Langmuir 2004, 20, 2707-2712
2707
In Situ Determination of the Thermodynamic Surface Properties of Chemically Modified Surfaces on a Local Scale: An Attempt with the Atomic Force Microscope Olivier Noel,* Maurice Brogly, Gilles Castelein, and Jacques Schultz Institut de Chimie des Surfaces et Interfaces (ICSI-CNRS-UPR 9069), BP 2488, 68057 Mulhouse Cedex, France Received May 22, 2003. In Final Form: December 29, 2003 We have monitored deflection-distance curves with an atomic force microscope (AFM) in contact mode, with a silicon nitride tip, on chemically modified silicon wafers, in the air. The wafers were modified on their surface by grafting self-assembled monolayers (SAMs) of different functional groups such as methyl, ester, amine, or methyl fluoride. A chemically modified surface with a functionalized hydroxyl group was also considered. Qualitative analysis allowed us to compare adhesive forces versus chemical features and surface energy. The systematic calibration procedure of the AFM measurements was performed to produce quantitative data. Our results show that the experimentally determined adhesive force or thermodynamic work of adhesion increases linearly with the total surface energy determined with contact angles measured with different liquids. The influence of capillary condensation of atmospheric water vapor at the tipsample interface on the measured forces is discussed. Quantitative assessment values were used to determine in situ the SAM-tip thermodynamic work of adhesion on a local scale, which have been found to be in good agreement with quoted values. Finally, the determination of the surface energy of the silicon wafer deduced from the thermodynamic work of adhesion is also proposed and compared with the theoretical value.
I. Introduction Self-assembled monolayers (SAMs) are considered as model surfaces in the way that they show well-ordered structures and allow a control of interactions at a microscopic scale. Investigations dealing with the chemical and mechanical properties of these molecular films toward technological applications in biotechnology, opto- and microelectronics, tribology, lubricants, and photolithography are numerous. There are many insightful techniques to get insights of the surface properties of SAMs. Nevertheless, the main advantage of the atomic force microscope (AFM) on other techniques is that accurate measurements of normal surface forces at a quasi-atomic scale resolution can be achieved. This paper focuses on the quantitative probing of chemical surface properties of SAMs with the AFM. Different SAMs, including functionalized alkyltrichlorosilanes or alkylmethoxysilanes grafted on an oxidized silicon wafer [Si(100) surface] were synthesized. Silicon wafers offer a smoother plane and a more organized and homogeneous structure compared to gold substrates1 for SAM elaboration. Each SAM system consists of three main parts. The first part is the headgroup, which chemisorbs to the surface. In the second part, an alkyl chain forms, in the ideal case, a homogeneous well-packed ordered monolayer. The terminal group at the end of the chain is supposed to interact with the probe through mainly van der Waals (VDW) and capillary interactions. We are fully dependent on the synthesis conditions, and rather than monolayers, we could obtain multilayers or not fully extended molecules grafting on the substrate. In this study, we mainly focus on the chemical surface properties of the SAMs and not on their structures. Nevertheless, SAM structure will be briefly * Corresponding author: e-mail
[email protected]. (1) Sung, M. M.; Kluth, G. J.; Yauw, O. W.; Maboudian, R. Langmuir 1997, 13, 6164-6168.
described on the basis of ellipsometric, contact angle, and X-ray photoelectron spectroscopy (XPS) experiments. Methyl (CH3), amine (NH2), methyl fluoride (CF3), and ester [CO(OCH3)] end-group SAMs, as well as hydroxyl (OH) functionalized surfaces, were synthesized and compared. For technical reasons, all the AFM measurements were done in the air at about 20 °C with a silicon nitride tip (Si3N4). At these conditions, the influence of the capillary forces cannot be neglected. According to previous studies, for a relative humidity, which lies between 20-40%, discontinuities could appear in the force measurements as a result of the formation of a capillary meniscus, depending on the surface chemistry.2 Because capillary forces can critically affect the measurements, the relative humidity was recorded for each experiment and found to be around 30%. Finally, to determine the intrinsic surface properties of our SAMs, the contributions of these capillary effects have been discussed. A unique feature of this paper is that, although there are numerous publications dealing with surface force measurements with the AFM, few of them insist on the experimental difficulties related to the AFM technique itself. A lack of precautions in recording force-distance curves can lead to irreproducibility or to a misinterpretation.3 Provided uncertainties, which have been listed and minimized, we propose in situ quantitative tip-sample thermodynamic work of adhesion determinations at a local scale, which are compared to macroscopic surface properties related to the surface chemistry. II. Experimental Section A. Materials. Silicon wafers (100; supplied by MAT Technology, France) polished on one side were used as the substrate for the SAM film grafting. In this paper, “as-received silicon (2) Mingyan, H.; Blum, A. S.; Aston, E.; Buenviaje, C.; Overney, R. M; Luginbu¨hl, R. J. Chem. Phys. 2001, 114, 1355-1360. (3) Haugstad, G.; Gladfelter, W. L. Ultramicroscopy 1994, 54, 3140.
10.1021/la034884m CCC: $27.50 © 2004 American Chemical Society Published on Web 03/03/2004
2708
Langmuir, Vol. 20, No. 7, 2004
Noel et al.
Table 1. Wetting Properties with Watera and Wetting Hysteresis end groups
θe (deg) (2°
θa (deg) (2°
θr (deg) (2°
wetting hysteresis (deg)
CF3 CH3 Sias-received COOR NH2 OH
106 103 78 71 57 6
117 111 81 75 59
86 101 70 59 21
31 10 11 16 38
Table 2. Wetting Properties with TCPa and Contact Angle Measurements with Water and TCP on the AFM Tip end groups
θe (deg) (2°
CF3 CH3 Sias-received
67 66 47
a
θe is an equilibrium contact angle, θa is an advancing contact angle, and θr is a receeding contact angle.
AFM tip a
(Sias-received or Siwafer)” refers to a silicon wafer previously cleaned in hexane and then in ethanol in an ultrasonic bath. That means a contaminated layer still remains on the surface. Four organosilane grafts (supplied by ABCR Karlsruhe, Germany) were used for the elaboration of homogeneous model surfaces on the substrate. Two hydrophobic model surfaces were prepared by using hexadecyltrichlorosilane (C16H42O3Si or SiCH3) and 1H,1H,2H,2H-perfluorodecylmethyldichlorosilane (C11H7Cl2F17Si or SiCF3), and two hydrophilic model surfaces were prepared by using (6-aminohexyl)aminopropyltrimethoxysilane (C12H30N2O3Si or SiNH2) and 2-(carbomethoxy)ethyltrichlorosilane (C4H7Cl3O2Si or Siester). All other chemicals used in the chemical handling (cleaning, synthesis) were of reagent grade or better (supplied by Aldrich). Bidistilled and deionized water was used for the contact angle measurements. B. Sample Preparation. B.1. Preparation of the Oxidized Silica Surface. Before grafting, the substrates must be chemically modified to get a hydrophilic surface (SiO2). The silicon surface was first cleaned with ethanol and dried with nitrogen before oxidation. Oxidized surfaces were obtained after cleaning by submersing the substrate in a warm piranha (60 °C) solution [3:7 (v/v) 30% H2O2/H2SO4 mixture] for about 30 min to keep a smooth surface and then thoroughly rinsing with deionized and bidistilled water. Just before being grafted with silane, the wafers are dried with nitrogen. This treatment produces a high hydroxyl group density on the surface (SiOH groups), to which functional silanes will adsorb upon hydrolysis.4 Surfaces modified with hydroxyl groups (SiOH) were synthesized with this method and immediately probed to avoid contamination of the surface by the environment due to the high reactivity of SiOH groups. B.2. Functionalized SAMs. Procedure. One of the numerous techniques to obtain SAMs, the vapor-phase molecular selfassembling technique,5 slightly improved in the laboratory,4 was used. The lack of solvent prevents the SAMs from incorporating small-solvent-molecule contamination and defects. Moreover, a previous study6 showed that the molecular films prepared with this method are more homogeneous. In this method, the oxidized silica surfaces are placed above a previously de-aired solution of a 100 µL of organosilane/3 mL of paraffin mixture. The vaporphase deposition of the molecular film on the substrate is performed in a vacuum chamber (50 min at 5 × 10-3 Torr) at room temperature. B.3. Monolayer Characterization. Contact angle measurement is an effective way to characterize surface hydrophilicity and hydrophobicity. Wetting properties with water in a thermostated room (20 °C) are measured with a Kru¨ss G2 system. The volume of the water drop in air lies between 1 and 2 µL, and the measurement chamber was previously saturated with water vapor. The equilibrium static angle, and advancing and receeding angles are reported in Table 1. It is interesting to note the expected correlation between contact angle measurements and the hydrophobic-hydrophilic character of the end group of the grafts. Such evolution confirms the formation of SAMs on the silicon wafers. Moreover, the (4) Mougin, K.; Haidara, H.; Castelein, G. Colloids Surf., A 2001, 193, 231-237. (5) Chaudhury, M. K.; Whitesides, G. M. Science 1992, 255, 12301232. (6) Vonna, L. Ph.D. Thesis, Universite´ de Mulhouse, Mulhouse, France, 1999.
θe (deg) (2°
end groups COOR NH2 OH
37 28 25
θe (deg) water (2°
θe (deg) TCP (2°
80
42
θe is an equilibrium contact angle. Table 3. Theoretical Surface Energiesa
end groups CF3 CH3 Sias-received COOR NH2 OH AFM tip
-2 -2 -2 γds (mJ‚m-2) γnd s (mJ‚m ) γs (mJ‚m ) (1 mJ‚m
21 21 25 34 38 39 29
0 1 7 9 15 37 6
21 22 32 43 53 76 35
γdS is the dispersive component (due to VDW contributions), and γnd S is the nondispersive (or polar) component of the surface energy. γs is the total energy. a
wetting hysteresis is low (except with the SiCF3 and SiNH2 surfaces). This means that our functionalized surfaces show high compacity and homogeneity on a macroscopic scale. Wetting properties with tricresyl phosphate (TCP) were also determined in the same conditions as with water (Table 2). We have also reported in Table 2 the contact angle measurements with water and TCP on the AFM tip. These results show that contact angle measurements obtained on the AFM tip and on the Sias-received wafer are very close. From contact angle measurements with water and TCP, theoretical surface energy components were calculated from the Good-Girifalco7 theories (Table 3). Table 3 shows that the nondispersive components are negligible for all the surfaces, except for the SiNH2 and SiOH surfaces. In these last cases, interactions with water through hydrogen bonding can occur, and, thus, capillary forces have to be considered. Moreover, the total surface energy increases with the hydrophilicity of the surfaces. Surface morphologies of the different types of layers were studied by the AFM (Nanoscope III, Digital Instruments) in the tapping mode. The SAM’s topographic and phase contrast AFM images (image size: 1 µm × 1 µm) confirmed that no aggregates were formed and showed a complete homogeneous recovery of the grafts. Moreover, the root-mean-square roughness of 0.15 nm deduced from the AFM images, which was the expected value for the substrate roughness, confirmed that our grafts were well ordered and packed on micrometer scales (this is also the case for the SiCF3 and SiNH2 surfaces). XPS (Leybold LHS 11) with a takeoff angle of 15° was also carried out. Comparison of the spectra related to the different types of molecules grafted on the silicon wafer with the spectra obtained on the silicon wafer confirms the effective grafting. Moreover, no peaks related to the chloride element have been detected regarding the spectra related to the chlorinated grafts. This confirms the efficiency of the covalent grafting and of the surface cleaning procedure. Finally, the thicknesses of the oxide and organic films were determined by ellipsometry measurements by fitting both the refractive index and the thickness of the film (Table 4). These results are in good agreement with the literature for the CH3 and the NH2 grafted films8,9 and confirm that only one monolayer (7) Girifalco, L. A.; Good, R. J. J. Phys. Chem. 1957, 61, 904. (8) Wood, J.; Sharma, R. Langmuir 1994, 10, 2307. (9) Mougin, K. Ph.D. Thesis, Unviversite´ de Mulhouse, Mulhouse, France, 2001.
Surface Properties of Chemically Modified Surfaces
Langmuir, Vol. 20, No. 7, 2004 2709
Figure 1. Schematic representation of a DD curve. When k < ∂F/∂D, the tip jumps in or off the surface, because of mechanical instabilities (k is the cantilever spring constant, F is the tip-sample interaction force, and D is the tip-sample distance). Table 4. Experimental and Theoretical11 Thicknesses of the Molecular Films (for a Fully Extended Graft Normal to the Surface Plane) As Measured by Ellipsometry end groups
thicknesses of the grafted organic films (Å)
theoretical thicknesses (Å)
CH3 COOR NH2
21 5 10
22.5 6 15
was grafted in these cases. These thicknesses were determined by previously measuring the thickness of the oxide layer formed during the 30-min piranha treatment. A 3-nm reproducible value was obtained for this oxide layer thickness. Even though important effort was made to characterize the grafted surfaces (XPS, ellipsometry, wetting, AFM images), one can suspect that the grafting density and the morphology of the monolayer may affect the comparison of these surfaces in terms of the quantitative adhesion force. To prevent such effects, evolution of the properties with the grafting time of the monolayers was investigated in previous studies.10 In the present study, the grafting time has been chosen to get stable layers versus time. C. Adhesion Force Measurements with the AFM. C.1. Deflection-Distance (DD) Curves Characteristics. Force measurements with the AFM, in the contact mode, consist of detecting the deflection of the spring (cantilever) bearing a tip at its end while interacting with the sample surface.12 The cantilever deflection is detected with an optical device consisting of a laser and four quadrant photodiodes (or photodetector). Provided that the normal spring constant of the cantilever is known, an interacting force can be calculated using Hooke’s law. At the same time, the tip is moved vertically forward and backward with a piezoelectric ceramic. Thus, it is possible to obtain DD curves and then force-distance curves. The DD curves were performed in the air with an available commercial apparatus (Nanoscope IIIa D3000, DI). A schematic representation of a DD curve obtained when probing a hard surface is reported in Figure 1. One should notice that jump-in observed during the approach is usually due to mechanical instabilities and is not considered in this paper. C.2. AFM Calibration. C.2.1. Determination of the Spring Constant of the Cantilever. One of the main difficulties in obtaining quantitative measurements is related to the determination of the spring constant of the cantilever. There are numerous papers describing different methods to get that value.13-15 In our case, we focused on two nondestructive methods, (10) Elzein, T.; Brogly, M.; Schultz, J. Surf. Interface Anal. 2003, 35 (2), 231-236. (11) Porter, M. D.; Bright, T. B.; Alara, D. L.; Chidsey, C. E. D. J. Am. Chem. Soc. 1987, 109, 3559-3568. (12) Cappella, B.; Dietler, G. Surf. Sci. Rep. 1999, 34, 1-104. (13) Cleveland, J. P.; Manne, S.; Bocek, D.; Hansma, P. K. Rev. Sci. Instrum. 1993, 64, 403-405. (14) Hutter, J. L.; Bechhoefer, J. Rev. Sci. Instrum. 1993, 64, 18681873.
Figure 2. Nonlinearity of the four quadrant photodiodes. Variation of the indentation slope is the difference between the experimental slope and a slope of 1, which is expected on a rigid surface. An error of about 60% occurs on the slope, when the detected photodiode tension is equal to (8 V. which have been rigorously studied and compared: the first one is related to the use of rectangular reference cantilevers16 (calibrated with the resonant frequency method and supplied by NanoMetrology) and the second is based on the thermal fluctuations measurement of the cantilever.17 Three cantilevers have been specially selected to give the same force on a reference silicon wafer. Then both calibration methods have been used and compared to calibrate the three cantilevers used in this study. These were triangular-shaped cantilevers (supplied by Nanosensor, Germany) and had an effective 0.30 ( 0.03 N‚m-1 spring constant, which is a value far from the one specified by the supplier (0.58 N‚m-1). C.2.2. Nonlinearity of the Four Quadrant Photodiodes. Haugstad and Gladfelter3 first mentioned the nonlinearity of the optical detector, which is the consequence of a nonhomogeneous spreading of the laser spot on the detector. Figure 2 shows the slope of the “cantilever deflection versus piezo displacement” curve during the loading or unloading (zones C or D) of a DD curve (and considering that there is no nonlinearity at the middle of the photo detector) versus the tension (V) measured by the photodetector. From this result, one can notice that the domain of linearity of the detector lies between +3 and -3 V. If nonlinearity is not taken into account, the error on the quantitative results can be significant. All the measurements have been done in the linearity domain. C.2.3. Tip Radius Determination. To deduce the work of adhesion from our experimental values, it was necessary to have a good estimation of the radius of the probe and, thus, of the tip-sample contact area. Villarrubia18 has developed a numerical method to determine tip radius. Nevertheless, for convenience (15) Senden, T. J.; Ducker, W. A. Langmuir 1994, 10, 1003-1004. (16) Torii, A.; Sasaki, M.; Hane, K.; Okuma, S. Meas. Sci. Technol. 1996, 7, 179-184. (17) Levy, R.; Maaloum, M. Nanotechnology Bristol 2002, 13, 3337. (18) Villarrubia, J. S. Surf. Sci. 1994, 321, 287-300.
2710
Langmuir, Vol. 20, No. 7, 2004
Noel et al.
Figure 4. Adhesive forces (line) and parameter related to hydrophilic properties of the surfaces (contact angles with water). Figure 3. DD force curves on SAMs (during pull-off). reasons, we deduced the radius on the basis of the shape revealed on a scanning electron microscope picture taken at the end of the measurements. Another experimental technique of tip radius estimation using nanoindentation experiments on a model polymer (perfectly elastic) has also been used.19 Both methods have approximately given the same tip radius for all the used cantilevers. Our tip radius was estimated to be 50 ( 5 nm. Finally, checking regularly and randomly the adhesion force on a reference silicon wafer allows us to verify contamination of the tip or change of the tip radius during the measurements. C.2.4. Scan Speed of the Cantilever. The actuator used to move the cantilever vertically backward and forward shows a hysteresis and nonlinearity in its vertical displacement. This phenomenon can be studied by reporting the slope of the contact zones (zones C and D) versus the amplitude of the contact zone and the scan speed. During the experiments, the actuator is considered thermally stable. We observed that a discrepancy appears for very low scan speeds (60 nm‚s-1). For higher scan speeds (18 µm‚s-1), the viscosity of the environment could be significant. A speed of about 6 µm‚s-1 is a good compromise for our actuator. C.3. Adhesion Force Measurements. Before monitoring DD curves on SAMs, the actuator and the cantilever were thermally stabilized. The laser spot in contact with the tip was positioned in such a way that tangential forces, due to frictional forces, were minimized.20 Moreover, SAMs were cleaned by immersing them in an ultrasonic bath (in different solvents) to remove organic contaminations. Samples were then dried with nitrogen, just before analysis. Finally, to check the noncontamination of the tip during the experiments, systematic DD curves were performed on a reference substrate (silicon wafer cleaned in ethanol and dried with nitrogen). We carried out these measurements on the basis of the previously described prerequisites so that the force measurements could be compared. The reported results are an average of about 100 DD curves for each sample, to have a good statistical estimation of our values. Before and after each experiment, the quality of the tip is evaluated by recording the DD curve on a reference surface. When the tip is contaminated, a new tip is used and characterized. In that way, we can select tips with about the same radius and the same spring constant to compare the experimental values. We mention that tip contamination occurs little in comparison with the great number of realized DD curves. C.4. Estimation of the Uncertainties on the Physical Values Related to the Experimental Pull-Off Force Measurements. If we consider that the main uncertainties in a pull-off force measurement with AFM are due to the determination of the spring constant and the radius of the tip, we can approximate an error of 10% on the measurement of an adhesion force (due to the uncertainty of the method of calibration of the tip as specified by ref 16), and a relative error of 20% on the determination of the work of adhesion and surface tensions (due to a 10% error on the determination of the tip radius).19 The uncertainty on the tip radius was evaluated on the basis of nanoindentation experiments with such a tip on a reference elastic polymer network. Nevertheless, even if 20% of the error is high, one must (19) Noel, O.; Brogly, M.; Castelein, G.; Schultz, J. Mechanical Properties derived from Nanostructuring Materials; MRS Symposium Proceedings Edition; Materials Research Society: Warrendale, PA, 2003; Vol. 778. (20) Basire, C. Ph.D. Thesis, ESPCI, Paris, France, 1997.
consider this error as a systematic error. Indeed, all measurements (over 100 for a given surface) were performed with the same tip of the same radius. Then experimental results are significant and the contribution of the different grafts can be discriminated (very low standard deviation, cf. Table 3).
III. Results and Discussion A. Measurements of the Surface Properties with the AFM. A.1. Adhesion Force (Fadhesion) Measurements. In Figure 3, we have reported experimental DD curves during pull-off obtained from the different grafted substrates. Figure 3 shows that AFM measurements in our conditions are sensitive to a chemical modification of the layers. When jump-off contact occurs, we measure the corresponding pull-off deflection (Dpull-off). It also shows that pull-off deflection values increase in the following order: Dpull-offSiCF3 < Dpull-offSiCH3 < Dpull-offSi as-received < Dpull-offSi ester < Dpull-offSiNH2 < Dpull-offSiOH. Comparative graphs of the pulloff forces versus a parameter related to the hydrophilic properties of the surfaces (contact angles with water) are reported in Figure 4. As expected, the pull-off deflection and, thus, the adhesion force value increases with the hydrophilicity of the surface (the contact angle with water decreases). Moreover, considering the standard deviation, the measurements can be reproduced. A.2. In Situ Thermodynamic Work of Adhesion (W0). DMT21 contact mechanics theories show that there is a relationship between the adhesion force (or pull-off force) and the thermodynamic work of adhesion (W0). Considering a contact between a rigid sphere (R: tip radius) and an elastic plane, the relationship is (for experiments conducted at fixed loading conditions)
DMT theory: W0 ) -
Fadhesion 2πR
(1)
DMT theory is rather suitable for systems with low adhesion, low deformation, and small tip radii. This supports the idea that the contact of our system is welldescribed with the DMT theory. Regarding Table 5, one can notice that the experimental works of adhesion calculated with the DMT theory (relationship 1) are coherent and significant. Indeed, the usual work of adhesion values lie between 40 and 70 mJ‚m-2 for organic-organic contacts such as between two silinated silica22 and between 40 and 145 mJ‚m-2 for a contact between raw materials such as silica and silinated silica. These thermodynamic works of adhesion W0 have been compared to the surface energies reported in Table 3 (Figure 5). Figure 5 shows that W0 increases linearly (correlation factor is 0.99) with γs. Nevertheless, this tendency depends on if we consider the (21) Derjaguin, B. V.; Muller, V. M.; Toporov, Y. P. J. Colloid Interface Sci. 1975, 53, 314. (22) Papirer, E.; Balard, H.; Sidqi, M. J. Colloid Interface Sci. 1993, 159, 238-242.
Surface Properties of Chemically Modified Surfaces
Langmuir, Vol. 20, No. 7, 2004 2711 Table 6. Experimental Adhesion Forces,a Theoretical Capillary Forcesb, VDW Forces,c and W0 Due Only to the VDW Contribution SAMs
experimental Fadhesion (nN)
theoretical Fcap (nN)
FVDW (nN)
W0 (mJ‚m-2)
SiNH2 SiOH
33.3 44.6
17 26.6
16.3 18
52 60
a Reported from Figure 4. b Calculated from eq 3. c Calculated from eq 2.
Figure 5. Correlation between normalized W0 and the surface hydrophilicity. Table 5. Thermodynamic Works of Adhesion (W0) Determined from DMT Theory (R ) 50 nm and k ) 0.30 N‚m) surfaces
W0 (mJ‚m-2) with DMT
surfaces
W0 (mJ‚m-2) with DMT
SiCF3 SiCH3 Siwafer
36 ( 13 37 ( 14 45 ( 15
SiCOOR SiNH2 SiOH
48 ( 18 106 ( 40 142 ( 56
Figure 6. Influence of the capillary forces on the thermodynamic works of adhesion for highly hydrophilic surfaces.
SiNH2 and SiOH points. Contact angle measurements with water (Table 1) also show that these surfaces are highly hydrophilic compared to others. Thus, a capillary bridge is certainly formed between the SiNH2 and SiOH surfaces and the tip, and the pull-off mechanism could be different from the other surfaces. This could explain why the corresponding points for SiNH2 and SiOH (in Figure 5) do not fit the linear extrapolation. A.3. Discussion on the Capillary Force (Fcap) Influence in the AFM measurements. A.3.1 General Introduction. When performing AFM experiments in the air, force measurements include the contribution of VDW and capillary forces. In particular, the total adhesion force is given by the following expression:
Fadhesion ) Fcap + FVDW
(2)
The capillary force depends on a meniscus formed between the two surfaces and is given by23
Fcap ) 2πRγ[cos(θs) + cos(θt)]
(3)
where γ is the surface energy of the liquid (0.072 J‚m-2 for water), and θs and θt are the respective contact angles (with water) of the substrate and the tip. According to expression 2, capillary forces could be predominant in the adhesion force measurement. However, Fisher and Israelachvili24 showed that if
γl < γs cos(θ)
(4)
(γl and γs are the liquid and solid surface energies and θ is the contact angle with the liquid) then adhesion is dominated by the solid-solid interactions rather than by the capillary forces. Concerning all our systems, the criterion of Fisher and Israelachvili24 is verified (except with amine and hydroxyl functional group grafting). A.3.4. Thermodynamic Work of Adhesion (W0) Due to VDW Contributions for the Highly Hydrophilic Systems SiNH2 and SiOH. For highly hydrophilic grafted SAMs, an asymmetric capillary bridge is formed. From expression 2, it is possible to determine the VDW forces (FVDW) knowing the capillary forces (given by expression 3) and (23) Riedo, E.; Levy, F.; Brune, H. Phys. Rev. Lett. 2002, 88, 185505. (24) Fisher, L. R.; Israelachvili, J. N. Colloids Surf. 1981, 3, 303319.
the experimental measured adhesion force (Figure 4). By applying the DMT theory, we can deduce the only contribution of the VDW forces to the thermodynamic work of adhesion:
W0 ) -
FVDW 2πR
(5)
W0 values for the SiNH2 and SiOH surfaces have been reported in Table 6. Regarding the usual values of thermodynamic work of adhesion for organic-organic contacts, the calculated values of Table 6 are still relevant and are congruent with the literature.25,26 Finally, we have reported the W0 due only to the VDW contributions versus γs (Figure 6). This Figure shows that highly hydrophilic surfaces behave thermodynamically like the other surfaces provided that only intermolecular force contributions are considered in the thermodynamic work of adhesion determination. Moreover, this study seems to confirm that W0 increases linearly with the hydrophilicity of the surfaces. A.4. In Situ Surface Energy Determination of a Silicon As-Received Wafer. The surface free energy γ is related to the work of adhesion W0 (no capillary forces) by the following expression:
W0 ) γx + γtip - γx-tip
(6)
where γx and γtip are the surface free energies of the sample and of the tip. γx-tip is the interfacial tension at the tipsample interface. In this study, W0 is the thermodynamic work of adhesion calculated from the DMT theory and experimental values (Table 5). If both of the material’s (Sias-received and Si3N4) contact can be roughly said to be similar (in that case γx-tip ) 0 J‚m-2), the expression of W0 is reduced to
W0 ) 2γSias-received
(7)
In that way, knowing W0 (Table 5), we have an estimation of the surface free energy (γSias-received) of our experimental system, which is equal to 23 mJ‚m-2 for Sias-received. The (25) Siebold, A.; Walliser, A.; Nardin, M.; Oppliger, M.; Schultz, J. J. Colloid Interface Sci. 1997, 186, 60-70. (26) Rogier, B.; Nanse, G.; Nardin, M.; Baud, G.; Jacquet, M.; Schultz, J. J. Adhes. Sci. Technol. 2000, 14, 339-350.
2712
Langmuir, Vol. 20, No. 7, 2004
value obtained is in good agreement with the theoretical one (γSias-received,calculated ) 25 mJ‚m-2). A future study on chemically modified SAMs should involve the synthesis of a similar chemically modified grafted tip, to measure the exact values of the surface free energy of the other systems. In that way, the γx-tip term of expression 6 would be roughly equal to 0 J‚m-2 and would give directly the surface free energy of the substrate. IV. Conclusion Adhesion forces, on different chemically modified SAMs, have been measured by using the AFM in contact mode in the air. A qualitative study has established that there is a correlation between the adhesion force and the wetting properties of the surfaces. A rigorous and systematic calibration protocol was established and done systematically before and after each experiment. Experiments conducted in this way provide reasonable quantitative force measurements that can be discussed. From the DMT contact mechanics theory, we could deduce directly the experimental work of adhesion (W0) between the tip and the substrate. Facing such assumptions, we have proposed
Noel et al.
a calculation to separate the VDW contribution from the capillary force contribution in the experimental adhesion force. Then, the thermodynamic work of adhesion (W0,Sias-received) and the surface free energy (γSias-received) of silicon as received, by considering only the VDW contribution (W0,Sias-received ) 45 mJ‚m-2 and γSias-received ) 23 mJ‚m-2), were calculated. A comparison between the experimental and the theoretical values that have been obtained seems to confirm that probing the surface with the AFM is a valuable technique for determining quantitative chemical properties. Finally, another result of this study is that W0 seems to increase linearly with the hydrophilicity of the surfaces. Acknowledgment. We are grateful to Hamidou Haidara, Karine Mougin, and Laurent Vonna for sharing their knowledge concerning SAMs synthesis, to Raphael Le´vy and Mounir Maaloum for discussions about spring constant calibration, and to Christian Fre´tigny for general discussions. LA034884M
