Cleaning and Hydrophilization of Atomic Force ... - ACS Publications

Role of surface morphology on desorption kinetics of water molecules from uncoated silicon microcantilever. M. Raghuramaiah , K. Prabakar , S. Tripura...
0 downloads 0 Views 200KB Size
Download PDF
J. Phys. Chem. B 2006, 110, 25975-25981

25975

Cleaning and Hydrophilization of Atomic Force Microscopy Silicon Probes L. Sirghi,* O. Kylia´ n, D. Gilliland, G. Ceccone, and F. Rossi European Commission, Institute for Health and Consumer Protection, TP-203, BMS, Via E. Fermi, 21020 Ispra (VA), Italy ReceiVed: May 30, 2006; In Final Form: October 9, 2006

The silicon surface of commercial atomic force microscopy (AFM) probes loses its hydrophilicity by adsorption of airborne and package-released hydrophobic organic contaminants. Cleaning of the probes by acid piranha solution or discharge plasma removes the contaminants and renders very hydrophilic probe surfaces. Timeof-flight secondary-ion mass spectroscopy and X-ray photoelectron spectroscopy investigations showed that the native silicon oxide films on the AFM probe surfaces are completely covered by organic contaminants for the as-received AFM probes, while the cleaning methods effectively remove much of the hydrocarbons and silicon oils to reveal the underlying oxidized silicon of the probes. Cleaning procedures drastically affect the results of adhesive force measurements in water and air. Thus, cleaning of silicon surfaces of the AFM probe and sample cancelled the adhesive force in deionized water. The significant adhesive force values observed before cleaning can be attributed to formation of a bridge of hydrophobic material at the AFM tip-sample contact in water. On the other hand, cleaning of the AFM tip and sample surfaces results in a significant increase of the adhesive force in air. The presence of water soluble contaminants at the tipsample contact lowers the capillary pressure in the water bridge formed by capillary condensation at the AFM tip-sample contact, and this consequently lowers the adhesive force.

Introduction Atomic force microscopy (AFM) measurements of surface forces are strongly affected by the chemistry of the AFM tip and sample surfaces.1,2 The chemistry of the AFM probe surfaces can be controlled by several surface functionalization techniques including coating with thiols, silanes, polymer layers, or attachment of carbon nanotubes or colloidal particles.1,3-6 However, because of the surface adsorption of airborne and package-released organic contaminants,7 the surface chemistry of most of the as-received commercial AFM probes remains unknown. Cleaning of the AFM probe surfaces is the simplest probe treatment that can be done to control the probe surface chemistry. Various cleaning methods for silicon and silicon nitride AFM probes are currently used in different laboratories for removing the thin films of the contaminant organic molecules adsorbed on the probe surfaces8-11 or for removing contaminant particles attached to the tips of the AFM probes.12,13 Contamination through adhesion of contaminant nanoparticles to the tips of the AFM probes occurs mainly during scanning of dirty AFM sample surfaces, and its effect on the force measurements and topography images is easily observable.13 In contrast, it is difficult to ascertain the contamination resulting from adsorption of airborne and package-released organic contaminants. Usually, contaminations of the AFM tip that occur during the measurements are revealed by significant variations in the tip-sample adhesive force.14 Because of the small size of the AFM tips, direct characterization of the tip surface chemistry is difficult and in most of the cases the quality of the surface cleaning is probed indirectly by the reproducibility of the force measurements. Bonaccurso and Gillies15 showed that it is possible to reveal contamination on AFM cantilevers by measuring the * To whom correspondence should be addressed. Tel: 39-332786561. Fax: +39-332785787. E-mail: [email protected] or [email protected].

contact angle of sessile water microdrops on their surface or by tip-air bubble force measurements in water. The effectiveness of different cleaning methods used for removing of organic contaminants adsorbed on silicon and silicon oxide surfaces was ascertained by time-of-flight-secondary ion mass spectroscopy (TOF-SIMS) and X-ray photoelectron spectroscopy (XPS),7 scanning auger microscopy,12 Fourier-transformed infrared attenuated total reflectance,16 and in situ spectroscopic ellipsometry.17 It is worthwhile to note that however good the cleaning of the AFM probe and sample surfaces is, the contamination by absorption of airborne organic molecules during the AFM measurements in air is inevitable. Contamination through organic molecule adsorption can easily occur during the AFM measurements in liquid as well. Although it is unanimously recognized that the presence of contaminant molecules on the AFM probe and sample surfaces can greatly affect the measurements of AFM tip-sample adhesion and friction forces,14,18 the effect of contaminants on these forces is not well-understood. The present work describes the effect of surface cleaning on the hydrophilicity of commercial AFM silicon probes and on the adhesive force measurements in water and air. Two cleaning methods, by acid piranha solution and by cathode plasma of a glow discharge in air at low pressure, are analyzed. A method to screen out the discharge electric field in the volume around the AFM probe during its cleaning in a classical plasma sputtering cleaner is described. Results of TOF-SIMS and XPS investigations of the AFM probe surfaces before and after cleaning are described and discussed. Water contact measurements for the as-supplied and cleaned AFM probes are done on sessile droplets of deionized water deposited on the cantilever base of the AFM probes. Results of AFM adhesive force measurements performed in water and in air with as-received and cleaned tips are described and discussed. The adhesive force between contaminated silicon surfaces in

10.1021/jp063327g CCC: $33.50 © 2006 American Chemical Society Published on Web 11/22/2006

25976 J. Phys. Chem. B, Vol. 110, No. 51, 2006

Figure 1. Sketch of the plasma device used for cleaning the AFM probes.

water is explained by the formation of a capillary bridge of hydrophobic material between the AFM tip and the sample surfaces. A similar explanation was used by Kanda et al.19 to explain the very large adhesion force between silica and mica surfaces in an aqueous solution of n-propanol at high concentration. The effect of surface cleaning on AFM measurements of the adhesion force in air can be explained by the reduced influence that contaminant molecules can have on the capillary pressure in the water bridge formed by capillary condensation at the tip-sample contact in air. Materials and Methods Surface Cleaning. Two surface-cleaning methods were used. In the first method, the AFM silicon probes (CSG11 and NSG11 provided by NT-MDT Co., and PPP-BSI provided by Nanosensors) were immersed for 30 min in acid piranha solution, 70% v H2SO4, and 30% v H2O2, at a temperature of 50 °C. In the second method, the AFM silicon probes and silicon wafers were placed on the cathode of a commercial plasma sputtering/ cleaning system (Edwards Sputter Coater S150B) working in the cleaning mode. The cathode was a tantalum disk with diameter of 10 cm. To avoid destruction of the AFM probes, a shield (a rectangular tantalum electrode with length of 3 cm and width of 1 cm, in electrical contact with the cathode and bent over the AFM probe to create a gap of 2 mm with the cathode surface) was used to isolate the AFM probe from the high electric field of the cathode fall. Figure 1 shows a sketch of the plasma-cleaning system used to clean the silicon surfaces of the AFM probes. The shield creates an electric field-free volume where the oxygen radicals produced in the negative glow plasma in air can diffuse in through the open sides. The silicon surface cleaning is attributed to the strong oxidizing power of the oxygen radicals produced in the plasma. The silicon AFM probes were cleaned for 10 min in negative glow plasma produced in air at a pressure of 900 Pa, a discharge voltage of 350 V, and a discharge current intensity of 5 mA. Scanning electron microscopy (SEM) images of the AFM probe tips before and after plasma cleaning showed that the tips of the AFM probes were not damaged as a result of this cleaning procedure. However, it should be mentioned that the surface of the cleaned AFM probes is prone to be contaminated through adhesion of particles collected from ambient air or during scanning of dust-contaminated sample surfaces (see the SEM image in the Supporting Information, which shows a dust particle attached on the surface of a plasma-cleaned AFM tip). The cleaning makes the surface of the AFM probes very hydrophilic, and this increases the adhesion force between the dust particles and the probe surface. The silicon AFM samples [p-Si(100) wafers, resistivity > 100 Ω cm, provided by MaTecK Material-Technologie & Kristalle GmbH] were ultrasonically cleaned in acetone (5 min), ethanol (5 min), and deionized water (5 min) and dried in highpurity nitrogen gas flow. For adhesive force measurements in deionized water, the silicon samples were further cleaned by plasma to improve their surface hydrophilicity. The water

Sirghi et al. contact angle of the AFM silicon samples after cleaning in acetone, ethanol, and deionized water was 52 ( 2°. After plasma cleaning, the water contact angle of the silicon samples decreased to about 4°. The damage caused to the reflex coatings of the AFM probes by the cleaning methods has been evaluated by measurements of the total intensity of the laser light beam reflected by the as-supplied and cleaned AFM probes. For the plasma-cleaning method, the reflectivity of the aluminum reflex coating (on the AFM probes PPP-BSI provided by Nanosensors) decreased to a low as 50% of the initial values, while the reflectivity of the gold reflex coating (on the AFM probes NSG11 and CSG11 provided by NT-MDT Co.) decreased to no less than 90% of the starting values. The cleaning by piranha solution did not result in a significant decrease of the reflectivity of the AFM probe reflex coatings. Contact Angle Measurements. The water contact angle of the as-supplied and cleaned AFM probes was measured with a goniometer (Digidrop from GBX Instruments Co.) for very small (volume of about 0.1 µL) sessile droplets of deionized water deposited on the base of the AFM probes. For each surface, the contact angle measurements were repeated five times and the average value was taken into account. TOF-SIMS and XPS Measurements. The chemical analysis of the probe surfaces was conducted in a combined XPS and TOF-SIMS surface analysis facility. This system is comprised of an AXIS ULTRA XPS Spectrometer (Kratos Analytical, United Kingdom) and an ION-TOF (IV) TOF-SIMS system linked together through an ultrahigh vacuum (UHV) sample exchange chamber. This system permits the UHV transfer of samples between the two instruments; thus, it was possible to analyze the probes by XPS and TOF-SIMS without exposing the samples to atmosphere contamination. The XPS instrument calibration was performed using a clean pure Au/Cu sample and pure Ag sample (99.99%). Measured values for electron binding energies were 83.98 ( 0.05 and 932.67 ( 0.05 eV. Samples were irradiated with monochromatic Al KR X-rays (hν ) 1486.6 eV) using an X-ray spot size of 100 µm2 and a take-off angle of 90° with respect to the sample surface. An electron flood gun was used to compensate for surface charging, and all spectra were corrected by setting the C1s hydrocarbon peak to 285.00 eV. For each sample, a survey spectrum (0-1150 eV), from which the surface chemical compositions (atom %) were determined, was recorded at a pass energy of 160 eV. In addition, one set of high-resolution spectra (PE ) 20 eV) was also recorded on each sample. The total acquisition time was kept below 10 min to avoid any possible X-ray-induced damage. The data were processed using the Vision2 software (Kratos) and a commercial software package (Casa XPS v2.3.10). Sample compositions were obtained from the survey spectra after linear background subtraction and using the relative sensitivity factors included in the software derived from Scofield cross-sections. This method is estimated to give an accuracy of 10% in the measurement of elemental compositions. Curve fitting of C1s peaks was carried out using the same initial parameters and interpeak constraints to reduce scattering. The C1s envelope was fitted with the Gaussian-Lorentian function (G/L ) 30) and variable full width half-maximum. The TOF-SIMS analysis was conducted using an ION-TOF (IV) TOF-SIMS system equipped with a 25 keV liquid metal ion source (LMIS) operating with either Bi+ (spectra shown in Figure 2) or Ga+ (spectra shown Figure 3) ions. Analyses were obtained from square areas of 250 × 250 µm2 in high mass resolution burst mode (resolution M/∆M > 6000). The total ion

Atomic Force Microscopy Silicon Probes

Figure 2. Positive ion TOF-SIMS (Bi+ primary ion gun) spectra of silicon surfaces of the as-supplied (A), piranha-cleaned (B), and plasmacleaned (C) AFM probes provided by NT-MDT Co. (CSG11).

beam dose was limited to less than 2 × 1012 ions/cm-2 and thus was within the static SIMS regime. AFM Measurements. All of the AFM measurements were performed by a commercial atomic force microscope, Solver P47H from NT-MDT Co., Russia, with the high-resolution scanner SC 103. For the measurements in deionized water, we used the open liquid cell unit AU028. The water cell was built by catching a big drop of MilliQ water (resistance of 17 MΩ cm-1 and pH 5.7) between the hydrophilic surfaces of the silicon sample and the glass holder of the AFM probe. For the measurements performed in air, the humidity and temperature of the ambient air were monitored by a high precision hygrometer (HM 34 from Vaisala, Finland). The force-scanner extension curves in water and in air were acquired at a speed of one force curve per second on an array of 100 points distributed over an area of 1 µm × 1 µm on the sample surface. The spatial resolution of scanner extension distance was 0.1 nm. The acquired force-scanner extension curves were processed by homemade software to compute the tip-sample adhesive force and to determine the dependence of tip-sample interaction force vs the tip-sample distance. Results Organic Contaminants on Commercial AFM Probes. Surfaces of the as-received and cleaned AFM silicon probes

J. Phys. Chem. B, Vol. 110, No. 51, 2006 25977 provided by two manufacturers, NT-MDT Co. and Nanosensors were characterized by TOF-SIMS and XPS. In the case of probes subjected to either of the cleaning procedures, the treated samples were placed under high vacuum conditions within 15 min of the completion of cleaning. The positive ion spectrum of the as-received probe supplied by NT-MDT Co. (Figure 2A) indicates the presence of molecular fragments (CH3Si+ and C3H9Si+) that are characteristic of surface contamination by poly(dimethylsiloxane) (PDMS). The same results were found by Lo et al.7 for the as-received AFM probes supplied by Park Scientific Instruments. The positive ion spectra (Bi+ primary ions) of piranha-cleaned AFM probe (Figure 2B) and plasmacleaned AFM probe (Figure 2C) show that both cleaning methods successfully removed the PDMS contaminants. The residual contaminants on the cleaned probe are composed of a mainly hydrocarbon material (CxHy) but with the underlying probe now being detectable as shown by the higher silicon ion peak together with the appearance of boron in the spectra. For the as-received AFM probes supplied by Nanosensors, the main contaminant materials identified on the probe surface were airborne hydrocarbons with little or no trace of PDMS. Figure 3 shows the TOF-SIMS positive ion spectra (Ga+ primary ions) of an as-received AFM probe surface supplied by Nanosensors (part A) as compared with the spectra of the same probe cleaned by either plasma (part B) or acid piranha solution (part C). It should be noted that the TOF-SIMS spectra shown in Figure 2 were obtained using a Bi+ LMIS while those shown in Figure 3 were measured using a Ga+ LMIS. While this experimental difference does not in any way influence the identification of contaminate materials, the spectra obtained using the different systems should not be compared quantitatively due to the significant differences in secondary molecular ion yields obtained using the two different sources. In particular, the Bi+ source has a higher ion yield for larger (organic) molecular fragments and as such shows greater sensitivity to the presence of organic contaminants than the Ga+ LMIS. XPS analysis of the Nanosensors probes shows principally the presence of C, O, and Si with the normalized atomic concentrations being reported in Table 1 (see also the XPS core level spectra of Si and C in Supporting Information). These values, obtained from an analysis depth of around 10 nm, comprise a signal from the contaminant layer, the underlying native oxide on the silicon, and some contribution from the bulk

Figure 3. Positive ion TOF-SIMS (Ga+ primary ion gun) spectra of silicon surfaces of the as-supplied (A), piranha-cleaned (B), and plasmacleaned (C) AFM probes provided by Nanosensors (PPP-BSI). (D) The spectrum for the piranha-cleaned AFM probe after storage in ambient air for 6 days.

25978 J. Phys. Chem. B, Vol. 110, No. 51, 2006

Sirghi et al.

TABLE 1: Elemental Composition of the AFM Probe Surfaces (Average of Three Measurements; Standard Deviation in Brackets) sample

C (atom %)

O (atom %)

Si (atom %)

as-received air plasma piranha solution

58.4 (8.8) 19.9 (6.2) 28 (5.8)

13.06 (1.7) 32.8 (4.9) 23.8 (2.5)

28 (7.5) 47.24 (3.7) 48 (6.3)

silicon of the probe. Using the more surface sensitive TOFSIMS method (depth of analysis ) 1-2 nm), the mass spectra show mainly hydrocarbon peaks (CxHy) with only very weak silicon-containing peaks and no trace of boron from the substrate. From this, we can deduce that the contaminant layer is mainly hydrocarbon in nature and probably completely covers the underlying probe with a layer thickness that is near to the analysis depth of the TOF-SIMS technique. By deconvolution of the C1s core level XPS spectra, we observe two components C0 (285 eV) and C1 (285.5 eV), which can be attributed to hydrocarbon bonding (C-C/C-H) and C-O single bonding, respectively. The high proportion of the C-H component (98%) confirms the dominance of hydrocarbon contamination observed in the TOF-SIMS spectra. Following the cleaning procedures, a significant reduction of the carbon content is observed with corresponding increases in the silicon and oxygen concentrations. In this respect, the plasma cleaning seems to be the most effective treatment for carbon reduction. TOF-SIMS spectra show large decreases in the CxHy peaks, increased Si and SiOx peaks, and the appearance of small-intensity boron peaks from the probe. These peaks are most intense in the plasma-treated sample. The main difference in the mass spectra obtained from the two different cleaning methods was the appearance of a variety of CxHyOz type peaks in the plasma-treated probes while the piranha-treated sample continues to show the presence of mainly a CxHy type contamination. Confirmation of this difference can be seen in the high-resolution XPS spectra of the carbon C1s peak (see the Supporting Information) in which, in the case of plasma treatment, shows a large increase in the C-O component together with the appearance of a third component attributable to carbon doubly bonded to oxygen at about 289 eV. In contrast to this, the piranha-treated surface shows no CdO component and only a minor increase in the C-O. While these two cleaning methods are effective, the removal of the airborne hydrocarbon contaminants from the AFM probe surface is not permanent. Normal storage of the cleaned probes in air results in coverage of the probe surface with a layer of hydrocarbon after only a few days. Positive ion spectra of the piranha-cleaned AFM probe recorded after the probe had been stored in ambient laboratory air for 6 days (Figure 3D) show that the surface is already contaminated with a sufficient amount of the airborne hydrocarbons as to mask almost completely the SiOH+ peaks in the TOF-SIM spectra. Long-time storage of the Nanosensors probes on silicone mat in plastic boxes results in contamination with PDMS. TOF-SIMS analysis (spectrum not shown here) of the probe surface after 6 months of storage revealed a complete coverage of the surface with PDMS contaminant. Hydrophilicity of the AFM Silicon Probes. The hydrophilicity of the AFM silicon probes has been characterized by water contact angle measurements. Because the TOF-SIMS and XPS investigations did not show significant differences in the chemistry of the tip, cantilever, and cantilever base surfaces of the AFM probes, we consider that water contact angle measurements on sessile macroscopic water droplets on the bases of the AFM probes are relevant for the evaluation of the hydro-

Figure 4. Time variation of water contact angle of the AFM probe surfaces during the plasma-cleaning process and storage in air.

philicity of the AFM tip surface. Figure 4 shows the time variation of the water contact angle of the surfaces of the AFM probes (NSG11 provided by NT-MDT Co. and PPP-BSI provided by Nanosensors) during the plasma-cleaning process and then during storage in laboratory air. The water contact angle of the AFM probes decreases quickly as a result of oxidation of hydrophobic contaminants on the probe surfaces. After about 10 min of cleaning, the water contact angle of both AFM probes decreased below 10°. Cleaning of probes in piranha solution also resulted in a rapid decrease of water contact angles to values below 10° (result not shown here). It appeared that the contact angle of plasma-cleaned surface is a little smaller than that of the piranha-cleaned surface. This observation is in agreement with the XPS/TOF-SIMS analyses, which showed slightly less carbon-containing contaminants on the plasmacleaned surfaces. The improvement of the AFM probe surface hydrophilicity is not permanent. As a result of storage in air, the probe surface hydrophilicity worsens due to the absorption of airborne hydrophobic contaminants (hydrocarbons). Thus, the water contact angles of both AFM probes increased to about 70° in 50 days. Measurements performed after longer storage times (results not shown here) showed a much slower increase of probe water contact angle. However, the water contact angle of the probe surface during storage in air did not reach the value of the water contact angle of the as-received AFM probes. This may be attributed to the difference in the contamination of the as-received probes and cleaned and air-exposed probes. For the NT-MDT Co. probes, while the as-received probes were contaminated mainly by silicon oils, the cleaned and air-exposed probe was contaminated mainly by airborne hydrocarbons. Adhesive Force in Water. Typical force-distance curves measured in deionized water before and after plasma cleaning of the AFM probe and sample surface are presented in Figure 5a,b, respectively. The AFM probe used in this experiment was CSG11 from NT-MDT, and it had the nominal cantilever spring constant of 0.1 N/m and the nominal tip curvature radius of 10 nm. The force-distance curves for the as-received AFM silicon probe and the silicon sample showed a snap-in instability on the approach part at a tip-sample distance of about 2 nm and a snap-out instability on the retract part, at a tip-sample distance of about 6 nm. This pattern of the force-distance curves indicates formation of a bridge of contamination material at the tip-sample contact. The snap-in instability on the approach part of the force-distance curve corresponds to coalescence of the layers of contaminants on the tip and sample surfaces. Therefore, the snap-in distance can be considered as a rough indicator of the thickness of contaminant layers on the AFM tip and sample surfaces,20 which in this case is less than 2 nm. While the tip is pushed further toward the sample surface, the contaminant material gathers at the contact and forms a meniscus that surrounds the contact. When the tip is retracted, the bridge of

Atomic Force Microscopy Silicon Probes

J. Phys. Chem. B, Vol. 110, No. 51, 2006 25979

Figure 7. Histograms of the tip-sample adhesive force measured in deionized water before and after cleaning of the AFM probe (CSG11 provided by NT-MDT Co.) and sample surfaces.

of a bridge of different material at the contact raises a meniscus adhesive force that is roughly computed as:21

F ) 4π ‚ σwc ‚ R ‚ cos θ

Figure 5. Typical force-distance curves measured on silicon in deionized water with as-supplied (a) and plasma-cleaned (b) AFM probe (CSG11 provided by NT-MDT), respectively.

Figure 6. Sketches of a bridge of organic contaminants (OC) in water (a) and of a bridge of water with soluble OC in air (b).

contaminants in water is elongated until it breaks up at certain tip-sample separation distance. This behavior of the bridge of contaminants at the tip-sample contact in water is similar to the behavior of the water bridge formed by capillary condensation of the water vapor at the tip-sample contact in ambient air. Figure 6a shows a sketch of the formation of a bridge of hydrophobic contaminants at the tip-sample contact in water, while Figure 6b shows a similar sketch of formation of a water bridge at the tip-sample contact in air. Generally, formation

(1)

where F is the adhesive force, σwc is the surface free energy of the contaminant-water meniscus interface, and R is the equivalent radius of the contact (in this case, because the sample surface is flat, the radius of the AFM tip). Taking into account values of 0.28 nN for F, 10 nm for R, and 0.5 for cos θ, a value of 0.0044 N/m is computed for the σwc, which is much smaller than the surface free energy of water. This small value of σwc is an indication that the organic contaminant molecules are not very hydrophobic. Cleaning of the AFM probe and sample surfaces in plasma resulted in cancellation of any adhesive force between surfaces. Figure 5b shows that there is no long- or short-range force between cleaned silicon surfaces in water. The absence of any electrostatic double layer force on the force-distance curves shows that either contaminated or clean silicon surfaces do not electrically charge in the deionized water. A statistical analysis of the adhesive force values measured for a set of 100 force-distance curves for the as-received AFM probe and sample and for the plasma-cleaned AFM probe and sample led to the histograms presented in Figure 7. A Gaussian distribution with the most probable value at 287 pN and the width of 174 pN was found for the adhesive force values measured with the as-received AFM probe on the silicon sample ultrasonically cleaned in acetone, ethanol, and deionized water. For the plasma-cleaned AFM probe and plasma-cleaned silicon sample, the adhesive force value distribution showed a Gauss distribution with the most probable value at 0 pN and the width of 17.7 pN. There was practically no adhesion force between the plasma-cleaned silicon surfaces, the dispersion of the adhesive force values in this case being caused by the noise in the AFM force detection system. Although the adhesive force between contaminated AFM silicon probe and sample surfaces is small, its value (287 pN) is important for the case of AFM measurements on biological samples, where typical values of forces may be below 100 pN.22 Adhesive Force in Air. Typical force-distance curves measured in ambient air (RH ) 33% and 24 °C), before and after plasma cleaning of a silicon AFM probe (NSG11 from NT-MDT with a cantilever spring constant of 15 N/m and a nominal tip curvature radius of 10 nm) are presented in Figure 8a,b, respectively. Prior to the adhesive force measurements, the surface of the silicon sample was cleaned by acetone, ethanol, and deionized water and dried in pure nitrogen. As a result of plasma cleaning of the AFM probe, the tip-sample adhesive force increased about three times. A statistical analysis of the adhesive force values measured for a set of 100 forcedistance curves for the as-supplied AFM probe and for plasma-

25980 J. Phys. Chem. B, Vol. 110, No. 51, 2006

Figure 8. Typical force-distance curves measured on a silicon sample in air (RH ) 33%, t ) 24 °C) with as-supplied (a) and plasma-cleaned (b) AFM probe (NSG11 provided by NT-MDT), respectively.

Figure 9. Histograms of tip-sample adhesive force measured on a silicon sample in air (RH ) 33%, t ) 24 °C) before and after plasma cleaning of the AFM probe (NSG11 provided by NT-MDT).

cleaned AFM probe led to the histograms presented in Figure 9. A Gaussian distribution with the most probable value of 9.2 nN and the width of 5.3 nN was found for the adhesive force values measured with the as-supplied AFM probe. After plasma cleaning of the AFM probe, the adhesive force values increased showing a Gaussian distribution with the most probable value of 29.4 nN and a width of 5.3 nN. A similar effect of adhesive force increase was observed when the AFM probes were cleaned in acid piranha solution (results not shown here). We observed an increase of the AFM tip-sample adhesive force even in the case in which the AFM probes were cleaned by imbedding them in hot (45 °C) ethanol and then in deionized water (see the adhesive force histogram SI.5 in the Supporting Information). The same behavior of the adhesive force in air was observed for the case of Nanosensors probes. Cleaning of the probe in piranha solution resulted in an increase of about 1.5 times of the adhesive force between the probe tip and a glass sample (see the adhesive force histogram SI.3 in the Supporting Information). Also, cleaning of the probe by plasma resulted in an increase of 1.45 times of the adhesive force between the probe tip and a silicon sample (see the adhesive force histogram SI.4 in the Supporting Information).

Sirghi et al. Because the main component of the AFM tip-sample adhesive force in the ambient air is the capillary force of the liquid bridge formed by capillary condensation of water at the tip-sample contact,23 the increase of the adhesion force in these experiments is attributed to the effect of surface cleaning on the properties of the capillary bridge. The silicon surface of the as-supplied AFM probe is covered by a layer of hydrophobic contaminants and water. When the AFM tip comes in contact with the sample surface, the two layers of adsorbed water and contaminant molecules coalesce and form a bridge. Moreover, the water vapor in the ambient air condenses into the small gap formed by the AFM tip and sample surfaces and contributes to the bridge formation. The moment of bridge formation is registered as snap-in instability on the approach part of the force-distance curve. The tip-sample distance at which the snap-in instability occurs indicates the maximum possible value of the thickness of the adsorbed layers of water and/or contaminant molecules on the AFM tip and sample surfaces (here estimated to about 2 nm on each surface).21 The forcedistance curves in Figure 8 show no noticeable difference in the snap-in distance, a fact that is an indication that our AFM probe was too soft for a precise measurement of water and contaminant layer thickness. The snap-out instability, which is usually observed on the retract part of the force curves, was avoided in these experiments by using an AFM probe with a hard cantilever (large spring constant).24 Thus, during the tip retraction, the AFM tip-sample interaction force decreased monotonically to zero for the either cleaned or as-received AFM probes. The monotonic decrease of the attraction force by the increase of the tip-sample distance is related to the elongation and breakup of the liquid bridge formed by capillary condensation at the tip-sample contact.11 Recently, it was shown that the distance at which the liquid bridge breaks up is related to the volume of the bridge.11,25 Interestingly, although the cleaning of the AFM probe resulted in an important increase of the adhesive force, the bridge breakup distance remained the same (about 10 nm). This means that the volume of the liquid bridge formed at the tip-sample contact remained roughly the same. The increase of the capillary force created by a bridge of pure water in equilibrium with water vapor in air is commonly understood as an increase of the amount of water at the tipsample contact, but it seems that this is not the case in this experiment. The chemical potential equilibrium equation assumed for a concave water meniscus in thermodynamic equilibrium with the undersaturated water vapor indicates a pressure reduction, ∆p ) pv - pl, in the liquid phase,

∆p )

RT ln(pv/pvs) Vm

(2)

where pv is the water vapor pressure, pl is the pressure of water in the liquid bridge, R is the universal gas constant, T is the absolute temperature, Vm is the molar volume of water, and pvs is the water vapor saturation pressure. The ratio pv/pvs expressed as a percentage determines the air relative humidity (RH). The water vapor in ambient air is usually undersaturated (pv < pvs), and the meniscus that constitutes the interface of the liquid bridge with the vapor phase in air is concave, which means that the pressure in the liquid bridge is smaller than the pressure in vapor (∆p ) pv - pl < 0). It is this pressure difference that raises the capillary force of the bridge. If A is the transversal area of the water bridge, the capillary force is

F ) ∆p ‚ A

(3)

Atomic Force Microscopy Silicon Probes

J. Phys. Chem. B, Vol. 110, No. 51, 2006 25981

Because the bridge volume does not change much as result of the AFM probe cleaning, it is reasonable to consider that A is constant and the cleaning-induced increase of F (in absolute value) is caused mainly by the increase of ∆p (in absolute value). This increase of ∆p can be understood if some of the organic contaminant molecules on the AFM probe and sample surfaces are considered soluble in water. Thus, the bridge formed at the contact of the contaminated AFM tip and sample surfaces can be considered as formed by an aqueous solution of organic contaminants (see the sketch in Figure 6b). In accordance with Raoult’s low, the solute contaminants lower the saturation pressure of water vapor to pvs′ < pvs, which means that they lower (in absolute value) ∆p to

∆p′ )

RT ln(pv/pvs′) Vm

(4)

Another way to understand the decrease in the capillary pressure is to consider that the solute organic contaminant molecules create an osmotic pressure, posm, in the liquid bridge that diminished the capillary pressure of the bridge

∆p′ ) ∆p - posm

(5)

The presence of water soluble contaminants at the AFM tipsample contact lowers the capillary pressure of the liquid bridge meniscus and the capillary component of the tip-sample adhesion force. Correspondingly, the removal of the organic contaminants from the AFM tip surface results in an increase of ∆p and the capillary force. Conclusion The surfaces of AFM silicon probes provided by NT-MDT Co. (CSG11 and NSG11) and Nanosensors (PPP-BSI) before and after cleaning were investigated by TOF-SIMS, XPS, and water contact angle measurements. The AFM probes were cleaned either in acid piranha solution or cathode plasma of a glow discharge in air at low pressure. For the plasma cleaning, a simple method to isolate the AFM probes from the high electric field of the cathode fall was described. It was shown that both cleaning methods successfully remove most of the organic contaminants from the surfaces of the AFM probes. Water contact angle measurements showed that the surface cleaning results in a drastic increase of AFM probe surface hydrophilicity. However, the hydrophilicity of the cleaned probes decreases within a few days due to the adsorptions of airborne contaminants. Cleaning of silicon surfaces of the AFM probes and samples results in important changes in the tipsample adhesive force in deionized water and air. While cleaning reduced the adhesive force between silicon surfaces in deionized water to zero, it drastically increased the adhesive force in air. The adhesive force between contaminated silicon surfaces in water may be explained by formation of a capillary bridge of hydrophobic contaminants between tip and sample surfaces. The cleaning operation removed the layer of organic contaminants adsorbed on the silicon surfaces of the AFM tip and samples and results in a zero adhesive force between cleaned silicon surfaces in deionized water. On the other hand, the cleaning operation increases the hydrophilicity of the AFM tip and sample surfaces and the tip-sample adhesive force in air. To explain this effect, we considered the effect of water soluble contaminant

molecules on the capillary pressure in the liquid bridge formed through capillary condensation at AFM tip-sample contact in ambient air. Thus, it is considered that the organic contaminants create an osmotic pressure inside the water bridge, which lowers the capillary pressure and, consequently, the capillary force. Supporting Information Available: C1s core XPS level spectra of the as-received (a) after air plasma cleaning (b) and after piranha solution cleaning AFM probe (PPP-BSI from Nanosensors); Si2p XPS core level spectra of the as-received (a) after air plasma cleaning (b) and after piranha solution cleaning AFM probe (PPP-BSI from Nanosensors); value distribution probabilities of the tip-sample adhesive force in air (RH ) 33% and t ) 24 °C) for the as-received and piranha solution cleaned AFM probe (PPP-BSI supplied by Nanosensors) and glass sample (Corning Glass Works); value distribution probabilities of the tip-sample adhesive force in air (RH ) 40% and t ) 26 °C) for the as-received and piranha solution cleaned AFM probe (PPP-BSI supplied by Nanosensors) and silicon sample; value distribution probabilities of the adhesive force in air (RH ) 30% and t ) 21 °C) for the as-received and ethanol-cleaned AFM probe (CSG11 from NT-MDT Co.) and silicon sample; SEM image showing a dust particle attached to the very hydrophilic surface of the tip of a plasma-cleaned AFM probe (CSG 11 from NT-MDT Co.), taken 24 h after the probe cleaning. This material is available free of charge via the Internet at http://pubs.acs.org. References and Notes (1) Frisbie, C. D.; Rozsnyai, L. F.; Nay, A.; Wrighton, M. S.; Lieber, C. M. Science 1994, 265, 2071. (2) Janshoff, A.; Neitzert, M.; Oberdorfer, Y.; Fuchs, H. Angew. Chem., Int. Ed. 2000, 39, 3213. (3) Luginbuhl, R.; Szuchmacher, A.; Garrison, M. D.; Lhoest, J.-B.; Overney, R. M.; Ratner, B. D. Ultramicroscopy 2000, 82, 171. (4) Cheung, C. L.; Hafner, J. H.; Lieber, C. M. PNAS 2000, 97, 3809. (5) Knapp, H. F.; Stemmer, A. Surf. Interface Anal. 1999, 27, 324. (6) Vakarelski, I. U.; Higashitani, K. Langmuir 2006, 22, 2931. (7) Lo, Y.-S.; Huefner, N. D.; Chan, W. S.; Dryden, P.; Hagenhoff, B.; Beebe, T. P., Jr. Langmuir 1999, 15, 6522. (8) Nakagawa, T.; Soga, M. Jpn. J. Appl. Phys. 1997, 36, 5226. (9) Senden, T. J.; Drummond, C. J. Colloids Surf. A 1995, 94, 29. (10) Tomitori, M.; Arai, T. Appl. Surf. Sci. 1999, 140, 432. (11) Sirghi, L.; Szoszkiewicz, R.; Riedo, E. Langmuir 2006, 22, 1093. (12) Costa, C. A. R.; Radovanovic, E.; Teixeira-Neto, E.; Goncalves, M. C.; Galembeck, F. Acta. Microscopica 2003, 12, 71. (13) Nie, H.-Y.; McIntyre, N. S. Langmuir 2001, 17, 432. (14) Thundat, T.; Zheng, X.-Y.; Chen, G. Y.; Sharp, S. L.; Warmack, R. J. Appl. Phys. Lett. 1993, 63, 2150. (15) Bonaccurso, E.; Gillies, G. Langmuir 2004, 20, 11824. (16) Choi, K.; Ghosh, S.; Lim, J.; Lee, C. M. Appl. Surf. Sci. 2003, 206, 355. (17) Foster, Ch.; Schnabel, F.; Weih, P.; Stauden, Th.; Ambacher, O.; Pezoldt, J. Thin Solid Films 2004, 455-456, 695. (18) Enachescu, M.; Carpick, R. W.; Ogletree, D. F.; Salmeron, M. J. Appl. Phys. 2004, 95, 7694. (19) Kanda, Y.; Nakamura, T.; Higashitani, K. Colloid Surf., A 1998, 139, 55. (20) Mate, C. M.; Lorenz, M. R.; Novotny, V. J. J. Chem. Phys. 1989, 90, 7550. (21) Israelachvili, J. N. Intermolecular and Surface Forces, 2nd ed.; Academic Press: New York, 1992. (22) Lavery, R.; Lebrune, A.; Allemand, J.-F.; Bensimon, D.; Croquette, V. J. Phys.: Condens. Matter 2002, 14, R383. (23) Xiao, X.; Qian, L. Langmuir 2000, 16, 8153. (24) Wiesendanger, R. Scanning Probe Microscopy and Spectroscopy. Methods and Applications; Cambridge University Press: Cambridge, 1994; p 235. (25) Maeda, N.; Israelachvili, J. N.; Kohonen, M. M. PNAS 2003, 100, 803.