four massfractions of AP25 were used as dentin adhesives.Scotchbond (3M ESPE, St. Paul, MN, USA) was used as the

980 d e n t a l m a t e r i a

(0.05%) of AP25, which is approximately 16% higher than thatof the TiO2-free resin. According to ANOVA, no significant dif-ferences were found among flexural modulus of other resinscompositions. In addition, the error bars of flexural modulus,especially at the mass fraction of 0.5%, indicate large uncer-tainty in the samples, which might be caused by insufficientcuring due to the low transparency of the samples.

In order to achieve the DC equivalent to those of the resindisks used in indentation, the light radiation time had to beincreased from 1 min per side to 2 min per side. Even 2 min perside was still not sufficient for resins with high mass fractionof TiO2 NPs including 0.08% and 0.5%, but for consistency, allof the resin bars were cured 2 min per side.

Without extending the irradiation time, significantdecrease in flexural modulus was apparent. The flexuralmodulus of resin with 0.05% mass fraction of AP25 NPsafter both 1 min per side and 2 min per side light radiationwas determined. The modulus from 3-point bending testsafter 1 min per side light radiation (1.1 ± 0.1 GPa) is onlyapproximately one third of its 2 min per side counterpart(3.3 ± 0.2 GPa). Without a consistent method to polymerizelow transparency resins, mechanical tests that require thicksamples or high mass fraction of AP25 NPs will be difficult toperform.

The great improvement of modulus and hardness are morelikely benefit from the excellent mechanical properties andunique photoactivities of TiO2 NPs. In addition, the attacheddouble bonds on AP25 NPs ensure that the NPs are fastenedinto the resin network. In return, the AP25 NPs created micro-crosslinked hard points in the network. The AP25 NPs couldalso generate more crosslinks. Even if two acrylates on a singleAP25 particle react, additional crosslinks are produced. Thehigher the DC reaches, the more crosslinks will form. In thiscase, the AP25 NPs are not simple additives to the resins, theymake a big difference in terms of the number of crosslinks andpolymer chain mobility.

3.6. Shear bond strength (SBS)

In order to avoid low transparency and to demonstrate clini-cal applications of these AP25 NP-resins, the AP25-containingresins were used as dental adhesives to bond tooth and com-posites. The bonding agents were generally a 50 �m layer ofresin between composites and dentin. A gold-standard three-step bonding procedure (acid etching, priming and dentaladhesive application) was used to adhere resin composite tothe ground dentin surface. In order to optimize the numberof samples, the experimental resins were applied only in thethird step as bonding agents. The materials and methods inthe first two steps of the bonding procedure were developedat our research center [32,33]. Total six mass fractions of AP25(0%, 0.02%, 0.08%, 0.10%, 0.12%, and 0.50%) were tested. TheSBS results are shown in Fig. 7. The resins with 0.02% and0.5% mass fractions of AP25 have the same mean SBS valueas that of the pure resin. The rest compositions generatedhigher SBS than the TiO2-free resin. The 0.1%-AP25 resin has

the maximum mean SBS value (24.5 ± 6.0 MPa) that is approx-imately 30% higher than that of the pure resin. As expected,a larger amount of AP25 NPs in the dental adhesive did notproduce stronger SBS due to the agglomeration of NPs and

control.

increased opacity. When apply the resin with 0.5% mass frac-tion of AP25 as bonding agent, it was hard to make smoothsurface due to its high viscosity and large amount of particles.Large agglomerates of NPs in the 0.5%-AP25 resin monomersformed precipitates in 2 days, while no precipitates were foundin the 0.1%-AP25 resin monomers after 3 months. This indi-cates a good dispersion of AP25 NP in the monomers at themass fraction of 0.1%.

The SBS test is a simple, often-used screening test for thedental adhesives. The experimental dental adhesives in thisstudy were subjected to loads causing adhesive failure at thedentin–adhesive interface or an occasional cohesive failure inthe dentin. The debonding loads of the composite from thedentin surface were in the range of 100–450 N. Under suchhigh loads, the SBS of AP25 NP-resin adhesives were equal oreven 30% greater than that of NP-free resin adhesive and thecommercial control. When dentin adhesives are used in con-junction with composite restorative materials to restore teethin clinical practice, the restorations, are subjected to loadsgenerated by biting, grinding and chewing that are generallylower than the maximum loads. As dentin adhesives, the AP25NP-resins have not only higher (or equal) SBS but also greaterelastic modulus and hardness than the NP-free resins. The lab-oratory performance of NP-resins demonstrated in this studycould lead to improved dental adhesives for clinical applica-tions.

4. Conclusions

In summary, the commercially available NPs (P25) were suc-cessfully modified by attaching a short chain acid withcarbon–carbon double bonds. These acid modified TiO2 NPs(AP25 NPs) were agglomerates with relatively narrow size dis-tribution in ethanol. The performance of the studied dental

resins was improved dramatically by adding a small amountof AP25 NPs: (1) the DC of the mixture of BisGMA and TEGDMAwas improved by approximately 5% by adding of 0.08% mass

2 7

ftam(m

fffrleA

bietdii

A

FtAhw1

r

d e n t a l m a t e r i a l s

raction of AP25 NPs; (2) the elastic modulus and hardness ofhe BisGMA and TEGDMA mixture was enhanced by ≈48% bydding 0.06% mass fraction of AP25 NPs, and its hardness wasore than doubled by adding 0.06% (mass fraction) of AP25;

3) the mean SBS was increased approximately 30% when 0.1%ass fraction of AP25 NPs was added to the dentin adhesive.An optimal mass fraction of AP25 NPs for the best per-

ormance of the resins was established. This optimal massraction is likely the maximum amount of AP25 NPs withoutorming large aggregates. The effectiveness of the AP25 NPs iselated to the distribution of the NPs and the amount of acry-ate functionalities on the surface of TiO2 NPs. For resins withvenly distributed AP25 NPs, the higher the concentration ofP25 NPs was the better the modulus and hardness.

Above all, the enhanced performance of the dental resinsenefits from the strong mechanical properties, photoactiv-

ty, and crosslinking potential of modified TiO2 NPs. Furtherxperiments are needed to understand the mechanism behindhe great enhancement in the mechanical properties. Theurability and stability of the TiO2-containing resins are also

mportant to know in order to apply them as dental adhesivesn clinical practice.

cknowledgements

inancial support was provided through an ADA Founda-ion startup lab fund. We would like to thank Drs. Josephntonucci, Sabine H. Dickens and Rafael L. Bowen for theirelpful discussions. The FTIR-RM that was used in thisork was purchased with NIH Shared Instrumentation Grant

S10RR22650-01A2 awarded to N.E.

e f e r e n c e s

[1] Chen X, Mao SS. Titanium dioxide nanomaterials: synthesis,properties, modifications, and applications. Chem Rev2007;107(July (7)):2891–959.

[2] Carp O, Huisman CL, Reller A. Photoinduced reactivity oftitanium dioxide. Prog Solid State Chem 2004;32(1–2):33–177.

[3] Mills A, Lee SK, Lepre A. Photodecomposition of ozonesensitised by a film of titanium dioxide on glass. JPhotochem Photobiol A 2003;155(February (1–3)):199–205.

[4] Mills A, LeHunte S. An overview of semiconductorphotocatalysis. J Photochem Photobiol A 1997;108(July(1)):1–35.

[5] Zhang ZB, Wang CC, Zakaria R, Ying JY. Role of particle sizein nanocrystalline TiO2-based photocatalysts. J Phys Chem B1998;102(December (52)):10871–8.

[6] Feng XJ, Zhai J, Jiang L. The fabrication and switchablesuperhydrophobicity of TiO2 nanorod films. Angew ChemInt Ed 2005;44(32):5115–8.

[7] Fujishima A, Honda K. Electrochemical photolysis of waterat a semiconductor electrode. Nature 1972;238(5358):37–8.

[8] Koziej D, Fischer F, Kranzlin N, Caseri WR, Niederberger M.Nonaqueous TiO2 nanoparticle synthesis: a versatile basisfor the fabrication of self-supporting, transparent, andUV-absorbing composite films. ACS Appl Mater Interfaces

2009;1(May (5)):1097–104.

[9] Ferracane JL. Current trends in dental composites. Crit RevOral Biol Med 1995;6(November (4)):302–18.

( 2 0 1 1 ) 972–982 981

[10] Van Landuyt KL, Snauwaert J, De Munck J, Peurnans M,Yoshida Y, Poitevin A, et al. Systematic review of thechemical composition of contemporary dental adhesives.Biomaterials 2007;28(September (26)):3757–85.

[11] Breschi L, Mazzoni A, Ruggeri A, Cadenaro M, Di Lenarda R,Dorigo ED. Dental adhesion review: aging and stability of thebonded interface. Dent Mater 2008;24(January (1)):90–101.

[12] De Munck J, Van Landuyt K, Peumans M, Poitevin A,Lambrechts P, Braem M, et al. A critical review of thedurability of adhesion to tooth tissue: methods and results. JDent Res 2005 Feb;84(2):118–32.

[13] Yu B, Ahn JS, Lim JI, Lee YK. Influence of TiO2 nanoparticleson the optical properties of resin composites. Dent Mater2009;25(September (9)):1142–7.

[14] Klapdohr S, Moszner N. New inorganic components fordental filling composites. Monatsh Chem 2005;136(January(1)):21–45.

[15] De Palma R, Peeters S, Van Bael MJ, Van den Rul H, Bonroy K,Laureyn W, et al. Silane ligand exchange to makehydrophobic superparamagnetic nanoparticleswater-dispersible. Chem Mater 2007;19(April (7)):1821–31.

[16] Frankamp BL, Fischer NO, Hong R, Srivastava S, Rotello VM.Surface modification using cubic silsesquioxane ligands.Facile synthesis of water-soluble metal oxide nanoparticles.Chem Mater 2006;18(February (4)):956–9.

[17] Rajh T, Saponjic Z, Liu JQ, Dimitrijevic NM, Scherer NF,Vega-Arroyo M, et al. Charge transfer across thenanocrystalline-DNA interface: probing DNA recognition.Nano Lett 2004;4(June (6)):1017–23.

[18] Vo DQ, Kim EJ, Kim S. Surface modification of hydrophobicnanocrystals using short-chain carboxylic acids. J ColloidInterface Sci 2009;337(September (1)):75–80.

[19] Koppel DE. Analysis of macromolecular polydispersity inintensity correlation spectroscopy—method of cumulants. JChem Phys 1972;57(11):4814–20.

[20] Riseman J, Kirkwood JG. The intrinsic viscosity, translationaland rotatory diffusion constants of rod-likemacromoleculres in solution. J Chem Phys 1950;18(4):512–6.

[21] Provencher SW. A constrained regularization method forinverting data represented by linear algebraic orintegral-equations. Comput Phys Commun1982;27(3):213–27.

[22] Provencher SW. CONTIN: a general-purpose contrainedregularization program for inverting noisy linear algebraicand integral-equations. Comput Phys Commun1982;27(3):229–42.

[23] Eidelman N, Raghavan D, Forster AM, Amis EJ, Karim A.Combinatorial approach to characterizing epoxy curing.Macromol Rapid Commun 2004;25(January (1)):259–63.

[24] Lin NJ, Drzal PL, Lin-Gibson S. Two-dimensional gradientplatforms for rapid assessment of dental polymers: achemical, mechanical and biological evaluation. Dent Mater2007;23(10):1211–20.

[25] Sun J, Eidelman N, Lin-Gibson S. 3D mapping ofpolymerization shrinkage using X-ray micro-computedtomography to predict microleakage. Dent Mater2009;25(March (3)):314–20.

[26] Eidelman N, Simon CG. Characterization of combinatorialpolymer blend composition gradients by FTIRmicrospectroscopy. J Res Natl Inst Stand Technol2004;109(March–April (2)):219–31.

[27] VanLandingham MR. Review of instrumented indentation. JRes Natl Inst Stand Technol 2003;108(July–August (4)):249–65.

[28] Masouras K, Silikas N, Watts DC. Correlation of filler contentand elastic properties of resin-composites. Dent Mater

2008;24(July (7)):932–9.

[29] Nunes RAX, Costa VC, Calado VMD, Branco JRT. Wear,friction, and microhardness of a thermal sprayed PET—poly

l s 2

982 d e n t a l m a t e r i a

(ethylene terephthalate) coating. Mater Res-Ibero-Am JMater 2009;12(April–June (2)):121–5.

[30] Johnson PM, Stafford CM. Surface indentation arrays forhigh-throughput analysis of viscoelastic material properties.Rev Sci Instrum 2009;80(October (10)):8.

[31] Johnson PM, Howarter JA, Stafford CM. Time-dependentrelaxations in thin hydroxypropyl cellulose coatings: therole of buried interfaces. Macromolecules 2011;44(9):3529–34.

[32] Hussain LA, Dickens SH, Bowen RL. Shear bond strength ofexperimental methacrylated betacyclodextrin-basedformulations. Biomaterials 2005;26(June (18)):3973–9.

[33] Dickens SH, Milos MF. Relationship of dentin shear bondstrengths to different laboratory test designs. Am J Dent2002;15(June (3)):185–92.

[34] Liu G, Wang XW, Chen ZG, Cheng HM, Lu GQ. The role ofcrystal phase in determining photocatalytic activity ofnitrogen doped TiO2. J Colloid Interface Sci 2009;329(January(2)):331–8.

[35] Sun JR, Fang R, Lin N, Eidelman N, Lin-Gibson S.

Nondestructive quantification of leakage at thetooth–composite interface and its correlation with materialperformance parameters. Biomaterials 2009;30(September(27)):4457–62.

7 ( 2 0 1 1 ) 972–982

[36] Sun J, Lin-Gibson S. X-ray microcomputed tomography formeasuring polymerization shrinkage of polymeric dentalcomposites. Dent Mater 2008;24(2):228–34.

[37] Wang J, Ni XY. Interfacial structure of poly(methylmethacrylate)/TiO2 nanocomposites prepared throughphotocatalytic polymerization. J Appl Polym Sci2008;108(June (6)):3552–8.

[38] Oliver WC, Pharr GM. Measurement of hardness and elasticmodulus by instrumented indentation: advances inunderstanding and refinements to methodology. J Mater Res2004;19(January (1)):3–20.

[39] Kinney JH, Marshall SJ, Marshall GW. The mechanicalproperties of human dentin: a critical review andre-evaluation of the dental literature. Crit Rev Oral Biol Med2003;14(January (1)):13–29.

[40] Amitay-Sadovsky E, Wagner HD. Evaluation of Young’smodulus of polymers from Knoop microindentation tests.Polymer 1998;39(May (11)):2387–90.

[41] Kinloch AJ, Taylor AC. The mechanical properties andfracture behaviour of epoxy-inorganic micro- andnano-composites. J Mater Sci 2006;41(June (11)):3271–97.

[42] Wilson KS, Allen AJ, Washburn NR, Antonucci JM.Interphase effects in dental nanocomposites investigated bysmall-angle neutron scattering. Wiley Subscription Services,Inc., A Wiley Company; 2007, pp 113–23.