You are on page 1of 5

Langmuir 2008, 24, 1577-1581 1577

Manipulation of Gold Nanoparticles: Influence of Surface Chemistry,
Temperature, and Environment (Vacuum versus Ambient
Atmosphere)†
K. Mougin,*,‡ E. Gnecco,§ A. Rao,§ M. T. Cuberes,| S. Jayaraman,⊥ E. W. McFarland,#
H. Haidara,‡ and E. Meyer§
ICSI-CNRS, 15 Rue Jean Starcky, 68057 Mulhouse, France,
Institute of Physics, UniVersity of Basel and NCCR “Nanoscale Science”,
Klingelbergstrasse 82, 4056 Basel, Switzerland,
Laboratorio de Nanotécnicas, UniVersidad de Castilla-La Mancha,
Plaza Manuel Meca 1, 13400 Almadén, Spain,
Corning Incorporated, 1 Science Center Road, Corning, New York 14831,
and Chemical Engineering Department, UniVersity of CaliforniasSanta Barbara,
Santa Barbara, California 93106-5080

ReceiVed September 20, 2007. In Final Form: October 24, 2007

We have manipulated raw and functionalized gold nanoparticles (with a mean diameter of 25 nm) on silicon
substrates with dynamic atomic force microscopy (AFM). Under ambient conditions, the particles stick to silicon until
a critical amplitude is reached by the oscillations of the probing tip. Beyond that threshold, the particles start to follow
different directions, depending on their geometry and adhesion to the substrate. Higher and lower mobility were
observed when the gold particles were coated with methyl- and hydroxyl-terminated thiol groups, respectively, which
suggests that the adhesion of the particles to the substrate is strongly reduced by the presence of hydrophobic interfaces.
Under ultrahigh vacuum conditions, where the water layer is absent, the particles did not move, even when operating
the atomic force microscope in contact mode. We have also investigated the influence of the temperature (up to 150
°C) and of the geometrical arrangement of the particles on the manipulation process. Whereas thermal activation has
an important effect in enhancing the mobility of the particles, we did not find differences when manipulating ordered
versus random distributions of particles.

1. Introduction A compromise between the two techniques is tapping mode AFM.
Nanomanipulation is a complex problem, where mechanical Here, the phase shift of the cantilever oscillations with respect
and chemical properties of substrates, probing tools, and nano- to the external periodic excitation can be used to estimate the
objects (“particles”) are combined and different results are dissipated energy. This method was recently used by Ritter and
expected depending on the environmental conditions. Despite co-workers to manipulate antimony particles on a graphite surface
the fundamental and practical interest of this problem, systematic in air.3
investigations on nanomanipulation are still scarce. A reason for Gold particles are very attractive for different reasons. For
that is the difficulties in quantifying the dynamical processes instance, they are ideal electrodes for molecular electronics.4
occurring while manipulating, that is, collisions between probing Gold clusters below 5 nm in size deposited onto thin metal oxides
tips and particles, friction between particles and substrates, also exhibit unexpected highly catalytic activity (not obtained
electrostatic interactions among all of them, and so forth. The with bulk metal) for different reactions, from combustion to
most accurate manipulation studies of nanoparticles have been hydrogenation, reduction, and so forth.5,6 Coated with organic
performed by scanning tunneling microscopy (STM). In a pioneer molecules, gold nanoparticles can be used for DNA assays in
experiment, one of us moved single C60 molecules along the genomics,7,8 as signal amplifiers for biological recognition or for
steps of a Cu(111) surface using a scanning tunneling microscope quantitation of tags in biological assays. To utilize and optimize
in ultrahigh vacuum (UHV).1 Unfortunately, despite the accurate the chemical/physical properties of nanosized gold particles, a
level of control obtained with STM, the energy dissipated in the large spectrum of research has been done on the control of the
manipulation process cannot be estimated with this technique. size,9,10 shape,11-14 surface chemistry,15-18 and aggregation
Manipulation of large C60 islands on NaCl was performed by (3) Ritter, C.; Heyde, M.; Stegemann, B.; Rademann, K.; Schwarz, U. D.
Lüthi et al. using contact atomic force microscopy (AFM).2 Even Phys. ReV. B 2005, 71, 085405.
(4) Adams, D. M.; Brus, L.; Chidsey, C. D.; Creager, S.; Creutz, C.; Kagan,
if the shear between the islands and crystal surface can be derived C. R.; Kamat, P. V.; Lieberman, M.; Lindsay, S.; Marcus, R. A.; Metzger, R. M.;
from the frictional forces experienced by the AFM tip while Michel-Beyerle, M. E.; Miller, J. R.; Newton, M. D.; Rolison, D. R.; Sankey, O.;
Schanze, K. S.; Yardley, J.; Zhu, X. J. Phys. Chem. B 2003, 107, 6668.
scanning, the applicability of contact AFM to nanomanipulation (5) Haruta, M. Catech 2002, 6, 102.
is limited to relatively large objects (tens of nanometers in size). (6) Valden, M.; Lai, X.; Goodman, D. W. Science 1998, 281, 1647.
(7) Wang, J. Anal. Chim. Acta 2003, 500, 247.
† Part of the Molecular and Surface Forces special issue. (8) Penn, S. G.; He, L.; Natan, M. J. Curr. Opin. Chem. Biol. 2003, 7, 609.
‡ ICSI-CNRS. (9) Frens, G. Nature (London), Phys. Sci. 1973, 241, 20.
§ University of Basel and NCCR “Nanoscale Science”. (10) Grabar, K. C.; Allison, K. J.; Baker, B. E.; Bright, R. M.; Brown, K. R.;
| Universidad de Castilla-La Mancha. Freeman, R. G.; Fox, A. P.; Keating, C. D.; Musick, M. D.; Natan, M. J. Langmuir
⊥ Corning Incorporated. 1996, 12, 2353.
# University of CaliforniasSanta Barbara.
(11) Nikoobakht, B.; El-Sayed, M. A. Chem. Mater. 2003, 15, 1957.
(12) Nehl, C. L.; Liao, H.; Hafner, J. H. Nanoletters. 2006, 6, 683.
(1) Cuberes, M. T.; Schlitter, R. R.; Gimzewski, J. K. Appl. Phys. Lett. 1996, (13) Sau, T. K.; Murphy, C. J. Langmuir, 2004, 20, 6414.
69, 3016. (14) Brown, K. R.; Walter, D. G.; Natan, M. J. Chem. Mater. 2000, 12, 306.
(2) Lüthi, R.; Meyer, E.; Haefke, H.; Howald, L.; Gutmannsbauer, W.; (15) Brust, M.; Walker, M.; Bethell, D.; Sciffrin, D. J.; Whyman, R. J. Chem.
Güntherodt, H.-J. Science 1994, 266, 1979. Soc., Chem. Commun. 1994, 801-802.

10.1021/la702921v CCC: $40.75 © 2008 American Chemical Society
Published on Web 01/18/2008

-H. 125. 193. 4. Haidara. 240. C. The solution resulted in a a net hydrophilic character. decanted. a unique concentration of 0.-Y. Am. S. morphology of nanoparticles. Preece. 19.. S. Baeck. S. H. Jpn. T. K. (18) Sun. -CH3) or hydrophilic group (hydroxyl. Hydroxyl-stabilized or methyl-thiol. the for 20 min to pellet the nanoparticles. Soc. AFM images of 25 nm diameter gold nanoparticles Nanoparticle Synthesis. By measuring the phase shift while scanning. we report on the precise manipulation of gold nanoparticles (approximately 25 nm diameter) adsorbed onto silicon wafers in the tapping mode.23 protocol basically involved particle adsorption by immersing the samples for around 20 min in the suspension bath. T. J. Mendes. Frame sizes: 3 particles (HAuCl4. F.. Leggettt. was 25 ( 5 nm (Figure 1). Ruiz.. Z. Chen. Gold nanoparticles were either raw or coated with self-assembled monolayers ending with a hydrophobic (methyl. C.10.. T.22 The Au NP-coated silicon wafer was prepared using a D. Jaramillo. either as isolated clusters19 or single or multilayer coatings. J.-H. vinylpyridine). K. . 1997. 125. S. 2003. Random Adsorption. S. 231. Am. and then it was stored Coated Gold Nanoparticles. by reducing HAuCl4.-C.. 14. these nanocolloidal particles directly obtained from thiol (methyl.21. For the adsorption Aldrich and used as received. W. Critchley. Lu. 2006.. Hanwell. M.. 2008 Mougin et al. D. Chen.. Haidara. (22) Liu.25 Au nanoparticles were encap- (19) Mougin. M. 3H2O) supplied by ABCR. R. H. R. Y. The experimental version of two common syntheses. C. M.. R. 6639. (24) Jaramillo. 1993. Cuenya. J. (17) Lin. J. Zhang. Rafailovich. An amount of 300 µL of the expected the following. 2003. 2. and 1 µm. Dodecanethiol for methyl-terminated at 4 °C until use. S. In redispersed in tetrahydrofuran. Diegoli. D. synthesis are referred to as “raw nanoparticles (raw NPs)”. B 2004. Nanoparticle Adsorption. Shull. experiments. as determined from transmission Finally.. No..-H. Evans.. which.-M. The suspension was stabilized with citric acid trisodium (provided by Aldrich). -OH). Parikh. In such way.or hydroxyl-terminated) was added to the solution. We have also addressed the role of the environment and thermal activation with manipulation experiments performed in UHV and at different temperatures. Brydson. The average diameter of the synthesized monolayers (referred to as CH3) and 11-mercapto-1-undecanol for nanoparticles is 25 ( 5 nm. White. Preparation and Experimental Conditions Figure 2.21. imparts the negative charge nanoparticles were then purified from surfactant excess and other of the citrate ions to the nanogold particle surface. A. size of these nanogold particles. McFarland... K. N. respectively. P. Nano Lett. Wu..03 wt % nanoparticles in stabilized gold nanoparticles were synthesized according to a modified the aqueous or organic dispersion was used. Richardson. the dialyzed solution was centrifuged and particles were electron microscopy (TEM) images.. Figure 1. Phys. McFarland. ranging from 20 up to 150 °C. Evans..10. S. TEM image (400 nm × 400 nm) of 25 nm diameter gold nanoparticles.. 9566. micelle encapsulation method. Samples were provided by McFar- (16) Johnson.. Kellock... Castelein.21. Chem. Colloid Interface Sci. The solution was deposited onto the silicon wafer (21) Kunz. 2134. Lin. B. Baeck. H... J.22 was centrifuged at 7000 rpm was maintained at 20 ( 1 °C. H. Colloids Surf... 24.. Germany. Soc. R.. Ulman. Upon this initial adsorption stage. 7148.. K. G.. 1147. 18. K. Soc. (25) Cuenya. Ren. Raw Gold Nanoparticles. C. Langmuir 2003. J.22 The average reactants by dialysis during one week. 12928. J.. D. J. land’s group at the Univeristy of CaliforniasSanta Barbara (UCSB).24. 9450. W. Chem. Chen. although It was sonicated and stirred for approximately 2 h to allow the they bear some residual stabilizing citrate ions which impart to them grafting reaction to reach completion. 156. (23) Yee.. A. Langmuir 1998. J. Vol. slow color change from yellow to transparent. and the thick dispersion film to 1 mL of DI water to reduce the citric acid concentration. P. sulated by the diblock copolymer polystyrene-block-poly(2- (20) Mougin. Mao. B. and then resuspended samples were removed from the bath. F. G. the threshold value of the power dissipation needed for translating a single nanoparticle could be determined.. the hydroxyl groups (referred to as OH) were provided by Sigma. The colloidal deposited onto a silicon wafer: (a) controlled organization as suspension was made by reduction of an aqueous solution of nanogold described in the text and (b) random distribution.. Geometrical Organization. 345. A 2001. E. 6. we could investigate the influence of the hydrophobicity of the coatings on the mobility of the nanopar- ticles. R... The 25 nm diameter Au nanoparticles were synthesized as previ- 108. A.. ously..1578 Langmuir. Langmuir 2002. Chem. The remaining on the substrates was allowed to dry. whose temperature The raw nanosphere solution10. Gu. Chem. A.20 In the present work. S. L. J. J. E. Karlsruhe.

The same area was imaged several times while increasing the excitation amplitude of the cantilever between each scan. smooth translation almost AFM tip. 4. one can reasonably suppose the existence of an adsorbed thin Figure 5 shows the phase shift images corresponding to Figure water film at the NP/substrate interface under ambient conditions 3. Tamayo. the particles started to move.27 The base pressure was below 10-9 mbar. leading to a higher dissipated energy. Phys. area (Figure 4). and dried under nitrogen flow.. This can be understood assuming amplitude was overcome. A.. Surf. The motion that the particles sticking on the substrate are not rigidly “fixed” of the particles consisted of either abrupt “jumps” along the fast and that they dissipate energy in the vibrations induced by the scan direction (circle in Figure 3b). In such case. and the quality factor of the free cantilever. Haefke.3 particles were found to be aligned at two edges of the scanned For these raw NPs bearing residual stabilizing citrate ions. After being dip-coated. 312. Vol. substrate. H. the energy required (dissipated) in moving these particles. and φ is the phase shift caused by the interaction between the tip and the underlying particles or surface. The particles accumulate at two edges of the area. UHV Measurements. was kept constant by a feedback loop.. When a certain value of the when the particles start moving. the gold reported by Ritter et al. and not the oscillation amplitude of the tip. As a result. lowering the adhesion and therefore correspond to regions with higher and lower power dissipation. which results in an apparent lower height of them. Results and Discussion scanning downward. starting from left to right. When the particles move smoothly. the polymer was removed using an oxygen plasma treatment (Figure 2). the frequency shift from the out-of-contact resonance. on flat antimony particles. the oscillation amplitude of the tip. Manipulation Setup. J. (27) Howald. PPP-NCLR and CONT cantilevers from Nanosensors were used in the two cases. 1999. the resonance frequency. only the upper part of them is imaged. R. San Paulo. Nanoscope IV from Veeco and Mobile S from Nanosurf). quality factors around 800 and 600. MPP12100 from Veeco and PPP-NCLR from Nanosensors) were used. k. or a of 8 × 10-11 W/particle.. the spring constant. R. a silicon sample was dipped into a colloidal solution of raw dissipation is higher on the “fixed” particles compared to the nanoparticles. Meyer. Air Measurements. Zoom out of the area in Figure 3 (a) before and (b) after 3. Rudin. and Q are. No. Overney. Frame size: 4 µm. the brighter and darker areas in Figure 5 which acts as a lubricant layer. Lett. In the first series of measurements. it turns out that the power compared to these ambient humidity results. which is consistent with the values combination of both (ellipse in Figure 3b). 24. Due to the high quality factor in UHV. . around 120 and 190 kHz. We have also performed measurements in the contact mode. H. 27. f0. Aset.. f0. dried. the adhesion of raw (citric-coated) nanoparticles with the silicon Güntherodt.-J. Lüthi. and nominal spring constants of 5 and 48 N/m (respectively. Raw Particles on Silicon. During manipulation. Figure 4. R. 2008 1579 Figure 3. From the images. where the set point is determined by the normal load acting between the tip and sample. 1993. The images in UHV were acquired with a home-built atomic force microscope available at the University of Basel. Appl.Manipulation of Gold Nanoparticles Langmuir. However. The images in air were acquired with two commercial atomic force microscopes (Multimode... the power dissipation accompanying the tip-sample interaction can be determined using the formula26 Pdis ) kf0 4π (ApiezoAset sin φ - A2set Q ) (1) where Apiezo is the oscillation amplitude of a piezo element coupled to the cantilever. Gold particles on silicon. we estimated an average power dissipation along the slow scan direction (ellipse in Figure 3c). From eq 1. According to eq 1.. Frame size: 2 µm. H. L. Interface Anal. and repeatedly scanned at increasing silicon substrate. the dissipation becomes much lower excitation amplitude (Figure 3). 63. was used as the imaging parameter (NC-AFM). the absence of humidity while working under ultrahigh vacuum rather favors (26) Garcia. Rectangular silicon cantilevers with resonance frequencies. As respectively. respectively. 117. E.

. except for raw NPs (randomly distributed) for which the R2 value is ∼0. the thin adsorbed water film which forms on the silicon wafer acts as a lubricant when confined between the hydrophobized CH3-coated nanoparticles and the hydrophilic substrate. the mobility of the particles. Intuitively. Similar kinetic effects have been recognized in friction on hydrophilic surfaces measured with different scan velocities. We also observed that the manipulation of hydrophilic coated nanoparticles often results in damage to the tip due to the high adhesion forces at the interface. (32) Meredith. 97. and (light with T which accounts for a thermally activated process. and Eact(T) nanoparticules.1580 Langmuir. Macromolecules coated nanogold particles is more than 10 times higher than the 2000. J. and moving the particles again on a flat silicon substrate. since the temperature (thermal energy kT) acts as an obstacle in forming stable intermolecular bonds and water bridges between the particles and substrate. Intermolecular & Surface Forces. More surprising is the behavior of the OH- by coating the gold nanoparticles with OH. 24. Chem. which shows the average power (30) Yoshizawa. No. Haidara. 1993. as one could expect for the lowest nanoparticle investigated the role of hydrophobicity in the manipulation process substrate adhesion. F. Phys. dissipation required to start the motion of the particles. and this effect appears to be stronger on hydrophilic particles. G. as described in the previous section. Chen. CH3-coated nanoparticles. (dark gray column) CH3-coated NPs. H. summarized in Figure 6. (0) represents the energy input involved in the motion of the particle. H.. E. Israelachvili. 4. Castelein.. Phys.32 The gray column) raw NPs that are randomly distributed on the substrate. for temperature ranging from 20 to 150 °C. The energy required to move OH. Brune.28-31 Temperature Effects. These experimental data display a good linear fit with R2 values higher than 0. 33. one could expect this result. indicate 4128. A.. J. Y. (O) raw nanoparticles geometrically organized. no evident damage was observed on the working areas.. ∆Eact ) (E0 . Figure 6b corresponds to a logarithm plot of the dissipated power as a function of the inverse of temperature (1/T). This figure shows in all cases that power dissipation decreases with temperature. Karim. Smith. 2008 Mougin et al. Academic Press: San Diego. 2004. Lett. This Figure 6.-L. slopes of these linear fits correspond to ∆Eact/k.. P. 11. E. Levy. (29) Mougin. and (2) OH-coated nanoparticles This energy variation (slope) is higher for the methyl-coated NPs. As a next step. The dark spots on the top of the particles correspond to the phase signal switching from +180° to -180° and do not have any special physical meaning. where ∆Eact (b) Logarithm of the dissipated power in moving raw and coated represents the variation of the activation energy Eact(T) with NPs on a silicon wafer versus the inverse of temperature for (9) raw respect to a reference state E0. indicating a stronger decrease of the input energy with Functionalized Particles on Silicon. 88. CA. This strong adhesion between the silicon substrate and hydrophilic coated nanoparticles which primarily arises from intermolecular interactions may also involve a contribution from capillary bridges which form under ambient conditions between the substrate and the NPs on one hand.and CH3-terminated coated NPs which show a similar sensitivity of their mobility to thiol groups. Tribol. 185505. Amis. that the presence of a hydrophobic interface significantly enhances (31) Israelachvili.78. Lett. K. Figure 6a represents a histogram of the raw data of power dissipation versus temperature.. we temperature. J. 1992. and between the NPs on the other hand.. C. J. Vol.. A.Eact). 17. as has already been observed. J. during this temperature-dependent manipulation. In contrast.90 for all NP/substrate couples. . Figure 5. (a) Average power dissipation accompanying the onset linear behavior of log(dissipated power) versus (1/T) actually of motion of raw and coated nanoparticles on silicon in air versus temperature: (black column) raw NPs that are uniformly distributed corresponds to an exponential decay of the dissipated power on the substrate. 9747. ReV. The results (28) Riedo. reducing the adhesion between them. energy required to move CH3-coated particles. 2002. H.28-29 It is worth noting that. Phase shift images corresponding to Figure 3.

mobility of the particles is greatly enhanced by hydrophobic Vacuum Effects.Manipulation of Gold Nanoparticles Langmuir. the interparticle forces are of the order of long range interactions. This result particularly illustrates the lubricating role of the LA702921V . nanomanipulation process changes in the absence of humidity. even when imaged at the National Science Foundation. the their organization. In these conditions. Finally. their mobility is not manipulation of gold nanoparticles considering different coatings. Frame of particles is essentially affected in this limit by electrostatic size: 1 µm. and manipulation could not be achieved. The mobility Figure 7. this result means that so long as the particle number Conclusions density np is such as the interparticle distance dp ∼ (np)-1/2 is In conclusion. which is 70 nm for the random distribution and 100 nm for the ordered one. silicon after the sample was transferred in UHV without any Thermal activation plays an opposite role. It is thus normal in adsorbed water layer between the particle and the substrate in the absence of both physical contact and notable intermolecular both the free (Brownian) and externally driven motion of the forces between the particles that their mobility is independent nanoparticles. und the European Science applied between the tip and particles. Acknowledgment. This work was supported by the Swiss the particles could not be moved. No effects coming from the the particles. In other words. interactions arising from the residue of the synthesis of citric acid that may be adsorbed on the particles. we have investigated how the interlayers. Gold particles on silicon in ultrahigh vacuum. due to the size) were observed. In ambient conditions. At such a scale. obtained as described previously. Vol. the raw particles have an activation energy lying in between those of the more hydrophilic and hydrophobic nanoparticles. which could have changed the arrangement of between the particles and substrate. 4. Finally. formed by organic coatings deposited on the particles. Organization Effects. in contact mode. the Kommission zur Förderung 100 nm). The first and third columns in the series in Figure 6a show the power threshold accompanying the motion of the randomly and geometrically organized distributions of nanoparticles. Foundation Nanotribo Program. absence of convolution effects coming from the water layer which covers the particles under ambient conditions. the gold particles strongly adhered The topography image in Figure 7 shows the gold particles on to the silicon substrate. This result can be explained by the average distance between nano-objects. An explanation for this less expected behavior may come at least partly from the sensitivity of the adsorbed water layer to temperature in this more hydrophilic system. we performed systematic studies of the larger than the range of short-ranged forces. The shapes of the particles are well defined. of their organization (random or geometrically ordered). with forces of a few nanonewtons von Technologie und Innovation. no motion was observed. 24. The power dissipation at different temperatures is comparable in both cases. In the absence of humidity. reducing adhesion further treatment. affected by their mutual interaction and is thus independent of environments. and temperatures. 2008 1581 temperature (strong decrease of input energy with T). Also. No. the National Center of Competence maximum amplitude applicable with our system (in the order of in Research on Nanoscale Science. and ordering of nanoparticles on large scales (well above the particle the structure of some aggregates can be recognized.