Journal of Colloid and Interface Science238,136–146 (2001)doi:10.1006/jcis.2001.7506, available online at http://www.idealibrary.com on

Aqueous Amino Silane Modification of E-glass Surfaces

Hazel Watson,1 Anne Norstrom, Aasa Torrkulla, and Jarl Rosenholm

Department of Physical Chemistry, Aabo Akademi University, Porthansgatan 3-5, A

abo 20500, Finland

Received September 22, 2000; accepted February 27, 2001

The majority of work available in the literature examinesthe effect of epoxy silane, γ -aminopropyltrimethoxy silane,and methacrylate silane on E-glass surfaces. As alternatives tothe most commonly used silanes, we investigated two novelsilanes: γ -ureidopropyltriethoxy silane and Nβ(aminoethyl) γ -aminotrimethoxy silane and additionally an amino-functionalpolysiloxane. The ureido silane-treated E-glass fibers demonstrateda zeta potential similar to that of the untreated E-glass fibers, whichwas independent of deposition solution pH over the pH range in-vestigated. A moderately hydrophobic E-glass surface, which wassilane concentration dependent, was noted as being due to con-densed Si–O–Si bonds at the surface. The diamino silane demon-strated an extremely basic surface at the higher silane concentra-tions investigated. These outer surface layers were modified bychanging the pH and the concentration of the deposition solution.The polysiloxane produced an increase in the hydrophobicity ofthe E-glass fiber, especially when deposited from extremely basicsolutions. At the higher solution concentrations investigated, thesilanes and the siloxane was initially deposited in patches and anouter surface or “skin” was formed over these patches, giving theappearance of fully coated fibers. C© 2001 Academic Press

Key Words: silane; siloxane; zeta potential; contact angle; DRIFT;E-glass.

ilnm

tilb

t

sl

atendraly-

te,be-owand

n ofthe

lox-hy-

sur-thethesehes

fer-n be

arlyex-ed toface. Ind byaryin-

wasdif-tions

on-ngethe

INTRODUCTION

It has long been known that treating the surface of a sfiller (colloidal, particulate, or fibrous) with a coupling ageimproves the physical and chemical properties of a polymertrix filled with such materials (1–4). The match between pomer matrix and silane organofunctional group type and lenhas a bearing on the mechanical properties of the producPark and Subramanian suggested that if the silane was hcrosslinked, interpenetration of the silane/siloxane would beited, leading to reduced interfacial properties of a polymer/fimatrix (6).

Acid concentration (i.e., pH of the treatment solution) andcuring conditions used, can affect the molecular weight ofdeposited silane oligomers and the amount of silane depoonto the surface (7–9). The rate of hydrolysis of aminopropy

1 To whom correspondence should be addressed. Fax: +358 2 215 4E-mail: hazel.watson@abo.fi.

ion,

130021-9797/01 $35.00Copyright C© 2001 by Academic PressAll rights of reproduction in any form reserved.

icata-

ly-gth(5).ghlyim-er

hetheited

tri-

706.

ethoxy silane (APTES) was at a minimum at pH 7, while the rof condensation was at a minimum at pH 4 (10). Premachaet al. demonstrated a similar pH dependency for the hydrosis ofγ -ureidopropyltrimethoxy silane. They went on to nohowever, that condensation of this silane was at a maximumtween pH 9 and pH 10 (11). Leyden and Atwater showed hthe organic functional group affected both the condensationthe hydrolysis reactions (12).

Drying the E-glass fibers accelerated the condensatiothe surface silanols with the silane silanols, enhancingchemisorption. Washing the E-glass fibers stabilized the siane layer, thus improving the hydrolysis resistance and thedrophobicity of the surface (13, 14).

Research showed that ureido silanes were bonded to theface by only a few Si–O–Si linkages (15). The remainder ofsilane was present as a highly crosslinked mass attached tofew surface bonded silanes, resulting in firmly attached patcof a 3D polysiloxane network.

Silane manufacturers recommend silanes for use with difent surfaces on the basis that the organo-functional group caexpected to react with proposed matrix polymers (16, 17). Eresults indicated that the ureido silane produced entirely unpected results, with a condensed Si–O–Si surface as opposa ureido surface. Thus investigation of whether a ureido suris achievable from this silane warrants further investigationorder to ensure that the siloxane surface was not produceour deposition method, a silane with primary and secondamino functions (diamino silane) and a polysiloxane werevestigated alongside the ureido silane. The aim of this workto investigate the nature of the E-glass fiber surface withferent silane-based treatments, using a range of concentraand pHs.

EXPERIMENTAL

We investigated the effect of treatment solution pH and ccentration on two cationic silanes and a polysiloxane. The rachosen for investigation was pH 4 to pH 7, and additionallypolysiloxane molecule was investigated at pH 10.5.

Materials

Chopped industrial grade E-glass fibers and an emulssiloxane AR433, were kindly supplied by Ahlstr¨om Glassfibre

6

F

.

n

n4

aite

-

-l-

n-

ere

4,d

nd26ly,eerble

bens

edensd in

AQUEOUS AMINO SILANE MODI

TABLE 1Composition of E-glass Fibers Used

E-Glass, oxide Elements as determinedconstituents wt% by XPS %

SiO2 55 O1s 52.5Al2O3 + Fe2O3 14.5 C1s 21CaO 21.5 Si2p 17.4MgO 0.5 Ca2p 3.4Na2O+ K2O <1.0 B1s 0.4B2O3 7.5 Al2p 5.6

Oy, Karhula, Finland. The E-glass fibers were of diameter 11×10−6 m and length between 0.8× 10−3 and 1.0× 10−3 m(Table 1). The E-glass fiber samples were water cooled,mediately after the E-glass fibers were drawn from the spinbush.

The cationic aminosilanes used were,γ -ureidopropyltri-ethoxy silane (A1160) and Nβ-(aminoethyl)γ -aminopropyl-trimethoxy silane (A1126) (Fig. 1). The manufacturers didreveal the chemical structure of siloxane Flexichem AR(Fig. 1), which was a milky emulsion with a siloxane conteof approximately 37%, and an undisclosed amount of ethand methanol. Aspokem Oy (Finland) kindly donated the Wsilanes. All the silane products were used without further trment.

FIG. 1. Schematic structures for (a) silane A1160, (b) silane A1126, a(c) siloxane AR 433.

ICATION OF E-GLASS SURFACES 137

0

im-ing

ot33ntnolcoat-

TABLE 2pH and pKa Values of Aqueous Solutions of Silanes A1160

and A1126 and Siloxane AR433

A 1160 A1126 AR433

pH of a 0.15% solution 8.6 10.1 9.4pKa 8.6 9.4 9.0

Distilled water was passed thorough a Millipore Milli Q academic 10 deioniser to produce deionized water of 18.2 MÄ · cmat 25◦C (pH 5.8). Glacial acetic acid (99–100% “Baker Analysed” ACS reagent) and ammonium hydroxide (“Baker Anaysed” [AR] 25% NH4OH) were obtained from J. T. Baker.

Methods

Before treatment, the E-glass fibers were sonicated in deioized water for 10 min and then vacuum dried at 50◦C and 200 Torrfor 24 h. The silanes were added to water, and the pH and pKa

values were determined in aqueous solutions, the solutions wused immediately after pH adjustment (Table 2).

Aqueous solutions (400 ml) of each silane were made at pHpH 5, pH 6, and pH 7 using acetic acid to adjust the pH anadditionally AR433 solutions were altered to pH 10.5 with 1 Mammonium hydroxide. Concentrations of 0.005, 0.01, 0.10, a0.15 wt% were used for each material. Silanes A1160 and A11were supplied as 50 and 40% solutions in methanol, respectivewhile the siloxane AR433 had a solids content of 37%. Thamount of silane in each of the solutions was actually much lowthan the concentrations noted above. To enable a reasonacomparison to be made among the three materials they willreferred to as 0.005, 0.01, 0.10, and 0.15 wt% of silane solutio(Table 3).

E-glass fibers (30 g) were added to each solution and stirrfor 20 min. The solution was decanted and the E-glass fibers ovdried at 120◦C for 10 min, to condense the silane with the glassurface. The E-glass fibers were then washed and sonicate

TABLE 3Actual Weight Percentages of Silanes A1160 and A1126

and Siloxane AR433 Used

Weight as supplied Actual silane content Molar silane(wt%) (wt%) content

A 1160 0.005 0.0025 9.64× 10−5

0.01 0.005 1.89× 10−4

0.1 0.05 1.89× 10−3

0.15 0.075 2.84× 10−3

A 1126 0.005 0.0019 8.9× 10−5

0.01 0.0037 1.78× 10−4

0.1 0.037 1.78× 10−3

0.15 0.055 2.67× 10−3

AR 433 0.005 0.002 Proprietary0.01 0.004 formulae

nd0.1 0.040.15 0.059

N

e

ane

p

e

t

m

iatr

idla

ni

sla

s on

inalnoton,he

0)lassaneen-O–Siosi-

cor-ilerel-

pH.8,

de-ntial

oretion

tothe7

en,. Atas

ted

hatsedof

ver,otedat

e

138 WATSO

deionized water for 3 min to remove loosely bound silane (1After careful filtration the E-glass fibers were dried at 60◦C for24 h in a vacuum oven at 200 Torr.

Premachamdraet al. demonstrated that hydrolysis ofγ -ureidopropyltrimethoxy silane in a water/methanol medium wextremely rapid at pH 4.87 and pH 9.83. Between these extrea minimum rate of hydrolysis occurred at pH 7.44. They won to qualitatively demonstrate that condensation was at a mmum between pH 8.97 and pH 9.87 (11). At lower pHs a prepderance of silanol groups existed which condensed at a mlower rate. Thus over the pH range examined it is reasonto suggest that during the 20-min period, for which the silawere allowed to hydrolyze, little or no condensation occurrprior to deposition onto the E-glass fibers.

Streaming Potential (Zeta Potentialζ )

Streaming potential was expedited using an Anton Parr Etrokinetic Analyser (EKA), with an automated pH titration aparatus, which has been described elsewhere (19, 20).

The electrolyte solution used was 1.0 E+ 03 M KCl, at aconductivity of 2.7 E+ 01 mS/m at 23± 2◦C. Starting with theoriginal solution the pH of the electrolyte solution was adjustHigher pH values were obtained by the addition of 0.25 M KOand lower pH values by the addition of 0.25 M HCl. One sconformed to 0.5 of a pH unit. Two grams of carefully packE-glass fiber sample was used for each measurement. Sipacking density and overall consistency were required forery test; any differences could result in poor reproducibilThe equipment was carefully washed with KCl between etest. Mathematical treatment of the data used the FairbroMastin approximation (19, 20). A minimum of three measuments were carried out, to ensure reproducibility.

Contact Angle

A KSV Instruments Sigma 70 Wilhelmy balance was usto measure the advancing and receding contact angles. A sfiber was attached, vertically, to a wire hook, using double-siadhesive tape. The hook was suspended from an electrobawhich had a range of 0.25 mN and a resolution of 0.05µN. Thewetting liquid used was deionized water (pH 5.8). The beawas moved in the vertical direction by a DC motor; the balabeing automatically zeroed before each measurement. By raand lowering the beaker at a known rate, receding and advancontact angles were measured (21, 22).

X-Ray Photoelectron Spectroscopy (XPS)

XPS was carried out on a Perkin Elmer PHI 5400 apparausing an MgKα source. A 1-mm-diameter molybdenum mawas used to maintain the orientation and position of the E-gfibers and thus prevent shadowing. All the silanes were ex

ined at pH 4 and pH 7 and at concentrations 0.005 and 0.1Siloxane AR433 solutions were also investigated at pH 10.5

ET AL.

8).

asmesntaxi-on-uchbleesd,

lec--

d.H

eped

ilarev-ty.ch

her-e-

edngleednce,

kercesingcing

tus,k

assm-

RESULTS

Titration of the silanes against HNO3 produced stoichiometricbalance points, which were assigned to the functional groupthe silanes and the siloxane (Table 4).

Silane A1160

The stoichiometric balance point was assigned as the termureido group (Table 4). Measured IEPs (isoelectric point) didmatch the stoichiometric balance point measured by titratiindicating that there were no ureidic functional groups at tsurface (Table 5 and Fig. 2).

The two lowest concentrations of the ureido silane (A116reduced the contact angle to below that of the untreated E-gfibers; i.e., the surface became more hydrophilic due to silsilanols at the surface. At higher deposition solution conctrations, the contact angle increased as a condensed Si–surface formed. Contact angle was also independent of deption solution pH (Table 6).

At the base plateau (pH 9.7 of the zeta potential curves) arelation with contact angle was noted, which indicated that whhydrophobicity was augmented the surface charge remainedatively constant as pH rose (Fig. 3a). As deposition solutionincreased from 4 to 6, the zeta potential, at electrolyte pH 5correlated with contact angle gave similar results. The pH 7position solution demonstrated that an increase in zeta poteand contact angle occurred as the silane solution became mconcentrated (Fig. 3b). This was due to enhanced condensaof SiO at pH 7.

At high silane concentrations, there was sufficient nitrogenbe able to resolve the ESCA spectra into the peaks due todifferent nitrogen binding energies (Table 7). At pH 4 and pHfree nitrogen accounted for less than 10% of the total nitrogthe free nitrogen being assigned as the secondary aminepH 7 the ratio of hydrogen bonded to protonated nitrogen walmost 11:1. At pH 4 the ratio of hydrogen bonded to protonanitrogen had decreased to 1:1.

DRIFT versus iep and contact angle results confirmed tincreased deposition solution concentration produced increadeposition without appreciably changing the surface chargethe treated fibers (Figs. 4a and 4b). The contact angle, howeincreased markedly. Minor changes in surface charge were nat deposition solution pH 4 and low concentration, the point

TABLE 4Stoichiometric Balance Points of Silanes A1160 and A1126

and Siloxane AR433, Titrated against Nitric Acid

Stoichiometric balance Stoichiometric balancpoint 1 point 2

Silane A1160 6.2 (ureido group)

5%..

Silane A1126 8.1 (diamine group) 4.9 (silane triols)AR433 8.0 (amine group) 6.2 (benzene ring)

0.15

9.3

.15

AQUEOUS AMINO SILANE MODIFICATION OF E-GLASS SURFACES 139

TABLE 5IEP Values for Silanes A1160 and A1126 and Siloxane AR433

pH 4 pH 5 pH 6 pH 7 pH 10

IEP 0.005 0.01 0.1 0.15 0.005 0.01 0.1 0.15 0.005 0.01 0.1 0.15 0.005 0.01 0.1 0.15 0.005 0.01 0.1

A1160 4.7 4.4 4.6 4.5 4.7 5.0 4.6 5 4.4 4.1 4.4 4.4 4.4 4.2 4.4 4.7A1126 5.2 5.0 7.8 8.9 5.9 6.3 7.2 8.8 6.3 6.8 8.9 8.8 6 6.8 8.8 9.2AR433 5.4 5.2 5.9 5.7 5.5 6.0 6.5 6.3 5.4 6.0 6.9 6.6 5.6 6.0 7.3 7.0 4.5 4.5 8.3

TABLE 6Advancing Contact Angles for Silanes A1160 and A1126 and Siloxane AR433

Silane A1160 Silane A1126 Polysiloxane AR433

pH 0.005 0.01 0.1 0.15 0.0025 0.01 0.1 0.15 0.005 0.01 0.1 0

4 35 40 46 52 58 62 63 63 78 78 88 945 37 42 61 61 50 54 59 60 78 90 94 976 43 47 57 57 59 58 64 65 79 85 93 1017 38 41 54 53 41 47 64 63 78 84 94 101

10.5 49 60 104 107

FIG. 2. Zeta potential traces for silane A1160-treated E-glass fibers at four concentrations: 0.005, 0.01, 0.1 and 0.15 wt%; (a) deposition pH 4; (b) depositionpH 5; (c) deposition pH 6, and (d) deposition pH 7.

N

c

stres

n

a

as

eenal-idic,

sss-andd 6the

0.the

ho-uterandiday

140 WATSO

FIG. 3. (a) Advancing contact angle (◦) versus zeta potential at the eletrolyte base plateau (electrolyte pH 9.7) for silane A1160. (b) Advancing conangle (◦) versus zeta potential at electrolyte pH 5.8. The solid line indicatecreasing contact angle and constant zeta potential with increasing concenbetween deposition solution pH 4 and pH 6. The dotted line indicates thcrease in both zeta potential and contact angle at deposition pH 7. ForA1160.

which condensation would be expected to be at a minim(10, 11).

Silane A1126

Zeta Potential results indicated that as the silane concetion was increased, the outer layer consisted mainly of amfunctionality from the deposited silane. The amino functionity adsorbed H+ from the solution, producing NH+3 . The surface

TABLE 7Ratios of Hydrogen-Bonded, Protonated, and Free Nitrogen

for Silane A1160 (Measured by XPS)

Hydrogen- ProtonatedA1160 pH Free N% bonded N% N%

0.15% 4 6.4 47.2 46.20.15% 7 9.2 83.2 7.6

ET AL.

-tactin-ationin-

ilane

um

tra-inol-

FIG. 4. (a) Correlation of IEP and quantity of silane A1160 depositedmeasured by DRIFT. (b) Correlation of contact angle (◦) and quantity of silaneA1160 deposited as measured by DRIFT.

was basic, with highly negative zeta potentials and IEPs betw8 and 9 (Figs. 5a–5d). Titration of silane A1126 gave two bance points, the basic amine at pH 8.1 and a second, more acpoint, possibly that of the silanol triols, at pH 4.9.

At low silane concentrations metal oxides inherent to gla(SiO−, Al2O−3 , CaO−) were present in large concentrations. Disociation of the metal oxides, produced an acidic surfacehence a negative zeta potential, with an IEP of between 5 an(Figs. 5a–5d). The IEP moved toward more basic values assurface concentration of amino groups increased.

As a surface modifier A1126 was more efficient than A116The maximum changes in contact angle were achieved withtwo lowest concentration solutions, both producing a hydropbic surface, due to the amino groups being located on the osurface. Increase in contact angle was linear between pH 4pH 6. Low silane concentration at deposition solution pH 7 dnot increase the hydrophobic nature of the surface. This m

have been due to a mixed amino, silanol, and silicate surfaceand the deposition of oligomers (Table 6). At electrolyte pH 9.7

AQUEOUS AMINO SILANE MODIFICATION OF E-GLASS SURFACES 141

t deposit

th

llrion

tit

%

beon-ion

FIG. 5. Zeta potential traces for silane A1126-treated E-glass fibers apH 5; (c) deposition pH 6; and (d) deposition pH 7.

the zeta potential results demonstrated only slight increasesincreasing contact angle. The IEP was dependent upon bodeposition solution concentration and the pH.

Sufficient silane A1126 was deposited on the surface to athe nitrogen peaks at high and low concentration to be reso(ESCA). Two peaks were detected, hydrogen bonded, and pnated nitrogen at 399.8 and 401.7 eV, respectively; no free ngen was detected. Four times the amount of protonated nitrwas present in the low concentration samples, independepH. At high concentration the ratio of protonated nitrogenhydrogen bonded nitrogen was 2:1 (Table 8).

Silane A1126 appeared to show a negative correlabetween DRIFT and contact angle. An increased contac

TABLE 8Ratios of Hydrogen-Bonded and Protonated Nitrogen for Silane

A1126 (Measured by XPS)

A1126 % pH Hydrogen-bonded N% Protonated N

0.005 4 16.7 83.30.005 7 16.9 83.10.15 4 30.4 696

0.15 7 33.3 66.7

four concentrations: 0.005, 0.01, 0.1, and 0.15wt%; (a) deposition pH 4; (b)ion

withthe

lowvedoto-tro-gent ofto

on,an-

gle with a decreased intensity of the normalized CH2 peak at2925 cm−1, was noted (Fig. 6). At first these appeared tocontradictory, if one also includes the zeta potential versus ctact angle results the implication is that as deposition solut

FIG. 6. Correlation between contact angle and quantity of silane A1126deposited as measured by DRIFT.

N

d

ane5%,f 7bynal

io-tric

142 WATSO

concentration increased the surface became more basic. Ththe surface packing density increased so the CH2 was shieldedby the amine, resulting in a reduced CH2 signal.

Siloxane AR433

At 0.005% silane concentration the zeta potential andremained constant as pH was increased, up to and inclupH 10.5 (Figs. 7a–7e). At all pHs the IEP, of the 0.005% a

the 0.01% surface treated E-glass fibers, was at an electrolytepoint was assigned to the amino function and the mildly acidic pH of between 4 and 6, indicating that the dominant surface wasone to the aromatic functional groups.

FIG. 7. Zeta potential traces for siloxane AR 433-treated E-glass fibedeposition pH 5; (c) deposition pH 6; and (d) deposition pH 7; (e) depositio

ET AL.

us as

ieping

nd

ionizable metal oxides, with a few sparsely deposited siloxpatches (Figs. 7a and 7b). At concentrations of 0.1 and 0.1the IEP became more basic, rising from an electrolyte pH oto an electrolyte pH of 8.5, caused by removal of acid sitesadsorption onto silanols and by the basic nature of the functiogroups of the siloxane.

Titration of the siloxane gave two poorly defined stoichmetric points at pH 8.0 and pH 6.2. The basic stoichiome

rs at four concentrations: 0.005, 0.01, 0.1, and 0.15 wt%; (a) deposition pH 4; (b)n pH 10.5.

F

e

a

e

).ecaeh

tity

asdon-ore

ox-at

rednsis-

ofanollittle

AQUEOUS AMINO SILANE MODI

Low concentrations of this material had the greatest effincreasing the contact angle some 30◦ more than did the twosilane monomers. As a consequence the contact angle wmost twice that of the untreated E-glass fibers. The increascontact angle was linear with solution concentration betwpH 4 and pH 7, but independent of pH. The surface was vhydrophobic, or conversely, oleophilic.

TGA analysis of fibers treated with a 0.15% deposition sotion at pH 4 and at pH 10.5, indicated that at pH 10.5, 10 timmore silane was deposited than at pH 4 (results not shown

Between deposition solution pH 4 and pH 7, again differsurfaces were demonstrated, which were in the main due tocentration effects. A more negative zeta potential and increcontact angle were the result of larger amounts of silane bdeposited from the higher concentration solutions, while pHa minimal effect (Figs. 8a and 8b).

FIG. 8. Effect of increasing deposition solution pH and concentration

the relationship between contact angle (◦) and zeta potential at the base plateapH 10.1 (a) and at the acid plateau pH 3.5 (b) for siloxane AR433.

ICATION OF E-GLASS SURFACES 143

ct,

s al-e inen

ery

lu-es

nton-sedingad

on

FIG. 9. (a) Correlation of contact angle (◦) with quantity of siloxaneAR 433 deposited, as measured by DRIFT. (b) Correlation of IEP with quanof siloxane AR 433 deposited, as measured by DRIFT.

The plot of the normalized CH2 peak at 2925 cm−1 versusIEP indicated a gradual rise in IEP with surface coveragemonitored by CH2 intensity. A large jump between the 0.01 an0.1% silane solution treated fibers at pH 10.5, further demstrating the large quantities deposited at this pH from the mconcentrated deposition solutions (Fig. 9).

Aluminum and calcium were not detected in the 0.15% silane AR433 sample, which was deposited from a solutionpH 10.5.

DISCUSSION

Silane A1160

By nature A1160 was a cationic silane; however, it appeato behave as a nonionic silane. The results noted were cotent with a Si–O–Si surface. Even at pH 6–7 (the natural pHthe silane), where one would expect to see deposition of silmonomers and hence a silanol or ureido surface, there was

udifference in the zeta potentials. This implied that, at all de-position pHs the ureido silane deposited onto E-glass fiber as

N

et

i

i

n

tv

a

r

w

astdo

6

mt

a

na

T

wasitionobic) thesili-.005tra-rtionp,”

singtedilaneorcedThed

ine

tivede-

wasilanol

ightherof themer,

es,aneined.e,

thusandt theittheto bepH.%,ited,closeO–Siec-singposi-very

144 WATSO

aggregates with a Si–O–Si outer shell. The results presentePremachandraet al. indicated that condensates were unlikto have formed at pH 4 in a ureido silane deposition soluand thus any condensation must have occurred with and aE-glass surface.

ESCA results indicated that as deposition solution pH wreduced to pH 4, the hydrogen bonds could not be mainta(Table 8). Decreasing the deposition pH below pH 4 mayterfere with any hydrogen-bonded structure formed; allowthe silane to deposit with the ureido function outermost. Tamount of protonated nitrogen could be expected to increfurther in this scenario. Under these conditions the opportufor formation of a ureido surface would be maximized. At lowpH the “micelle like” structures may be destroyed allowingformation of a ureido surface. It should also be noted, howethat extremely low pH solutions have a deleterious effect onphysical nature of E-glass fibers (18, 23).

The zeta potentials of the silane A1160-treated E-glass fibwere independent of treatment solution pH and concentra(Fig. 2). Although silane A1160 is a cationic silane, on E-glfibers it behaved as a nonionic silane (zeta potential). Whenposited onto metal surfaces from aqueous solutions, the usilane had the ability to behave as a cationic silane as demstrated by van Ooij and by Puomi (24, 25). The conclusionthat the combination of E-glass fiber and this silane inducesouter layers to deposit “upside down,” with a highly crosslinkSi–O–Si outer surface. Aluminium from the E-glass fiber surfwas detected (XPS) in the upper layers of this silane depofrom more concentrated organic solutions. We propose thaaluminum induced the formation of very stable hydrogen bon“micelle-like” complexes with the ureido silane, which did nbreak down over the pH range of the deposition solutions, leing to a bilayer structure, with a siloxane outer surface andaluminum condensed into the siloxane network (26, 27).

Silane A1126

The surface basicity increased with deposition solution ccentration and with pH, this was confirmed by DRIFT (Fig.the amount of protonated nitrogen (ESCA, Table 8) and ctact angle results (Fig. 3b and Table 5). The increase in baity suggested that, at lower deposition pHs, some silanehave deposited with the silanols outermost. Additionally neu(pH 7) deposition conditions were more favorable for this basilane; at neutral deposition pHs the conditions were favourfor silane condensation (10)

At neutral deposition pH the surface basicity approachedstoichiometric balance point of the amine, indicating a totaamine surface. This was confirmed by a plot of the contact aversus zeta potential at pH 5.8, indicating that the surface chwas entirely due to the diamino functionality (Fig. 10a). Abo0.1% solution concentration and pH 6, the surface did notcome any more basic with enhanced deposition (Fig. 10b).

may have been a result of increased condensation in soluof the more concentrated systems as the pH increased to

ET AL.

d bylyiont the

asnedin-ngheaseity

erheer,

the

ers,tionssde-

eidoon-asthe

edceitedtheedtad-the

on-),

on-sic-ay

ralsicble

thellyglerge

vebe-his

indicated by Osterholtz and Pohl (10). Enhanced basicityproduced by a higher density of amino groups, as the depossolution concentration and pH increased a more hydrophsurface was produced. At the acid plateau (zeta potentialeffect of the acidic surface silanols and aluminum from thecate could be seen at the lower deposition concentrations (0and 0.01%) (Fig. 10c). At higher deposition solution concentions the surface was due to NH+3 , resulting in a much highezeta potential (Fig. 10c). Condensation of the silanols in soluat pH 7 formed oligomers, which then deposited “right way uresulting in an amino surface.

The zeta potential and DRIFT effects can be explained ua combination of Ishida’s and our own results. Ishida indicathat as silane concentration and hence packing of the spatches increased, the deposited silane molecules were fto change from the prone position to an upright position.reduction in the intensity of the CH2 peak can be explaineby increased shielding of the hydrocarbon chains by the amfunctionality as the packing density increased.

Siloxane AR433

The amino functionality resulted in an enhanced negacharge at high electrolyte solution pHs, especially whenposited from highly basic solutions (Fig. 7e).

At low concentrations and pH 10.5 the contact anglereduced almost to that of untreated E-glass fibers. Surface sgroups, which were produced by hydrolysis of the (SiO)m–CH3

and (SiO)n–CH3 groups, may have caused this effect. At hconcentrations the contact angle was higher than all the osurface treatments, possibly because of complete coveragesilane patches on the E-glass surface by the siloxane polyresulting in an extremely hydrophobic surface.

We propose that the siloxane initially deposited in patchwhich grew vertically, and that when large amounts of silwas deposited the upper layers of these patches were joResulting in E-glass fibers with an outer “skin” of siloxanbelow which was an open porous structure.

Siloxane AR433 was already hydrolyzed and condenseddeposition pH had no effect upon the state of hydrolysiscondensation of the deposited material, but was able to affecdynamic Si–O–Si↔ 2Si–OH. Different surfaces were implicin the plots of zeta potential versus contact angle, at bothacid and the base plateau (Fig. 8a). The surfaces appeareddependent upon both deposition solution concentration and

At low deposition solution concentration, 0.005 and 0.01and pH 10.5 very small amounts of the siloxane were deposas demonstrated by a low contact angle and zeta potentialto that of untreated fibers. Here the surface was due to Si–from the E-glass and from the polysiloxane (Fig. 8b). At eltrolyte pH 3.5 a pH-dependant correlation between increazeta potential and contact angle was observed. For the detion solution at pH 10.5 as the concentration was increased,

tion,7, aslarge amounts of silane were deposited, demonstrated by highcontact angle and zeta potential (Figs. 8a and 8b).

e

ion of

AQUEOUS AMINO SILANE MODIFICATION OF E-GLASS SURFACES 145

FIG. 10. (a) Effect of increasing deposition solution pH and concentration on the relationship between contact angle (◦) and zeta potential at the IEP for silanA1126. The area between the lines represents deposition of monomer and the hatched area deposition of oligomers. (b) Effect of increasing depositionsolution pHand concentration on the relationship between contact angle (◦) and zeta potential at electrolyte pH 5.8 for silane A1126. The solid line represents depositmonomer and the dotted line deposition of oligomers. (c) Effect of increasing deposition solution pH and concentration on the relationship between contact angle

(◦) and zeta potential at the acid plateau (pH 3.5). The solid line representing a surface of silicate and mixed orientation deposition and the hatched area an amino

ro

o

lnftd

ee

dp

thekedhy-olsE-in-

ragere-

in are-er,

surface consisting of oligomers and polymer for silane A1126.

We have seen little evidence in previous work of our own, oothers, which confirms complete surface coverage. We proptherefore that the silane was chemisorbed in a patchy fashiis generally accepted.

Aluminum and calcium have been detected in the upperers of all the condensed silanes, irrespective of the amousilane deposited, which has been interpreted as evidence oinclusion in the siloxane network. Aluminum was not detecin the upper layers of the sample of 0.15% AR433 depositepH 10.5%; being previously condensed it would be unlikelyhave aluminum and calcium in the upper layers.

The relationship between the deposition solution conctration and pH was clearer at pH 10.5, where an increasthe degree of hydrophobicity and amount deposited coulseen. At acid and neutral deposition solution pHs a gradual

independent increase with concentration was noted (Figs.and 9b).

ofse;

n as

ay-t oftheired

atto

n-in

beH-

CONCLUSIONS

The zeta potential of A1160 was independent of pH overrange measured. This was as a result of the highly crosslinSi–O–Si outer layer, at higher concentrations resulting in adrophobic surface. At low concentrations hydrophilic silanwere dominant. Unlike nonionic and cationic silane-treatedglass fibers there was little or no surface dissociation. Thecrease in hydrophobicity was solely due to increased coveof the E-glass surface and not to solution pH, unlike thesults presented by Nishyamaet al. (7–9). According to Park andSubramanian a highly crosslinked silane layer would resultreduction of the physical properties of a composite, due toduced interpenetration of the siloxane and the matrix polymensuing in a weak interphase (6).

9aCorrelation of the zeta potential, contact angle, and DRIFT re-sults for concentrated solutions of silane A1126 (0.1 and 0.15%),

N

u

tst

inap

iolsu

ic

e

k

.,

uct

ok,az,

-ess,

146 WATSO

indicated that the diamino silane was deposited “right wayand moreover that the outer surface or skin of the samplesdependant upon both deposition solution pH and concentraAt low concentrations (0.005 and 0.01%), isolated patchesilane increased the hydrophobicity of the surface. A negacorrelation between surface coverage, as measured by thesity of the CH2 peak (DRIFT), was noted. The correlation wthought to be due to shielding of the terminal amino grougiving the appearance of reduced deposition.

Silane AR433 likewise demonstrated three concentratdependant surfaces, one due mainly to silanols from the E-gwith a low concentration of deposited siloxane. A secondface indicated a surface consisting of siloxane functional groon the polymer. The third surface was apparent only at vhigh deposition pH (10.5), which was due to an outer “skof siloxane. We propose that at a deposition solution contration of 0.15% and pH 10.5 the siloxane initially depositedpatches and that further siloxane deposition was as an ovwhich joined the patches, giving the appearance of fully coaE-glass fibers. The structure below the outer surface or “swas open and porous.

ACKNOWLEDGMENTS

The authors thank TEKES and the Graduate School of Materialssearch (GSMR) for financial support for J. Rosenholm, A

a. Torkulla, and H.

Watson and Raisio Chemicals Ltd for A. Norstr¨om. Additionally we thankDr. H. M. Fagerholm for XPS measurements and useful discussions.

REFERENCES

1. Boudewijn, J. R., Schottens, J. R., and Brackman, J. C.,J. Adhesion Sci.Technol.52,115 (1995).

2. Dwight, D. W., Fowkes, F. M., Cole, D. A., Kulp, M. J., Sabat, P. JSalvati, L., Jr., and Huang, T. C.,J. Adhesion Sci. Technol. 4, 619 (1990).

ET AL.

p”wasion.of

iveten-ss,

n-ass,ur-ps

eryn”en-inrlaytedin”

Re-

3. Pluedemann, E., “Silane Coupling Agents”,2nd ed. chap. 4. Plenum Press,New York, 1991.

4. Daniels, M. W., and Francis, L. F.,J. Colloid Int. Sci.205,191 (1998).5. Tesoro, G., and Wu, Y.,J. Adhes. Sci. Technol.5(10), 791 (1991).6. Park, J. M., and Subramanian, R. V.,J. Adhes. Sci. Technol.5(6), 459 (1991).7. Nishiyama, N., Asakura, T., and Horie, K.,J. Colloid Int. Sci.124(1),

14(1988).8. Nishyama, N., Schick, R., and Ishida, H.,J. Colloid Int. Sci.143,146 (1991).9. Nishyama, N., Horrie, K., and Asakura, T.,J. Colloid Int. Sci.129, 113

(1989).10. Osterholz, F. D., and Pohl, E. R.,J. Adhesion Sci. Technol.6, 127 (1992).11. Premachandra, J. K., van Ooij, W. J., and Mark, J. E.,J. Adhesion Sci.

Technol.12(12), 1361 (1998).12. Leyden, D. E., and Atwater, J. B.,J. Adhesion Sci. Technol.5(10), 815

(1991).13. Plueddemann, E. P.,J. Paint Technol.40,1 (1986).14. Pape, P. J., and Plueddemann, E. P.,J. Adhesion Sci. Technol.10, 5

(1991).15. Watson, H., Norstr¨om, A., Root, A., Matisons, J., and Rosenholm, J

J. Adh. Sci. Technol., in press.16. Arkles, B., “Silanes, Silicones and Metal Organics”. Gelest 2000 prod

information.17. Silquest Silanes, Products and Applications Witco (1997).18. Watson, H., Norstr¨om, A., Engstr¨om, B., and Rosenholm, J.,Colloids and

Surfaces A, in press.19. Jacobasch, H. J., Baub¨ock, G., and Shurtz, J.,Colloid Polym. Sci.263,3

(1985).20. EKS 100. Control and evaluation software for EKA Instruction Handbo

Software version 4.25 Document A481B05-A Anton Paar GmbH GrAustria, 1997.

21. Sauer, B. B., and Carney, T. E.,Langmuir6, 1002 (1990).22. Bascom, W. D.,in “Modern Approaches to Wettability: Theory and Ap

plications” (Schrader, M. E., and Loeb, G. I., eds.), P. 329. Plenum PrNew York, 1992.

23. Elmer, T. H.,J. Am. Ceram. Soc.67,778 (1984).24. Yuan, W. E., and van Ooij, W. J.,J. Colloid Interface Sci.185,197 (1997).25. Puomi, P., and Fagerholm, H.,J. Adhes. Sci. Technol.,in press.26. Watson, H., Kaunisto, M., Gustafsson, J., and Paivarinta, J.,J. Colloid

., Interface Sci., in press.27. Wang, D., and Jones, F. R.,Surf. Interface Anal.20,457 (1993).