J Sol-Gel Sci Technol DOI 10.1007/s10971-015-3620-9

ORIGINAL PAPER

Hybrid sol–gel coating agents based on zirconium(IV) propoxide and epoxysilane Ingrid Milosˇev • Barbara Kapun • Peter Rodicˇ Jernej Iskra

•

Received: 31 October 2014 / Accepted: 7 January 2015 Ó Springer Science+Business Media New York 2015

Abstract The hybrid sol–gel was synthesized as a basis for preparing coatings for use as anticorrosion and antibacterial protectants of medical implants. Coatings were prepared using (3-glycidyloxypropyl)trimethoxysilane and zirconium(IV) propoxide as precursors, with acetic acid as a catalyst. The synthesis was followed using in situ Fourier transform infrared spectroscopy. Specific steps of the synthesis were followed by 1H nuclear magnetic resonance spectroscopy. The reactions of zirconium(IV) propoxide and (3-glycidyloxypropyl)trimethoxysilane with acetic acid differed; acetate was formed only in the reaction between zirconium(IV) propoxide and acetic acid. The reaction of the opening of the epoxy ring of GPTMS was followed, during both synthesis and curing, at room temperature and at 150 °C. The opening of epoxy ring was catalysed thermally. The thermal stability of the prepared sol was determined by thermogravimetric analysis.

Graphical Abstract

Ingrid Milosˇev and Barbara Kapun shares equally the first authorship. I. Milosˇev (&)  B. Kapun  P. Rodicˇ  J. Iskra Department of Physical and Organic Chemistry, Jozˇef Stefan Institute, Jamova c. 39, 1000 Ljubljana, Slovenia e-mail: [email protected] J. Iskra Centre of Excellence for Integrated Approaches in Chemistry and Biology of Proteins, CiPKeBIP, Jamova c. 39, 1000 Ljubljana, Slovenia

Keywords Hybrid sol–gel coating  Zirconium(IV) propoxide  (3-glycidyloxypropyl)trimethoxysilane (GPTMS)  FTIR  NMR

123

J Sol-Gel Sci Technol

1 Introduction Hybrid coatings for metal surfaces with crosslinking between inorganic and organic phases can be obtained by the use of functionalized silanes as cross linkers. One such widely used organically functionalized silane is (3-glycidyloxypropyl)trimethoxysilane (GPTMS; sometimes denoted as GPTS and GLYMO). It contains an epoxide functional group. Due to their chemical and thermal resistance and suitable mechanical properties, epoxy/silica hybrids have numerous applications, e.g. as scratch resistant, corrosion resistant, anti-fogging and hydrophobic coatings [1]. GPTMS can undergo several reactions such as hydrolysis of the methoxy group to give silanol groups, which can then be condensed to a 3D-branched siloxanesilica network, and opening and polymerization of its epoxy ring to form a linear poly(ethylene oxide) organic network. Opening the epoxide ring in GPTMS leads to the formation of diol, alkoxy alcohol and polyether products [2]. Non-catalyzed opening of the epoxy ring proceeds at elevated temperatures (thermal curing), while at room temperature the epoxy ring remains intact [3]. In order for the reaction to proceed at low temperatures, crosslinking agents such as amines [3–5] or boron trifluoride diethyl etherate (BF3OEt2) have to be added [1]. Epoxy-based inorganic–organic hybrid polymers prepared from GPTMS and DETA (diethylenetriamine) have been used [3]. The process of ring opening at elevated temperatures was dependent on the concentration of the added amine. In the case of a low content of DETA, compared to that of GPTMS, ring opening took place during the thermal treatment at 150 °C but, at high amine content, the reaction was possible even at room temperature [3]. The effect of water content was found to be crucial for the structure of the hybrid films formed from GPTMS and tetraethylorthosilicate (TEOS) [2]. At low organic content (i.e. low amount of GPTMS), low water content was sufficient for both hydrolysis and ring opening. At low water/ high organic content, water was consumed preferentially in the hydrolysis reaction and not in the ring opening reaction. The high organic content required higher water content in order to accomplish both hydrolysis and ring opening. Hydrolysis was more dependent on the pH than the condensation reaction [6] and at lower pH values was much faster than condensation [7]. The hydrolysis rate increases sharply as the molar ratio of water to alkoxide is increased [7–9]. The organic component of hybrid sol–gel coatings allows the use of low curing temperatures, while retaining flexibility and density. The introduction of inorganic components like metal alkoxides contributes to the mechanical properties of the hybrid coating, such as hardness and adhesion to the substrate [10–12]. Recently

123

we described new hybrid sol–gel coatings, based on zirconium(IV) tetrapropoxide (ZTP), TEOS, MAPTMS (3(trimethoxysilyl)propyl methacrylate) and methacrylic acid, that provide good corrosion protection of aluminium [13, 14]. Since metal alkoxides are often highly reactive it is necessary to use chelating agents (e.g. methacrylic acid, acetic acid, 2,4-pentanedione) to lower their reactivity [1]. In the case of organically functionalized silanes containing an epoxy group like GPTMS, it is important that the metalbased sols show relatively high activity for epoxide ring opening [15, 16]. Zirconium promotes thermal polymerization of epoxides to polyethylene oxide chains, while titanium promotes photochemical polymerization [16]. Zirconium alkoxide, as a Lewis acid, is able to catalyse the epoxy ring opening by amines [4, 5, 15, 17]. In addition to water content, temperature and crosslinking agent, the ratio of inorganic to organic components is also important. The addition of Zr promotes the formation of a more amorphous Si–O–Si network instead of the ring (SiO)4 unit typical of trialkoxysilanes like GPTMS [4, 18]. Zr also acts as a toughening agent and enhances fracture of the hybrid coating [4]. This is partially ascribed to the presence of the epoxide ring, which not only increases cross-linking within the network but also improves adhesion to the substrate, since the opened epoxy ring can form Si–O–C bonds and hydrogen bonds with the substrate. At high ratios of GPTMS to ZTP, i.e. increased organic content, the epoxy groups fill the pores within the Si/Zr network and reduce shrinkage of the coating and the tendency to crack formation [11]. Acetic acid [19] and diglycol [7] have been used as catalysts and chelating agents in the formation of pure ZTP sols. It was proposed that acetic acid substitutes for the alkoxy groups of Zr alkoxide and probably acts as coordinating bridging ligand [19]. Diluted nitric acid and ethyl acetoacetate [5], and acetic acid [4, 11, 12, 20] have been used in the synthesis of GPTMS/ZTP sols. The mechanism of reactions involved in the formation of GPTMS/ZTP sols has not been investigated in detail. We have therefore studied the synthesis of ZTP as a completely hydrolysable inorganic precursor, and of GPTMS as a partially hydrolysable, organically modified silane precursor. Acetic acid was used as catalyst and chelating agent. We used the reported basic synthetic route [12] and followed the synthesis using in situ Fourier transform infrared spectroscopy (FTIR) and nuclear magnetic spectroscopy (NMR) in order to determine the mechanisms of the chemical reactions involved. Special emphasis was given to the epoxy ring opening reaction. No cross-linking additives were added and curing was performed at room temperature or at 150 °C. The final goal was to develop the hybrid sol–gel for use as a protective coating in biomedical applications. The components used

J Sol-Gel Sci Technol

have been chosen to be suitable for application in human body, since ZTP, GTPMS and acetic acid are innocuous.

vigorous stirring at 300 rpm for 15 min to induce gelation, giving a hybrid sol–gel denoted as ZG. 2.1.3 Material

2 Experimental 2.1 Preparation of coatings 2.1.1 Chemicals for sol–gel The sol precursors used were zirconium(IV) propoxide, [ZTP: Zr(OPr)4, 70 wt%, in 1-propanol, Aldrich], (3glycidyloxypropyl)trimethoxy silane (GPTMS: C9H20O5Si, 98.0 %, Aldrich) and acetic acid (AcOH: C2H4O2, 100 %, AppliChem).

Titanium was supplied by Goodfellow, England, in the form of 2.0 mm thick foil. Samples were cut from the foil in the shape of 15 mm diameter discs. Before depositing, the substrates were ground mechanically under a stream of water with, successively, 320-, 500-, 800-, 1,000-, 1,200-, P2,400- and P,4000-grit SiC paper (supplied by Struers, Ballerup, Denmark). The samples were rinsed with distilled water, cleaned ultrasonically in ethanol for 2 min and dried in a stream of nitrogen. 2.1.4 The deposition of coatings

2.1.2 Synthesis of sols Inorganic–organic hybrid sol–gels were synthesized in two steps by mixing two, separately prepared, solutions, Sol 1 and Sol 2 (Fig. 1). All reactions were performed in a mini reactor of 25 mL volume at atmospheric pressure in the EasyMax 102 controller. The temperature of the reactor jacket, Tj, was kept constant at 25 °C ± 0.01. During synthesis, each solution was stirred using a propeller stirrer at a speed of 100 rpm. Sol 1 was prepared by mixing ZTP and acetic acid in the ratio 1:4. Water was then added in a 1:1 molar ratio (Zr:H2O = 1:5). Reaction was continued until no further change was seen by FTIR spectra, i.e. the sol was fully stabilized. Sol 2 was made by mixing GPTMS with acetic acid in a 1:3 molar ratio. Water was then added immediately to a molar ratio of GPTMS to water of 1:3. Each sol was stirred for 15 min and then mixed under

Sol ZG was deposited onto the substrate surface by a spincoater (Laurell—WS-650-23NPP/LITE/IND). The substrate was then rotated at a speed of 4,000 ± 2 rpm for 30 s. During rotation, excess applied sol was removed, leaving the substrate covered evenly with a homogeneous coating. After deposition, samples were cured thermally on a hot-plate at 150 °C for 1 h. 2.2 Methods for characterizing the sol–gel process and coatings 2.2.1 Fourier transform infrared spectroscopy Chemical changes during sol–gel synthesis and optimization of the process were analysed by in situ FTIR in the range from 2,800 to 600 cm-1. The spectra were recorded

Fig. 1 Experimental procedure for preparing sol ZG from two separately prepared sols

123

J Sol-Gel Sci Technol

by a ReactIRTM 45 instrument using a universal attenuated total reflectance (ATR) sampling accessory with a resolution of 8 cm-1, averaging 128 scans. An EasyMax 102 controller was used to control reaction conditions during the reaction. The instruments were controlled by iControl EasyMax 4.2 and iC IR 4.2 software. Intensities of spectra are shown as log absorbance in absorbance units. FTIR spectra and temperature were measured in situ every 1 min during the course of synthesis of each sol, as well as after their mixing to obtain a hybrid sol–gel. FTIR spectra of sols and deposited coating were recorded, with a resolution of 4 cm-1, by a Perkin Elmer Spectrum 100 instrument, using the ATR sampling accessory, averaging four scans in the range from 4,000 to 600 cm-1. The sol was analysed by depositing a drop of sol on the topplate of the instrument. The coating was analysed by mounting the coated substrate kinematically at the top-plate. Spectra are normalized and presented as absorbance. 2.2.2 Nuclear magnetic resonance 1

H NMR spectra were recorded in deuterated chloroform (CDCl3) solvent using a Varian Inova 300 MHz spectrometer operating at 300 MHz. Chemical shifts (d) are reported in ppm units with CDCl3 as internal standard. Values of chemical shifts were referenced externally to TMS (d = 0). Spectra were recorded at different steps of preparation for all sols.

3.1.1 Sol 1: mixing ZTP and AcOH ZTP in 1-propanol shows several characteristic absorbance bands in the range between 860 and 1,460 cm-1, the most intense being at 1,156, 1,130 and 1,072 cm-1 (Fig. 2a; Table 1). The region between 1,300 and 800 cm-1 is the most important region for studying the hydrolysis of ZTP. The bands at wavenumbers lower than 1,100 cm-1 are related mainly to the 1-propanol solvent. The bands at 1,460 and 1,381 cm-1 are related to the propyl group of ZTP and 1-PrOH. The band at 1,130 cm-1 was attributed to a combination of Zr–O–C and skeletal stretches [8, 21]. Acetic acid shows characteristic bands at wavenumbers higher than 1,200 cm-1, the most intense being at 1,714, 1,423 and 1,293 cm-1 (Table 1). The formation of Sol 1, including successive additions of AcOH and water to ZTP and the reactions of hydrolysis and condensation, were followed using in situ ATR-FTIR spectroscopy (Fig. 2). Acetic acid was used as the catalyst and chelating agent for ZTP. The addition of AcOH to ZTP resulted in decreased intensity of characteristic bands of ZTP at 1,156 and 1,130 cm-1, and appearance of new bands at 1,600, 1,450 and 700 cm-1. The band of the

2.2.3 Thermogravimetric analysis The thermal stability of the hybrid coatings was determined by thermogravimetric analysis (TGA) and differential thermal analysis (DTA). TGA measurements were performed using an STA 449 F3 NETZSCH instrument, Selb, Germany. The measurements were carried out from room temperature to 450 °C at a heating rate of 10 K/min. During measurements the instrument was purged with argon (50 mL/min). The sample was placed in a small electrically heated oven with a thermocouple (Al2O3). The oven was kept under argon inert atmosphere (30 mL/min). A DEFINE quadrupole mass spectrometer (QMS) was used to analyse the solvent residues.

3 Results and discussion 3.1 Optimization of the synthesis Two sols, ZTP/acetic acid and GPTMS/acetic acid, were formulated using the acetic acid/water based sol–gel process. The synthesis was optimized using in situ FTIR spectroscopy.

123

Fig. 2 a FTIR spectra of the ZTP precursor and of its solution following the addition of acetic acid and water (Sol 1). b In situ 3-D FTIR spectra of formation of Sol 1 as a function of time. The points at which AcOH and water were added are denoted by arrows

J Sol-Gel Sci Technol Table 1 Selected bands in the FTIR spectra of reference compounds and synthesized sols

The most intense bands are in bold type

Frequency (cm-1)

Signal ZTP in 1-propanol

860, 888, 909, 969, 1,003, 1,072, 1,130, 1,156, 1,381, 1,460

1-propanol

860, 888, 905, 968, 1,017, 1,055, 1,069, 1,099, 1,236, 1,384, 1,458

H2O

1,638

CH3COOH (AcOH)

1,047, 1,293, 1,423, 1,714, 1,756

GPTMS

780, 817, 910, 1,078, 1,192, 1,256

Zr(OAc)4

645, 1,449, 1,562 [21]

Si-OH

909, 944, 1,010

Si-O-Si

1,021, 1,090

ZTP-AcOH complex

696, 1,238, 1,457, 1,479, 1,553

1,479 and 1,457 cm-1 transformed into a band at 1,449 cm-1; that at 1,238 cm-1 was shifted to 1,266 cm-1. In the region of Zr–O–C bonds the signal at 1,130 cm-1 disappeared completely. The bands of AcOH, at above 1,700 cm-1, increased, indicating that free acetic acid is released in the reaction (Eq. 2a). It may also be assumed that PrOH is released due to substitution of OPr (Eq. 2b). Due to the appearance of the PrOH signal and the disappearance of that for PrO- in the first step, the most likely reaction is Eq. 2a with n = 4, while coordination of AcOH and PrOH to Zr is also possible. Fig. 3 In situ 3-D FTIR spectra recorded in the region between 1,500 and 1,660 cm-1 related to formation of acetate in Sol 1 after addition of acetic acid to ZTP

ZrðOAcÞn ðOPrÞ4n þ xH2 O !Zr(OAc)nx ðO PrÞ4n ðOHÞx þ xAcOH ð2aÞ

carbonyl group of AcOH (at 1,714 cm-1) appeared only a few minutes after the latter’s addition (Fig. 2b). Thus, nucleophilic substitution of propoxy by acetate groups (Eq. 1) takes place on the addition of AcOH to ZTP solution, in the first few minutes following mixing. ZrðO PrÞ4 þ 4AcOH ! ZrðOAcÞn ðO PrÞ4n þ n Pr OH ð1Þ Detailed analysis of the region between 1,500 and 1,660 cm-1, in the first 15 min following addition of AcOH (Fig. 3), shows that the reaction is multistep and that intermediate products are formed as indicated by bands 1,576, 1,566, 1,540, 1,640 cm-1, in order of their appearance. After stabilization, the product showed two bands at 1,596 and 1,543 cm-1 in that region, with further bands at 1,479, 1,457, 1,415 and 696 cm-1 (Fig. 2). These bands agree well with those reported as characteristic of Zr(OAc)4—at 1,562, 1,449 and 645 cm-1 [21]. In the region of C–O bonds only three bands at 1,054, 1,013 and 968 cm-1 remained that match those of PrOH (Fig. 2). After water is added (Fig. 2b), the band related to an intermediate product at 696 cm-1 disappeared and a new band appeared at lower wavenumbers. The other bands related to Zr(OAc)n(OPr)4-n at 1,543 and 1,596 cm-1 are transformed into one band centred at 1,553 cm-1, those at

ZrðOAcÞ4n ðOPrÞn þ xH2 O !Zr(OAc)4n ðO PrÞnx ðOHÞx þ xPrOH ð2bÞ For complete hydrolysis to ZrO2, 2 mol of water are necessary. When water alone is added to ZTP, a white suspension of ZrO2 is formed immediately. The presence of acetic acid prevents the formation of oxide by the formation of acetate. After partial hydrolysis (Eq. 2a, b), free acetic acid could catalyse condensation with the formation of a Zr–O–Zr linkage (Eq. 3). ZrðOAcÞn1 ðOPrÞ4n ðOH) + Zr(OAc)n ðO PrÞ4n ! ðPrO)4n ðAcOÞn1 ZrOZr(OAc)n1 ðO PrÞ4n þ AcOH ð3Þ To obtain further information about the processes during the formation of Sol 1, samples from each step of its preparation were analysed by NMR spectroscopy (Fig. 4; Table 2). It would be possible to follow the changes by 29Si NMR; however, 29Si is a low sensitivity nucleus and a broad background signal from the glass and quartz should be eliminated which makes the measurements time-consuming. 1H NMR was thus chosen for the analysis of the sol–gel processes. The solution of ZTP in DMSO-d6 was

123

J Sol-Gel Sci Technol Fig. 4 1H NMR spectra of the initial component ZTP, recorded in CDCl3 solution before and after the addition of acetic acid and water to give Sol 1

not stable and a white suspension was formed. ZTP was stable in CDCl3 solution, showing a broad signal for PrOH at 3.68 ppm and for propoxide at 3.91 ppm (Fig. 4). After adding AcOH to ZTP, only one signal remained for the Pr group, attributed to PrOH (3.62 ppm), although it was broadened. On the other hand, there were several signals in the region 1.8–2.1 ppm indicating that the methyl group of AcOH has several different forms. Another sample was taken after addition of water; however, solution of the Sol 1 in CDCl3 was heterogeneous which complicated the acquisition of NMR spectra. Nevertheless, the signal for the CH2 group of Pr became sharper, while that for the acyl

123

CH3 remained broadened. NMR analysis confirmed the initial observation by IR spectroscopy (Figs. 2, 3) that zirconium tetra acetate was preferentially formed in the first step and that it could be further coordinated by PrOH, as indicated by the broadened signal observed after addition of water. 3.1.2 Sol 2: mixing GPTMS and AcOH The most intense bands in the spectrum of GPTMS are at 1,192, 1,078, 910, 817 and 780 cm-1 (stretching of C–O, Si–O and Si–C bonds) (Fig. 5; Table 1). With the addition

J Sol-Gel Sci Technol Table 2 Chemical shifts of the 1H NMR spectra of the initial precursors Chemical

Chemical shifts—d (ppm)

70 wt% ZTP in 1-propanol (in CDCl3)

0.92 (t, 3H), 1.59 (m, 2H), 3.68 (br s, PrOH), 3.91 (br s, PrO-)

1-propanol (in CDCl3)

0.94 (t, 3H), 1.57 (m, 2H), 3.58 (t, 2H)

Acetic acid (in CDCl3) Acetic acid (in DMSOd6)

2.10 (s) 1.88 (s)

GTPMS (in CDCl3)

0.68 (m, 2H), 1.71 (m, 2H), 2.61 (dd, 1H), 2.79 (dd, 1H), 3.15 (m, 1H), 3.40 (dd, 1H), 3.47 (m, 2H), 3.57 (s, 9H), 3.79 (dd, 1H)

Methanol (in CDCl3)

3.43 (s)

Fig. 5 a FTIR spectra of GPTMS before and after the addition of acetic acid and and water. b 3-D FTIR spectra recorded as a function of time. The points at which AcOH and water were added are denoted by arrows

of AcOH the bands of GTPMS decreased in intensity, while bands related to AcOH appeared (1,756, 1,714, 1,413 and 1,293 cm-1). In contrast to Sol 1 (Fig. 2), bands related to the formation of acetate at 1,596 and 1,543 cm-1 were not observed, indicating that chemical reaction between GPTMS and AcOH does not occur under these conditions. The major difference in the IR spectra following addition of AcOH was observed in region of C–O

and Si–O bands between 900 and 1,100 cm-1, where new bands at 1,055, 1,010, 943 and 909 cm-1 appeared. Following the addition of water (Fig. 5a, b), significant changes in the IR spectra were observed. The most intense bands of GTPMS at 1,192, 1,078, 817 and 780 cm-1 decreased or disappeared. Those related to AcOH at 1,413 and 1,293 cm-1 were shifted to lower wavenumbers. The Si–O–Si band at 1,010 cm-1 increased further and shifted to 1,021 cm-1, due to networking within the sol. The band at 910 cm-1 (Si–O-, Si–OH) also increased. The presence of strong bands in the region of Si–O–Si, Si–O- and Si– OH bonds (1,100–850 cm-1) confirms that hydrolysis and condensation occurred following the addition of acetic acid and, even more so, following the addition of water, resulting in the formation of a siloxane network. Si–O–Si asymmetric stretching absorption bands in the 1,000–1,200 cm-1 region may be related to the different configurations—ring or network—of siloxane [4, 22–24]. Analogy can be made to silsesquioxanes (SSQs), organosilicon compounds with the empirical formula (RSiO1.5)n, where R can be hydrogen, methyl, phenyl, or higher weight organic groups. They may be composed of close cages (T6, T8), partially open cages (T7), ladder, or random network [24]. The configuration affects several properties, like refractive index and dielectric constant [24]. Asymmetric Si–O–Si stretch can combine into symmetric or asymmetric modes with respect to the inversion point through the centre of the (Si–O)4 ring. The presence of a mring-asym band around 1,150 cm-1 indicates a highly symmetrical structure with the (Si–O)4 ring subunits found in closed cages and partially open cage structures. The relative intensity of the mring-sym band around 1,050 cm-1 reflects the local symmetry around the Si–O-Si unit, the higher intensity indicating a less symmetric, random network structure. By analogy to SSQs in the present material, the band at 1,021 cm-1 can be ascribed to a random network, and that at 1,090 cm-1 to a more closed, cage-like configuration. The domination of the intensity of former band indicates that the hydrolysis and condensation of GPTMS leads to the formation of a less symmetric Si–O–Si random network configuration. GPTMS possesses an epoxy ring that is important for its activity. Opening the ring relieves the ring strain; the products usually being secondary alcohol, diols or polyethers. Acid can catalyse ring opening. In the spectrum of GPTMS, the bands at 1,256, 910 and 856 cm-1 are due to the epoxy ring (Fig. 5a). After addition of AcOH, these bands overlap with those of Si–O- and Si–OH in the lower wavenumber region, while the band at 1,256 cm-1 overlaps with those for acetate. Thus, based on FTIR spectra, it is difficult to state whether the reaction of hydrolysis proceeds through the splitting of the epoxy ring. Based on the significant changes in the Si–O- and Si–OH bands (910–940 cm-1), it is more

123

J Sol-Gel Sci Technol Fig. 6 1H NMR spectra of the initial component GPTMS before and after the addition of acetic acid and water giving Sol 2

likely that the hydrolysis proceeds mainly on the silicon atom and not on the epoxy ring. The initial reactions are presented in Eqs. 4a, b and 5a, b (where GP denotes 3-glycidoxypropyl). Each product can act as a reactant in a further condensation reaction i.e. Eq. 6. The following reactions are possible during the formation of Sol 2:

123

Hydrolysis of GPTMS at the Si centre: AcOH

ðMeOÞ3 SiGP þ nH2 O ! ðMeOÞ3n ðHOÞn SiGP þ nMeOH ðMeOÞ3 SiGP þ nAcOH ! ðMeOÞ3n ðAcOÞn SiGP þ nMeOH

ð4aÞ

ð4bÞ

J Sol-Gel Sci Technol

Hydrolysis of GPTMS at the epoxide ring: ð5aÞ

ð5bÞ

Condensation reaction ðMeOÞ3n ðHOÞn SiGP þ ðMeOÞ3 SiGP AcOH

! ðMeOÞ3n GPSiOSiGPðOMeÞ3n þ nMeOH

ð6Þ

In order to obtain more information about the reaction mechanism, especially concerning the competitive processes of hydrolysis at the Si atom and of opening of the epoxy ring, the process of formation of Sol 2 was followed by NMR spectroscopy (Fig. 6). The peaks at 2.61, 2.79 and 3.15 ppm related to the epoxy ring and the singlet at 3.57 ppm to the MeO groups in the spectrum of GPTMS (Table 2) are of special interest. Following the addition of AcOH to GPTMS, a sample was taken after 5 min of stirring and dissolved in CDCl3. The spectra show no significant changes, indicating that there is no reaction between GPTMS and AcOH (Eqs. 4b, 5b). The next sample was taken after the addition of water and the spectra show that hydrolysis take place only at the silane centre (Eq. 4a), as indicated by the presence of a peak related to methanol at 3.45 ppm in place of that for the SiOMe group at 3.57 ppm. On the other hand, the epoxy signals remained unchanged (Fig. 6). The release of methanol could not be confirmed by the FTIR spectra (Fig. 5), since the main MeOH band at 1,030 cm-1 overlaps with those of the Si– O–Si bands. It appears that, following the addition of water, most, if not all, the methoxy groups bonded to silicon are hydrolysed, resulting in the formation of siloxane groups (Si–OH) and release of methanol. In a separate experiment, GPTMS was mixed with water and analyzed by NMR spectroscopy. Hydrolysis of methoxy groups was found to occur only to a small extent, pointing to the conclusion that acetic acid catalyses the hydrolysis process. It became evident that, during the preparation of Sol 2, an epoxy ring opening reaction can be excluded (i.e. Eq. 5a, b), since the relevant peaks remained unchanged. 3.1.3 Combination of Sol 1 and Sol 2 forming ZG sol Bands related to acetic acid are present in the FTIR spectra for Sol 1 and Sol 2 and for the combined ZG sol (Fig. 7). Sol

1 is dominated by a band at 888 cm-1, bands at 1,070 and 1,098 cm-1 (could be attributed to Zr–O/Zr–O–Zr; in spectra they partially overlap with ethanol bands) and by band at 1,554 cm-1 (could be attributed to Zr acetate). Sol 2 is characterized by Si–O bands at 909 cm-1, Si–OH at 1,021 cm-1 and Si–O–Si at 1,090 cm-1. Sol ZG (Fig. 7) gives rise to spectra that, in general, resemble a combination of both sols, but with some important differences. The band at 1,553 cm-1 related to Zr-tetraacetate decreased, while the band of AcOH at wavenumbers higher than 1,700 cm-1

Fig. 7 a FTIR spectra for Sol 1, Sol 2 and the final sol ZG. b 3-D FTIR spectra recorded as a function of time; the points at which Sol 2 was added to Sol 1, and the point at which stable ZG sol was formed are denoted by arrows

123

J Sol-Gel Sci Technol Fig. 8 1H NMR spectra of Sol 1, Sol 2 and final Sol ZG

increased, indicating formation of free AcOH from acetate bonded within a Zr network. In the region between 800 and 940 cm-1 of the curve for ZG sol, bands originating from Si–O and Zr–O overlap. Overlapping of bands is observed also in the region between 950 and 1,150 cm-1 for Zr–O– Zr, Si–O–Si and Zr–O–Si bands [13, 14]. The separation between the asymmetric and symmetric stretches of the carboxyl group COO- (D) in acetic acidcoordinated compounds has been ascribed to the type of ligand formed [25]. A separation larger than that observed for ionic species (e.g. D = 162 cm-1 for sodium acetate) could indicate monodentate ligands while smaller

123

separations are considered characteristic of bidentate ligands (chelating or bridging) [25]. Separations significantly smaller than that for the ionic species (i.e. D = 105 cm-1) have been proposed to be related mainly to bidentate chelating coordination, but also to the presence of some bridging acetate [25]. In the present study Sol 1 exhibits an asymmetric COO- band at 1,553 cm-1 and a symmetric COO- band with a maximum at 1,479 cm-1 on the high-energy side of the peak (Fig. 2). The separation between the bands of 75 cm-1 is therefore indicative of the acetate acting as a bidentate, mainly chelating ligand. In the final sol ZG, the symmetric COO- band is shifted to

J Sol-Gel Sci Technol

Fig. 9 Normalized FTIR spectra for the sol ZG and the coating deposited on Ti substrate, following curing for 1 h at 150 °C

1,449 cm-1, leading to a separation of 105 cm-1 (Fig. 5). Similar results, D = 102 cm-1 [12] and D = 113 cm-1 [21] have been reported for chelating of ZTP with acetic acid, and D = 104 cm-1 [26] for chelating of ZTO and hexahydrophthalic anhydride. Sol ZG was also analysed by NMR spectroscopy. Unlike in Sol 1 and Sol 2, no new signals appeared (Fig. 8). The main difference is in the shape of the signals. Peaks for MeOH, PrOH and AcOH are all sharp, while those related to the glycidoxypropyl group are broadened. This could indicate that a siloxane network is already forming, resulting in the glycidoxypropyl group being in a different environment. In contrast, the AcOH signal is sharper than in Sol 1, in agreement with the results of IR spectroscopy showing that AcOH is being released during formation of the ZG sol. Signals of the epoxy ring at 2.61, 2.79 and 3.15 ppm are also present in spectra of the ZG sol, showing that, under the conditions of formation of ZG sol epoxy ring is still present in the sol. 3.2 The reaction at the epoxy ring FTIR spectra of ZG sol, before and after its deposition on the surface of Ti substrate, were recorded in the range between 4,000 and 700 cm-1 (Fig. 9). The spectrum of the sol shows typical bands related to the hydroxyl group at 3,600 cm-1 (not shown), to acetic acid at 1,714 cm-1, to acetate between 1,554 and 1,260 cm-1, and to Zr- and Sirelated bands at lower wavenumbers (Fig. 7). After deposition in the form of a coating, followed by curing at 150 °C, the spectrum changed. The bands for hydroxyl and acetic acid are decreased, due to release of the latter during the curing process. The typical shape of the spectrum related to Zr- and Si- bands is distorted, leaving only three

Fig. 10 Normalized FTIR spectra for the coating deposited on Ti substrate a as a function of curing time up to 31 days at room temperature, and b after 3 and 31 days curing at room temperature, and after curing for 1 h at 150 °C. Vertical lines represent the positions of the peaks related to the epoxy group

bands (at 1,564, 1,455 and 1,003 cm-1) related to the sol. Due to the process of condensation and formation of a condensed network, the bands overlap. Analysis by NMR showed (Figs. 6, 8) that the epoxy group remained intact after synthesis of Sol 2 and ZG sol. The question remains as to what happened to the epoxy group after deposition of

123

J Sol-Gel Sci Technol

After 30 days the bands were no longer visible. In contrast, bands related to the inorganic network (between 1,000 and 1,100 cm-1 and 1,400 and 1,500 cm-1) are formed at low temperature immediately after deposition. Spectra recorded 3 and 31 days after deposition for coatings cured at room temperature and at 150 °C are compared (Fig. 10b). For low-temperature cured coating, bands related to acetic acid (around 1,423 cm-1) and those of Zr-acetate (at 696 and 1,553 cm-1) decreased after 31 days due to its slow evaporation. After curing at 150 °C the acetic acid evaporated faster (see below) and the bands related to epoxy group disappeared immediately after curing. The shape of the spectra did not change significantly with time following curing. 3.3 TGA-DTG analysis

Fig. 11 a Thermogravimetric (TG), differential thermogravimetric analysis (DTG) and differential thermal analysis (DTA) curves for the ZG sol, presenting the weight loss as a function of temperature under nitrogen flow. b Thermogravimetric (TG) (left y-axis) and quadrupole mass spectrometer (QMS) curves (right y-axis) showing the release of the main residues of the ZG sol with increasing temperature. The characteristic molecular ion or fragment peak in the mass spectrum is denoted in brackets

the coating and its curing at 150 °C. The FTIR data are ambiguous regarding the presence of an epoxy ring, since the related bands at 856, 910 and 1,256 cm-1 overlap with Si–O-related (858, 906 cm-1) and acetate-related bands (1,293 cm-1). It has been reported [5] that, unless the amine crosslinker is added, the epoxy groups are present in the coating cured at room temperature, although decreasing with time, indicating progressive cross-linking. In order to determine whether the curing at 150 °C affects the reaction of the epoxy group, the coatings were deposited and cured at room temperature. FTIR spectra were followed for up to 31 days and compared with those recorded for coatings cured at 150 °C (Fig. 10). The epoxy related bands were clearly present immediately after deposition (Fig. 10a). Up to 24 days after deposition they were still present, although at a smaller intensity, indicating the occurrence of polymerization of the organic network, at a very slow rate.

123

Changes in weight of the ZG sol were determined at various temperatures in a controlled nitrogen atmosphere by thermo-gravimetric analysis. TG and DTA show typical curves of the progress of drying of the sol–gel (Fig. 11a). Up to 100 °C, the weight loss is due to evaporation of 1-propanol, methanol, water and acetic acid, as shown by QMS analysis (Fig. 11b). Weight loss was at a minimum at 105 °C. At higher temperatures the coating is stable. At approximately 380 °C a small peak denotes the release of water incorporated in the sol structure. The coating is stable up 440 °C but decomposes above this temperature. The shape of the DTA curve confirms that the curing process, with vaporisation of the coating components that proceeds up to 90 °C, is an endothermic process.

4 Conclusions Zirconium(IV) propoxide and organically modified silane precursor GPTMS containing the epoxy group were synthesized by a sol–gel procedure, using acetic acid as catalyst. ZTP reacts chemically with acetic acid through nucleophilic substitution of propoxy by acetate, forming Zr-tetraacetate as an intermediate product that could be further coordinated by propanol. After addition of water, hydrolysis and, later, condensation occur, accompanied by release of excess of acid and propanol. The condensed product contains Zr–O–Zr bonds and incorporates acetate in its structure in a bidentate chelate coordination. In contrast to ZTP, GPTMS does not react with acetic acid to form acetate. Its hydrolysis leads to the formation of Si–O– Si bonds. The position of the spectral bands leads to the conclusion that a less symmetric Si–O–Si random network configuration is formed, and not a more closed, cage-like configuration. Hydrolysis proceeds mainly at the silicon centre of the molecule and not at the epoxy part, the epoxy

J Sol-Gel Sci Technol

ring being shown to remain intact. The latter is related to the fact that no cross-linker was added to the sol to stimulate ring opening, while acetic acid is not strong enough to catalyse epoxy-ring opening. Almost complete hydrolysis of the methoxy group occurred, resulting in formation of Si–OH and methanol. Acetic acid acts as a catalyst of this reaction. The final sol, obtained by combining the two sols, contains Zr–O–Zr and Si–O–Si. Acetic acid, bonded as acetate to the Zr-network of Sol 1, is released but the majority remained, acting as a bidentate ligand within the Zr-network. The release of acetate would, presumably, also lead to the formation of Zr–O–Si bonds. On the other hand, a siloxane network may already be forming, leading to the glycidoxypropyl group being in a different environment and hence giving rise to broad peaks in NMR spectra. The epoxy group remained in the final sol. The sol is thermally stable up to 440 °C. After deposition of the sol in the form of a coating on titanium substrate, the curing was followed at room temperature and at 150 °C in order to investigate reaction of the epoxy group. During curing at room temperature this group was present for up to 24 days after deposition, but disappeared completely after 30 days. Polymerization of the organic network occurred slowly and acetate eventually evaporated from the coating. During curing at 150 °C, the latter process proceeded immediately, together with opening of the epoxy ring. Thus, in the absence of cross-linker, the reaction of ring opening is shown to be catalysed thermally. The sol, synthesized in the present work and deposited in the form of a coating on titanium substrate, will be the basis for anticorrosive and antimicrobial protection as will be shown in the second part of this study. Acknowledgments Financial support by the Slovene Research Agency is acknowledged (Grants Nos. P2-0148 and P1-0134). The authors thank the Centre of Excellence for Integrated Approaches in Chemistry and Biology of Proteins (CIPKeBiP) for use of the ReactIRTM 45 FTIR instrument and accompanying equipment, Prof. Barbara Malicˇ from the Jozˇef Stefan Institute and Centre of Excellence NAMASTE, Advanced Materials and Technologies for the Future, for the use of PerkinElmer Spectrum 100 instrument, and the National

NMR Centre at the National Institute of Chemistry. The authors thank Prof. R.H. Pain for proof reading the manuscript.

References 1. Innocenzi P, Brusation G, Guglielmi M (1999) Bertani. Chem Mater 11:1672–1679 2. Metroke TL, Kachurina O, Knobbe ET (2002) Prog Org Coat 44:295–305 3. Davis SR, Brough AR, Atkinskon A (2003) J Non-Cryst Solids 315:197–205 4. Oliver MS, Blohowiak KY, Dauskardt RH (2010) J Sol–Gel Sci Technol 55:360–368 5. Fontinha IR, Salta MM, Zheludkevich ML, Ferreira MGS (2013) Port Electrochim Acta 31:307–319 6. Riegel B, Blittersdorf S, Kiefer W, Hofacker S, Mu¨ller M, Schottner G (1998) J Non-Cryst Solids 226:76–84 7. Zhao J, Fan W, Wu D, Sun Y (2000) J Non-Cryst Solids 261:15–20 8. Rubio F, Rubio J, Oteo JL (1998) J Mater Sci Lett 17:1839–1842 9. Yoldas BE (1986) J Mater Sci 21:1080–1086 10. Schmidt H, Seiferling B (1986) MRS proceedings. doi:10.1557/ PROC-73-739 11. Feng Z, Liu Y, Thompson GE, Skeldon P (2010) Surf Interace Anal 42:306–310 12. Voevodin NN, Grebasch NT, Soto WS, Kasten LS, Grant JT, Arnold FE, Donley MS (2001) Prog Org Coat 41:287–293 13. Rodicˇ P, Iskra J, Milosˇev I (2014) J Sol–Gel Sci Technol 70:90–103 14. Rodicˇ P, Iskra J, Milosˇev I (2014) J Non-Cryst Solids 396–397: 25–35 15. Hoebbel D, Nacken M, Schmidt H (2000) J Sol–el Sci Technol 19:305–309 16. Philipp G, Schmidt H (1986) J Non-Cryst Solids 82:31–36 17. Jay RR (1964) Anal Chem 36:667–668 18. Matejka L, Dukh O, Brus J, Simonsick WJ, Meisnner B (2000) J Non-Cryst Solids 270:34–47 19. Yi G, Sayer M (1996) J Sol–Gel Sci Technol 6:65–74 20. Voevodin NN, Grebasch NT, Soto WS, Arnold FE, Donley MS (2001) Surf Coat Techol 140:24–28 21. Noonan GO, Ledford JS (1995) Chem Mater 7:1117–1123 22. Liu W-C, Yang C-C, Chen W-C, Dai B-T, Tsai M-S (2002) J Non-Cryst Solids 311:233–240 23. Albrect MG, Blanchette C (1998) J Electrochem Soc 145: 4019–4025 24. Park ES, Ru HW, Nguyen CV, Jaffe RL, Yoon DY (2008) Chem Mater 20:1548–1554 25. Deacon GB, Phillips RJ (1980) Coord Chem Rev 33:227–250 26. Ochi M, Nii D, Harada M (2011) Mater Chem Phys 129:424–432

123