2270
Langmuir 2004, 20, 2270-2273
Hydrolytic Stability of Organic Monolayers Supported on TiO2 and ZrO2 Stephen Marcinko and Alexander Y. Fadeev* Department of Chemistry and Biochemistry, Seton Hall University, South Orange, New Jersey 07079 Received May 27, 2003. In Final Form: October 14, 2003 The hydrolytic stability of C18 monolayers supported on TiO2 and ZrO2 was studied. Three types of monolayers were prepared from the following octadecyl modifiers: (1) octadecyldimethylchlorosilane (C18H37Si(CH3)2Cl); (2) octadecylsilane (C18H37SiH3); and (3) octadecylphosphonic acid (C18H37P(O)(OH)2). The hydrolysis of the surfaces prepared was studied under static conditions at 25 and 65 °C at pH 1-10. On the basis of the loss of grafted material, the stability of the monolayers fall in the following range: C18H37P(O)(OH)2 g C18H37SiH3 . C18H37Si(CH3)2Cl. At 25 °C, monolayers from C18H37P(O)(OH)2 showed only ∼2-5% loss in grafting density after one week at pH 1-10. The high stability of these monolayers was explained because of the strong interactions of the phosphonic acids with the substrates. Monolayers from C18H37Si(CH3)2Cl showed poor hydrolytic stability at any pH, which was explained because of the low stability of Ti-O-Si and Zr-O-Si bonds. Unlike monofunctional silanes, trifunctional silane (C18H37SiH3) yielded surfaces that showed good hydrolytic stability. This suggests that the stability of the monolayers from trifunctional silanes is primarily due to “horizontal” bonding (Si-O-Si or Si-OH...HOSi) rather than due to bonding with the matrix (M-O-Si). At 65 °C, all C18 surfaces become more susceptible to hydrolysis; however, the trend observed for 25 °C remained unchanged. Low-temperature nitrogen adsorption was used to study the adsorption properties of the monolayers as a function of their grafting density. The energy of adsorption interactions showed a significant increase as the grafting density of the monolayers decreased. The order of the alkyl groups in the monolayers, as assessed from CH2 stretching, decreased as the grafting density of the monolayers decreased.
Introduction Surface modification of titanium and zirconium dioxides is of great interest in view of the importance of these surfaces in biomaterials, catalysts, photocatalysts, pigments, fillers, and so forth. One of the most powerful approaches for the surface modification of metal oxides uses covalent attachment of organic monolayers. Although several synthetic approaches for the surface modification of TiO2 and ZrO2 have been reported, very little is known about the stability of different supported surfaces. Hydrolytic stability is the key parameter that determines the performance of final materials and it is critical in choosing a synthetic strategy for surface functionalization. The reaction of monofunctional silanes (R3SiX) has been reported by several groups.1-5 Different tri-alkylsilanes (with X ) Cl, OR, N(CH3)2, OH, H) react with the surface M-OH groups to yield M-O-Si chemical bonds:
The use of trifunctional organosilanes (RSiX3, X ) Cl, OCH3, H) have been reported.5,9-14 Under certain conditions, the reaction of long-chained trifunctional silanes gives monolayers that are closely packed and highly ordered. These monolayers are usually referred to as self-assembled monolayers (SAMs). Molecules in the SAMs are strongly bonded with their neighbors and, to a small extent, with the surface. The SAMs may be considered a 2-D siloxane network that is supported by the surface:9,10,14
The use of organophosphonic acids (RPO(OH)2)5,15-17 has been reported for titania, zirconia, and other metal
A similar mechanism (formation of Ti-O-Si bonds) was reported for the reaction of several tri- and tetrafunctional silane coupling agents (Si(OEt)4 and RSi(OEt)3) with single-crystal (110) TiO2 under UHV conditions.6-8 * Address correspondence to this author. (1) Amati, D.; Kova`ts, E.sz. Langmuir 1988, 4, 329. (2) Yatsuk, S. P.; Brey, V.V.; Chuiko, A. A. Russ. J. Phys. Chem. 1988, 62, 1940. (3) Yu, J.; El Rassi, Z. J. Chromatogr. 1993, 631, 91. (4) Fadeev, A. Y.; McCarthy, T. J. J. Am. Chem. Soc. 1999, 121, 12184. (5) Helmy, R.; Fadeev, A. Y. Langmuir 2002, 18, 8924.
(6) Gamble, L.; Hugenschmidt, M. B.; Campbell, C. T.; Jurgens, T. A.; Rogers, J. W. J. Am. Chem. Soc. 1993, 115, 12096. (7) Gamble, L.; Jung, L. S.; Campbell, C. T. Langmuir 1995, 11, 4505. (8) Gamble, L.; Henderson, M. A.; Campbell, C. T. J. Phys. Chem. B 1998, 102, 4536. (9) Moses, R. P.; Wier, L. M.; Lennox, J. C.; Finklea, H. O.; Lenhard, J. R.; Murray, R. W. Anal. Chem. 1978, 50, 576. (10) Finklea, H. O.; Murray, R. W. J. Phys. Chem. 1979, 83, 353. (11) Tsubokawa, N.; Kogure, A. J. Polym. Sci. 1991, 29, 697. (12) Mahon, M.; Wulser, K. W.; Langell, M. A. Langmuir 1991, 7, 486. (13) Xiao, S.-J.; Textor, M.; Spencer, N. D. Langmuir 1998, 14, 5507. (14) Fadeev, A. Y.; Helmy, R.; Marcinko, S. Langmuir 2002, 18, 7521. (15) Randon, J.; Blanc, P.; Paterson, R. J. Membr. Sci. 1995, 98, 119. (16) Gao, W.; Dickinson, L.; Grozinger, C.; Morin, F. G.; Reven, L. Langmuir 1996, 12, 6429. (17) Ramser, R. D.; Henriksen, P. N.; Gent, A. N. Surf. Sci. 1988, 203, 72.
10.1021/la034914l CCC: $27.50 © 2004 American Chemical Society Published on Web 02/07/2004
Hydrolytic Stability of Organic Monolayers
Langmuir, Vol. 20, No. 6, 2004 2271
oxides. Phosphonic acids interact strongly with metal oxides, probably through the formation of robust M-O-P bonds:
The use of alkylhydroxamic18 and alkylcarbonic19,20 acids for surface modification of titania and zirconia has also been reported. In the present work, we report the results on the hydrolytic stability of different C18-monolayers supported on TiO2 and ZrO2. Experimental Section General Information. Octadecyldimethylchlorosilane and octadecylsilane (Gelest, Tullytown, PA) and octadecylphosphonic acid (Oryza Labs, Chelmsford, MA) were used as received. HPLC grade solvents and all other chemicals were purchased from Aldrich and Fisher. Titanium dioxide (anatase, SN2(BET) ) 10 ( 1) was purchased from Aldrich. Zirconium dioxide (monoclinic, SN2(BET) ) 17 ( 1) was donated by Magnesium Electron, Inc. Infrared spectra were obtained with a Perkin-Elmer Spectrum One FTIR instrument with an MCT detector. Spectra were collected in reflectance mode using a Harrick Seagull accessory (Ossining, NY). All spectra were acquired at a 45° angle of incidence with 32 scans at 4 cm-1. Nitrogen adsorption isotherms were measured using the Coulter Omnisorp model 100CX automated gas sorption analyzer. Nitrogen was purchased from Airgas (Piscataway, NJ). The adsorption isotherms were measured to a maximum p/p0 of 0.50. Analysis of the adsorption isotherms was made using the BET equation:21
()
1 C-1 p p ) + nmC po n(po - p) nmC
(4)
where n is the amount adsorbed (mmol/g) at equilibrium pressure p, and nm is the monolayer capacity. C is the constant, which is related to average heat of adsorption:21
∆H = RT ln C
(5)
Grafting density (F, molecules/nm2) of the monolayers was determined from elemental analysis using the formula
F)
6 × 105‚(%C) 1 ‚ [2400‚nc - FW ‚(%C)] S(BET)
(6)
where %C is carbon percentage in the sample, nc is the number of carbon atoms present in grafted molecule, FW is the formula weight of grafted molecule, and S(BET) is the BET surface area of metal oxide. Chemical analysis was performed by Schwarzkopf Microanalytical Laboratories (Woodside, NY) using the ASTM method. Thermogravimetric analysis studies were performed using a TA Hi-Res 2950 Thermogravimetric Analyzer operated between room temperature and 800 °C at a rate of 10 °C/min. During the investigations, we found that there is a linear correlation between the TGA weight loss measured in the range 200-500 °C and carbon percentage in modified metal oxides (Figure 1). As can be seen from Figure 1, linearity was established over a range of 0.5-3 %C with an R2 value of 0.998. This correlation provides a convenient way for quick determination of grafting density directly from the TGA weight loss data (WL, wt %):
F)
6 × 105 ‚(0.886 × WL) 1 ‚ [2400 ‚nc - FW ‚(0.886 × WL)] S(BET)
(7)
(18) Folkers, J. P.; Gorman, C. B.; Laibinis, P. E.; Buchholz, S.; Whitesides, G. M.; Nuzzo, R. G. Langmuir 1995, 11, 813.
Figure 1. Correlation between carbon percentage and TGA weight loss for MO2-supported C18 monolayers. Straight lines and the regression coefficients were obtained via the leastsquares root method. Synthesis of Monolayers. The monolayers were prepared by the reactions of octadecyldimethylchlorosilane (C18H37Si(CH3)2Cl), octadecylsilane (C18H37SiH3), and octadecylphosphonic acid (C18H37PO(OH)2) with titanium and zirconium dioxides. Approximately 0.5 g of metal oxide powder was placed in a GC headspace vial and dried overnight in an oven at 120 °C. A 10mL solution of C18H37Si(CH3)2Cl, C18H37SiH3, or C18H37PO(OH)2 in toluene was injected into the vials using a syringe. The solutions contained 25 µmol of modifier per each meter square of the surface of metal oxide. This corresponds to a ∼three-fold excess in respect to the complete monolayer coverage of the surface with alkyl groups. The reaction vessels were placed in a temperaturecontrolled water bath at 80 °C for 120 h after the addition of the solution. The reactions were quenched by filtering on a Bu¨chner funnel with a fritted disk. The samples were washed with 3 × 20 mL of reagent grade toluene followed by 3 × 20 mL of acetone. The samples were allowed to dry on the filter and then transferred to a 60 °C oven and dried overnight. Hydrolysis of Monolayers. A 1:1 solution of THF and aqueous solution at each of the individual pH values (1-10) was added to a GC headspace vial containing 0.5 g of adsorbent. The reagents and their corresponding pH values were the following: HCl (0.1 M, pH ) 1.1 ( 0.1), potassium acid phthalate (0.05 M, pH ) 4.1 ( 0.1), DI water (pH ) 6.9 ( 0.1), tris-(hydroxymethyl)aminomethane (0.1 M, pH ) 10.3 ( 0.1). All pH measurements were performed using an Orion Research digital ionalyzer 501. The presence of THF in aqueous solutions was to ensure complete wetting of the hydrophobic adsorbents. A first series of the reaction vessels were left at room temperature for 1 week. A second series of reaction vessels were allowed to react for a given time (1-168 h) at 65 °C. Each series of samples were quenched by filtering on a Bu¨chner funnel with a fritted disk. The samples were then washed with THF (4 × 20 mL), dried on the filter, and placed in a 60 °C oven overnight.
Results and Discussion The C18 surfaces were prepared via reactions of nonporous TiO2 (anatase) and ZrO2 (monoclinic) powders with C18H37Si(CH3)2Cl (eq 1), C18H37SiH3 (eq 2), and C18H37P(O)(OH)2 (eq 3). The reactions were monitored by IR. IR spectra for all the reactions were consistent with the attachment of octadecyl-siloxanes (for silanes) and octadecylphosphonates (for phosphonic acid) to the surface. (19) Pawsey, S.; Yach, K.; Halla, J.; Reven, L. Langmuir 2000, 16, 3294. (20) Marguerettaz, X.; Fitzmaurice, D. Langmuir 1997, 13, 6769. (21) Brunauer, S., Emmet, P. H.; Teller, E. J. Am. Chem. Soc. 1938, 60, 309.
2272
Langmuir, Vol. 20, No. 6, 2004
Marcinko and Fadeev
Table 1. Grafting Density and Selected IR Stretches for C18 Monolayers Prepared in This Work TiO2
C18 modifier
grafting density, group/nm2
C18H37Si(CH3)2Cl C18H37SiH3 C18H37PO(OH)2
1.2 4.1 3.5
ZrO2 νa(CH2), cm-1
grafting density, group/nm2
νa(CH2), cm-1
2924 2918 2918
1.1 4.4 3.7
2925 2919 2919
Surfaces prepared from C18H37SiH3 and C18H37P(O)(OH)2 were characterized by close packing (∼3.5-4.4 molecules/ nm2) and highly ordered alkyl groups, as assessed from CH2 stretching. Surfaces prepared from C18H37Si(CH3)2Cl had a grafting density ∼1.5 molecules/nm2 and were rather disordered. Given the fact that the density of surface OH groups is ∼8 group/nm2 (anatase36) and ∼12 group/ nm2 (monoclynic zirconia),22 only a small fraction of these groups can be attached to the organic molecules in the monolayer. The characteristics of the C18 surfaces prepared are given in Table 1. For more details about the reactions (1-3) with TiO2 and ZrO2 and the characterization of the monolayers, see our previous works.5,14 Hydrolytic stability of the C18 monolayers prepared was tested under static conditions at 25 and 65 °C using waterTHF mixtures with pH ranging from 1 to 10. The loss of grafted material because of hydrolysis was determined through chemical analysis and TGA. Figure 2 presents the kinetics plots for the hydrolysis of the monolayers. Since there was no significant hydrolysis of the monolayers at 25 °C (with the exception of C18H37Si(CH3)2Cl), all kinetics experiments were carried out at 65 °C. As one can see from Figure 2, most of the kinetic plots level off after ∼50 h. For the convenience of data analysis, the extent of hydrolysis for all monolayers was determined after 1 week (168 h) as follows:
θHYDROLYZED ) 1 -
F F0
where F is the grafting density after hydrolysis and F0 is the grafting density prior to hydrolysis. Monolayers of C18H37Si(CH3)2Cl. The hydrolytic stability of these surfaces was quite low. Even at neutral pH and 25 °C, the monolayers were hydrolyzed almost completely (θHYDROLYZED > 0.9 for both TiO2 and ZrO2) within 3-24 h. Acidic and basic solutions hydrolyzed the monolayers even faster. (22) Nawrocki, J.; Rigney, M. P.; McCormick, A.; Carr, P. W. J. Chromatogr., A 1993, 657, 229. (23) Schindler, F.; Schmidbaur, H. Angew. Chem., Int. Ed. Engl. 1967, 6, 683. (24) Britt, D. W.; Hlady, V. J. Colloid Interface Sci. 1996, 178, 775. (25) Rye, R. R.; Nelson, G. C.; Dugger, M. T. Langmuir 1997, 13, 2965. (26) Parikh, A. N.; Allara, D. L.; Azouz, I. B.; Rondelez, F. J. Phys. Chem. 1994, 98, 7577. (27) Thomas, R. C.; Sun, L.; Crooks, M. Langmuir 1991, 7, 620. (28) Fadeev, A. Y.; McCarthy, T. J. Langmuir 2000, 16, 7268. (29) Kessel, C. R.; Granick, S. Langmuir 1991, 7, 532. (30) Allara, D. L.; Parikh, A. N.; Rondelez, F. Langmuir 1995, 11, 2357. (31) Gregg, S. J.; Sing, K. S. W. Adsorption, Surface Area and Porosity; Academic Press: New York, 1982. (32) Kazakevich, Y. V.; Fadeev, A. Y. Langmuir 2002, 18, 3117. (33) Snyder, R. G.; Straus, H. L.; Elliger, C. A. J. Phys. Chem. 1982, 86, 5145. (34) Nuzzo, R. G.; Dubois, L.H.; Allara, D. L. J. Am. Chem. Soc. 1990, 112, 558. (35) Parks, G. A. Chem Rev. 1965, 65, 177. (36) Parfitt, G. D. The surface of titanium dioxide. In Prog. Surf. Membr. Sci.; Cadenhead, D. A., Danielli, J. F., Eds.; Academic Press: New York-London, 1976; Vol. 11.
Figure 2. Kinetics of hydrolysis of ZrO2-supported C18 monolayers at 65 °C at different pH. Monolayers from C18H37SiH3 (diamonds), monolayers from C18H37P(O)(OH)2 (circles), monolayers from C18H37Si(CH3)2Cl (triangles).
The results obtained agree well with the literature,3,4,22 where low efficiency of monofucntional silanes for the purposes of surface functionalization of TiO2 and ZrO2 was reported. Other researchers,1,2 however, reported the preparation of closely packed organic surfaces on TiO2 and ZrO2 from monofunctional silanes. Close analysis of these and previous works1-4,22 shows that there are no contradictions between the data. Monolayers of monofunctional silanes are obtained because of the formation of siloxane (Si-O-M) bonds with the metal atom (M) in the metal oxide (eq 1). The siloxane bond, however, is stable only for M)Si and is not so stable for M)Ti and Zr.23 Apparently, monolayers with high surface coverage can be prepared using reaction 1, but one can hardly expect good stability of these surfaces in contact with water or water-organic solutions. Monolayers of C18H37SiH3. At 25 °C the monolayers of C18H37SiH3 showed almost no hydrolysis at pH 1-10. The values of θHYDROLYZED for the TiO2 and ZrO2 supported surfaces were ∼0.02-0.05. At 65 °C the monolayers become more susceptible to hydrolysis, especially at low
Hydrolytic Stability of Organic Monolayers
Langmuir, Vol. 20, No. 6, 2004 2273
Table 2. The Extent of Hydrolysis (θHydrolyzed) of the C18 Monolayers after 1 Week of Contact with Solutions of Different pHs at 65 °C C18H37Si(CH3)2Cl
C18H37SiH3
C18H37PO(OH)2
pH
TiO2 and ZrO2
TiO2
ZrO2
TiO2
ZrO2
1 4 7 10
g0.98a
0.85 0.45 0.25 0.30
0.78 0.40 0.25 0.31
0.20 0.15 0.16
0.13 0.05 0.05 0.15
g0.98a g0.98a g0.98a
a The amount of grafted materials present on the surface was less than the detection limit of chemical analysis.
pH. The values of θHYDROLYZED were ∼ 0.25-0.4 for pH 4-10 and ∼ 0.8 for pH 1 (Table 2). Table 2 shows that the hydrolytic stability of the monolayers from C18H37SiH3 is notably better than for the monolayers from monofunctional silanes. Also, the data in Table 2 suggests that metal oxides play no essential role in the hydrolytic stability of the monolayers. As mentioned in the previous section, the hydrolytic stability of Ti-O-Si or Zr-O-Si bonds is quite low and cannot explain the high stability of the monolayers from C18H37SiH3. This suggests that the hydrolytic stability of the monolayers from trifunctional silanes is primarily due to “horizontal” bonding, while bonding with the matrix plays very little (if any) role in the stability of these monolayers (eq 2). The results obtained are consistent with the structure proposed in the literature for selfassembled monolayers from trifunctional silanes.24-28 The strong bonding between molecules of alkylsilanes (lateral siloxane (Si-O-Si) and hydrogen bonds (Si-OH...HOSi), and van der Waals interactions between alkyl chains) and to a small extent bonding with the surface (M-O-Si) is a unique property of these monolayers. This, for instance, allows the preparation of monolayers of similar quality on different substrates, whether covalent bonds can form with the substrate, for example silica, mica and gold.29,30 Monolayers of C18H37PO(OH)2. The monolayers of C18H37P(O)(OH)2 demonstrated the best hydrolytic stability among the studied surfaces. The extent of hydrolysis (θHYDROLYZED) for the TiO2 and ZrO2 supported surfaces was ∼0.02-0.05. At 65 °C, the values of θHYDROLYZED were ∼0.05-0.15 for zirconia-supported surfaces and ∼0.150.2 for titania- supported surfaces (Table 2). High stability of these monolayers is explained because of strong specific interactions of phosphonic acid group with the surfaces of metal oxides. A somewhat higher hydrolytic stability of the monolayers supported on ZrO2 than those on TiO2 may be rationalized using isoelectric points (IEP) of the metal oxides.35 The surface of ZrO2 is more basic (IEP ∼ 7-11) than this of TiO2 (IEP ∼ 4-7), which results in a stronger bonding of phosphonic acids. Adsorption Properties and Order in the C18 Monolayers as a Function of θHYDROLYZED. The study of vapor phase adsorption provides insight into the structure of the monolayers at the solid-vapor interface. For bare metal oxide and C18-MO2, the nitrogen adsorption isotherms are Type II14 which, according to the classification,31 indicates physical adsorption of N2 on the surface. The intensity of the adsorption interactions can be assessed from the constant C of the BET equation, which is related to the average heat of adsorption (eq 5).21 The high value of the C constant (∼80) obtained for bare TiO2 and ZrO2 demonstrated the presence of high-energy, polar surfaces. Coating of the surface of metal oxides with hydrophobic alkyl groups resulted in a decrease of the
Figure 3. BET C constant for nitrogen adsorption on the MO2supported C18 surfaces as a function of grafting density. Monolayers of C18H37SiH3 on TiO2 (diamonds), monolayers of C18H37P(O)(OH)2 on ZrO2 (circles).
Figure 4. The position of the CH2-stretching for the ZrO2supported monolayers of C18H37P(O)(OH)2 as a function of their grafting density.
energy of the adsorption interactions and a drop of the C constants. For the monolayers of C18H37P(O)(OH)2 and C18H37SiH3, the values of C constants were around 12, indicating the presence of closely packed CH3 surfaces.1,32 Hydrolysis of the monolayers resulted in a notable increase of the C constant. Figure 3 shows that the BET C constants increased as the monolayer’s grafting density decreased. This trend is consistent with the appearance of polar adsorption centers (OH groups on the surface of metal oxide) on the surfaces with incomplete monolayer coverage. The ordering of alkyl groups in the monolayers can be assessed from the infrared spectra using the position of CH2 stretch. For completely disordered structures, it is close to that of a liquid alkane (νa ∼ 2924 cm-1). For wellordered monolayers, it is shifted to lower wavenumber approaching this for crystalline alkane (νa ∼ 2915-2918 cm-1).33,34 Table 1 shows that the monolayers of C18H37SiH3 and C18H37P(O)(OH)2 are highly ordered (νa ∼ 2918-2919 cm-1). As the grafting density of the monolayers decreases, the CH2 stretching shifts to higher wavenumbers (Figure 4), indicating an increase of the disorder of alkyl groups. Acknowledgment. Authors would like to acknowledge Dr. Y. V. Kazakevich for valuable assistance with adsorption measurements. The support from the NSF (CMS0304098) is acknowledged. The support from Seton Hall Research Council is acknowledged. We also would like to thank Jonathan Simmonds for his helpful participation. LA034914L
