- Email: [email protected]
Temperature-induced processes for size-selected metallic nanoparticles on surfaces
G Model
ARTICLE IN PRESS
APSUSC-33162; No. of Pages 4
Applied Surface Science xxx (2016) xxx–xxx
Contents lists available at ScienceDirect
Applied Surface Science journal homepage: www.elsevier.com/locate/apsusc
Temperature-induced processes for size-selected metallic nanoparticles on surfaces H. Bettermann ∗ , M. Werner, M. Getzlaff ∗∗ Institute of Applied Physics, University of Düsseldorf, Universitätsstr. 1, D-40225 Düsseldorf, Germany
a r t i c l e
i n f o
Article history: Received 29 December 2015 Received in revised form 22 April 2016 Accepted 23 April 2016 Available online xxx Keywords: Nanoparticles Melting behavior Temperature dependence Ostwald ripening Scanning tunneling microscopy (STM)
a b s t r a c t The melting behavior of Iron-Nickel alloy nanoparticles on W(110) was studied under UHV conditions as a function of heating temperature and heating duration. These particles were found to be stable at 423 K without evaporation or diffusion taking place. Unrolling carpet behavior occurs at higher temperatures. This creates ramified islands around the nanoparticles. Ostwald ripening at higher temperatures or longer heating times is creating compact islands. The melting of these nanoparticles opens the possibility for thin film growth of FeNi alloys. The formation of monolayer high islands is a strong contrast to Fe, Co, and FeCo alloy nanoparticles which are dominated by direct evaporation, single atom surface diffusion and anisotropic spreading. © 2016 Elsevier B.V. All rights reserved.
1. Introduction The size range between bulk material on the one hand and single atoms on the other hand is giving nanoparticles a unique set of properties and fascinating behavior. Having been under investigation for a long time, the materials and size are obviously important contributors to these properties. These also include the melting point and other thermodynamic characteristics. The stoichiometry is another important factor when dealing with alloy systems. While it is possible to grow nanoparticles directly on a substrate, our work is based on the creation of nanoparticles from the gas phase and their subsequent deposition. Previous research includes the crystallization and alignment of iron particles during deposition [1] and their thermal treatment [2,3]. While thermal treatment typically takes place on a scale of minutes and longer, the deposition and subsequent relaxation lasts only pico- to nanoseconds. Therefore the fine tuning of the kinetic energy is offering a different approach to the treatment of nanoparticles. The additional energy can lead to a complete recrystallization on impact within picoseconds [4]. In this contribution size selected supported nanoparticles made of iron-nickel alloy being prepared under UHV conditions and
∗ Principal corresponding author. ∗∗ Corresponding author. E-mail addresses: [email protected] (H. Bettermann), [email protected] (M. Getzlaff).
investigated in-situ by scanning tunneling microscopy will be discussed. This extends our previous work concerning temperature induced effects of transition metal nanoparticles on W(110).
2. Materials and methods 2.1. Nanoparticles Nanoparticles are generated in an Arc Cluster Ion Source (ACIS) [5] with a typical size between 5 and 15 nm. It consists of a hollow cathode which is made from Iron0.5 Nickel0.5 alloy. Argon is injected in the back of the cathode. An arc discharge strikes the inner wall and removes metal. The Argon gas leaves the cathode through a small nozzle and is taking the metal vapor with it. The gas is cooled in a supersonic expansion. Two skimmer stages with a roots and turbomolecular pump respectively remove argon gas and shape a narrow beam of nanoparticles. The additional injection of Helium can improve the cooling process. A high amount of singly charged particles in the beam and the narrow velocity distribution of the super-sonic expansion facilitate mass selection by a static electric field. This field is created in a quadrupole deflector with a maximum voltage of ±2 kV. The best possible mass resolution is M/M = 10%. The first pumping stage can be throttled to widen the velocity distribution. This allows for a wider mass distribution. Particle flux can be monitored at a grid after deflection. The subsequent deposition fulfills soft-landing condition at a kinetic energy below 0.1 eV per atom. The pressure
http://dx.doi.org/10.1016/j.apsusc.2016.04.161 0169-4332/© 2016 Elsevier B.V. All rights reserved.
Please cite this article in press as: H. Bettermann, et al., Temperature-induced processes for size-selected metallic nanoparticles on surfaces, Appl. Surf. Sci. (2016), http://dx.doi.org/10.1016/j.apsusc.2016.04.161
G Model APSUSC-33162; No. of Pages 4
ARTICLE IN PRESS
2
H. Bettermann et al. / Applied Surface Science xxx (2016) xxx–xxx
in the deposition chamber is in the 10−7 mbar regime during ACIS operation. 2.2. Substrate preparation The substrate for nanoparticle deposition is a tungsten single crystal with a (110) surface. Well known cleaning procedures are outlined by e.g. Bode et al. [6]. The W(110) crystal is cleaned before deposition in a separate chamber by electron bombardment heating (flashing) at T > 2000 K and p0 < 5 ×10−10 mbar. Carbon impurities are removed by oxygen annealing at T = 1500 K and poxygen = 10−7 mbar. Temperature is determined by narrow band pyrometric measurement at wavelength of 960 nm and emissivity of = 0.3. A resistive heater is integrated into the manipulator of the STM chamber. This allows tempering of samples between 350 K and 1100 K. 2.3. Scanning tunneling microscopy The nanoparticles are investigated in situ by means of scanning tunneling microscopy (STM). The microscope is a MicroSPM by Omicron. Base pressure is below 3 × 10−10 mbar. All measurements are done at room temperature. STM gives easy access to the height of nanoparticles in the size regime of ACIS. Additional information can be obtained about the structure of the top-facets. Determination of lateral size is limited due to tip-surface convolution. XY-calibration and imagedistortion were checked on carburized W(110). The R(15 × 3) reconstruction exhibits thin lines with a separation of 1.37 nm that are aligned at ±35◦ with respect to the W[001] axis [7]. Axis orientation was cross-checked by LEED while x-calibration is in perfect agreement with measurements of HOPG. Slow scanning speeds (typically 1000 nm/s at 1000 nm image size) were employed to keep the tip from crashing into nanoparticles and for providing low-drift conditions. Z-calibration is based on the well-known step height of W(110). 3. Theory The melting properties of nano-scaled systems were previously investigated. A behavior that is often observed is known as unrolling carpet. I.e. a layer of atoms forms around the nanoparticle or dot during melting. Reuter et al. [8] have shown this behavior for tempering of iron, nickel and cobalt dots that were grown on W(110). Similar behavior has been found for deposited iron nanoparticles. Other systems, like Pd [9] exhibit similar properties.
(a)
(b)
Fig. 1. Hard ball model for surface diffusion of islands. Unrolling carpet mechanism: (a) Atoms from the nanoparticle are moving on top of the monolayer until the reach it’s edge. Adapted from [8]. Anisotropic diffusion on bcc(110): (b) The free atom (1) can move along the 111 axes. Atom (2) may detach from the island. Atom (3) may move along the island until it reaches position (4). Adapted from [10].
The effect is driven by a strong interaction between substrate and the first monolayer whereas the interaction between first and second monolayer is comparably weak. The first monolayer is therefore stable. Single atoms from the central nanoparticle are transported by diffusion until they reach the edge of the first monolayer (see Fig. 1(a)). The monolayer is spreading because these atoms are subsequently trapped at the step edges. One factor that contributes to the shape of islands is the crystallographic structure of the substrate. Köhler et al. [10] describe a model for an anisotropic spreading of islands on a bcc(110) surface. The 111 axes are the easiest directions for hopping between adsorption sites (see Fig. 1(b), atom (1)). Hopping along the 001 direction is rather unlikely because it leads over an on-top position. I.e. an atom (2) that is attached at a 110 edge of an island will probably detach from the island while atoms (3) at 111 edges can move easily along the edge until they reach a relatively stable site with multiple nearest neighbors (4). Therefore islands are elongated along the [001] axis. 4. Results Fe0.5 Ni0.5 nanoparticles were deposited on W(110) and subsequently heated. This section will discuss the properties of these nanoparticles after deposition and their response to elevated temperatures. 4.1. Softlanding Fig. 2(a) shows the size distribution for different source parameters. Both the first stage pumping speed and the deflection voltage were varied. The nanoparticles considered for further treatment
Fig. 2. FeNi nanoparticles after deposition on W(110). (a) Height distribution for different source parameters. (b) STM image after deposition at reduced pumping speed. No preferred landing sites (+1.0 V, 0.8 nA).
Please cite this article in press as: H. Bettermann, et al., Temperature-induced processes for size-selected metallic nanoparticles on surfaces, Appl. Surf. Sci. (2016), http://dx.doi.org/10.1016/j.apsusc.2016.04.161
G Model APSUSC-33162; No. of Pages 4
ARTICLE IN PRESS H. Bettermann et al. / Applied Surface Science xxx (2016) xxx–xxx
3
Fig. 3. STM images after heating at 458 K (+1.0 V, 0.8 nA). (a) After 30 min. Unrolling carpet has created irregular shaped islands around these two nanoparticles. (b) After 120 min. Islands have developed into a compact form. Some material has agglomerated along the substrate’s step edge.
have an average height of 10.4±1.3 nm. The nanoparticles cover the whole substrate without any preference toward specific sites (Fig. 2(b)). No agglomeration at step edges occurs. They do not form larger clusters. Surface diffusion after deposition can therefore be excluded. 4.2. Tempering One hour of tempering to 423 K does not affect the size or shape of the nanoparticles. Islands with monoatomic height around some nanoparticles can be observed after 30 min at 458 K. This is due to unrolling carpet behavior. These islands have an irregular shape (see Fig. 3(a)). Longer tempering leads to diffusion of material away from the islands toward adjacent terrace steps. A clear separation between the tungsten step edge and the accumulated FeNi layer is visible. The islands attain elliptic shape with elongation along the substrate’s [001] direction. Some nanoparticles remain intact. This elongation is independent from the distance to tungsten step edges. This behavior is accelerated at higher temperatures. The increased surface diffusion can be observed in the formation of
small-sized islands which are spread evenly across the substrate (see Fig. 4(a)). A distinct separation between the step-edges of the substrate and of iron-nickel is visible. All nanoparticles disappear by tempering at 1010 K for 30 min (see Fig. 4(b)). The material from the nanoparticles has formed monolayer high islands and terraces. Vacancy-islands within terraces can be observed (not shown). The islands are evenly spread across the surface. Larger islands may indicate the site of molten nanoparticles. Furthermore no separation between step-edges of tungsten and iron-nickel monolayer is visible. The material contained in the nanoparticles is sufficient to cover the surface with roughly 1.6 monolayers. The observed coverage of islands is 23%. The observed coverage in the second layer might be lower than expected due to the creation of a closed layer adjacent to the terrace steps. 5. Discussion We have examined the response of iron-nickel nanoparticles to elevated temperature. These particles remain unaffected at 423 K. Unrolling carpet behavior was observed above 485 K. This carpet
Fig. 4. STM images after 30 min of heating. (a) At 581 K. Islands have spread along the step edges. New dots have spread along the surface (+1.0 V, 0.8 nA). (b) At 1010 K. Monolayer high terraces and islands (+1.0 V, 1.8 nA).
Please cite this article in press as: H. Bettermann, et al., Temperature-induced processes for size-selected metallic nanoparticles on surfaces, Appl. Surf. Sci. (2016), http://dx.doi.org/10.1016/j.apsusc.2016.04.161
G Model APSUSC-33162; No. of Pages 4
ARTICLE IN PRESS
4
H. Bettermann et al. / Applied Surface Science xxx (2016) xxx–xxx
Temperature induced surface diffusion leads to compact shapes. This effect is enhanced with increasing temperature. The melting behavior of iron-nickel nanoparticles on W(110) is in strong contrast to previous results concerning iron and cobalt. Cobalt and Iron0.5 Cobalt0.5 alloy particles do not exhibit unrolling carpet behavior at any temperature. Height reduction is only due to direct evaporation. A change in shape occurs only at very high temperatures (>1000 K), the particles are elongated along the tungsten’s [001] axis (see Fig. 6, [3]). This behavior can also be observed for iron particles. In addition iron has a critical temperature Tcrit = 640 K. Below that temperature evaporation without surface spreading is taking place. Tcrit is sufficient to cause a realignment from random to epitaxial orientation [1]. Unrolling carpet behavior can be observed above Tcrit [2]. Fig. 5. Development of nanoparticles and islands at different temperatures. 30 min of tempering each. Inset: Tempering at 458 K for different time intervals.
6. Conclusions We have shown that the melting point for iron-nickel nanoparticles is below 485 K. This temperature is close to the melting point of pure iron nanoparticles [1]. We could not observe any evaporation of material below that temperature. Unrolling carpet mechanism and surface diffusion are the driving processes to form the different shapes above the melting point. A slight anisotropy occurs during melting. References
Fig. 6. STM image of FeCo nanoparticles after heating for 1 h at 1000 K. The nanoparticles are spread along the [001] direction (+1.0 V, 0.1 nA).
is one monolayer high and can unroll even across terrace steps (see Fig. 3(a)). The time and temperature dependent development from nanoparticles to islands is shown in Fig. 5. The development to nanoparticles with a surrounding layer (unrolling nanoparticles) and subsequently to islands depends on the temperature and the duration of annealing. The islands initially formed have irregular shapes. This is in agreement with measurements of iron nanoparticles of similar size by Rosellen et al. [11]. Longer heating leads to compact island shapes. Accumulation of atoms at nearby step edges indicate that diffusion on substrate away from the nanoparticles is taking place. The remaining islands are elongated along the [001] axis. This is in good agreement with theoretical models for diffusion on bcc(110) (see Section 3).
[1] A. Kleibert, A. Voitkans, K.-H. Meiwes-Broer, Size-dependent alignment of Fe nanoparticles upon deposition onto W(110), Phys. Rev. B 81 (7) (2010), http:// dx.doi.org/10.1103/physrevb.81.073412. [2] W. Rosellen, C. Kleinhans, V. Hückelkamp, F. Bulut, A. Kleibert, J. Bansmann, M. Getzlaff, Influence of substrate and temperature on the shape of deposited Fe, Co, and FeCo nanoparticles, Phys. Status Solidi B 247 (5) (2010) 1032–1038, http://dx.doi.org/10.1002/pssb.200945569. [3] F. Bulut, W. Rosellen, M. Getzlaff, Structural properties of size-selected FeCo nanoparticles deposited on W(110), Appl. Phys. A 97 (1) (2009) 185–189, http://dx.doi.org/10.1007/s00339-009-5230-2. [4] W. Rosellen, H. Bettermann, T. Veltum, M. Getzlaff, Low energy impact of size selected FeCo nanoparticles with a W(110) surface, Physica E 44 (7-8) (2012) 1683–1686, http://dx.doi.org/10.1016/j.physe.2012.04.020. [5] R.-P. Methling, V. Senz, E.-D. Klinkenberg, T. Diederich, J. Tiggesbäumker, G. Holzhüter, J. Bansmann, K. Meiwes-Broer, Magnetic studies on mass-selected iron particles, Eur. Phys. J. D 16 (1) (2001) 173–176, http://dx.doi.org/10. 1007/s100530170085. [6] M. Bode, S. Krause, L. Berbil-Bautista, S. Heinze, R. Wiesendanger, On the preparation and electronic properties of clean W(110) surfaces, Surf. Sci. 601 (16) (2007) 3308–3314, http://dx.doi.org/10.1016/j.susc.2007.06.017. [7] M. Bode, R. Pascal, R. Wiesendanger, STM study of carbon-induced reconstructions on W(110): strong evidence for a surface lattice deformation, Surf. Sci. 344 (3) (1995) 185–191, http://dx.doi.org/10.1016/00396028(95)00873-x. [8] D. Reuter, G. Gerth, J. Kirschner, Anisotropic diffusion of 3 d metals on W(110): Competition between crystalline structure and surface steps, Phys. Rev. B 57 (4) (1998) 2520–2529, http://dx.doi.org/10.1103/physrevb.57.2520. [9] R. Butz, H. Wagner, Surface diffusion of Pd and Au on W single crystal planes, Surf. Sci. 87 (1) (1979) 69–84, http://dx.doi.org/10.1016/00396028(79)90170-5. [10] U. Köhler, C. Jensen, C. Wolf, A. Schindler, L. Brendel, D. Wolf, Investigation of homoepitaxial growth on bcc surfaces with STM and kinetic Monte Carlo simulation, Surf. Sci. 454-456 (2000) 676–680, http://dx.doi.org/10.1016/ s0039-6028(00)00151-5. [11] W. Rosellen, Strukturelle Eigenschaften massenselektierter 3d-Metall-Partikel, Ph.D. thesis, Universitäts-und Landesbibliothek der Heinrich-Heine-Universität Düsseldorf, 2013.
Please cite this article in press as: H. Bettermann, et al., Temperature-induced processes for size-selected metallic nanoparticles on surfaces, Appl. Surf. Sci. (2016), http://dx.doi.org/10.1016/j.apsusc.2016.04.161
ARTICLE IN PRESS
APSUSC-33162; No. of Pages 4
Applied Surface Science xxx (2016) xxx–xxx
Contents lists available at ScienceDirect
Applied Surface Science journal homepage: www.elsevier.com/locate/apsusc
Temperature-induced processes for size-selected metallic nanoparticles on surfaces H. Bettermann ∗ , M. Werner, M. Getzlaff ∗∗ Institute of Applied Physics, University of Düsseldorf, Universitätsstr. 1, D-40225 Düsseldorf, Germany
a r t i c l e
i n f o
Article history: Received 29 December 2015 Received in revised form 22 April 2016 Accepted 23 April 2016 Available online xxx Keywords: Nanoparticles Melting behavior Temperature dependence Ostwald ripening Scanning tunneling microscopy (STM)
a b s t r a c t The melting behavior of Iron-Nickel alloy nanoparticles on W(110) was studied under UHV conditions as a function of heating temperature and heating duration. These particles were found to be stable at 423 K without evaporation or diffusion taking place. Unrolling carpet behavior occurs at higher temperatures. This creates ramified islands around the nanoparticles. Ostwald ripening at higher temperatures or longer heating times is creating compact islands. The melting of these nanoparticles opens the possibility for thin film growth of FeNi alloys. The formation of monolayer high islands is a strong contrast to Fe, Co, and FeCo alloy nanoparticles which are dominated by direct evaporation, single atom surface diffusion and anisotropic spreading. © 2016 Elsevier B.V. All rights reserved.
1. Introduction The size range between bulk material on the one hand and single atoms on the other hand is giving nanoparticles a unique set of properties and fascinating behavior. Having been under investigation for a long time, the materials and size are obviously important contributors to these properties. These also include the melting point and other thermodynamic characteristics. The stoichiometry is another important factor when dealing with alloy systems. While it is possible to grow nanoparticles directly on a substrate, our work is based on the creation of nanoparticles from the gas phase and their subsequent deposition. Previous research includes the crystallization and alignment of iron particles during deposition [1] and their thermal treatment [2,3]. While thermal treatment typically takes place on a scale of minutes and longer, the deposition and subsequent relaxation lasts only pico- to nanoseconds. Therefore the fine tuning of the kinetic energy is offering a different approach to the treatment of nanoparticles. The additional energy can lead to a complete recrystallization on impact within picoseconds [4]. In this contribution size selected supported nanoparticles made of iron-nickel alloy being prepared under UHV conditions and
∗ Principal corresponding author. ∗∗ Corresponding author. E-mail addresses: [email protected] (H. Bettermann), [email protected] (M. Getzlaff).
investigated in-situ by scanning tunneling microscopy will be discussed. This extends our previous work concerning temperature induced effects of transition metal nanoparticles on W(110).
2. Materials and methods 2.1. Nanoparticles Nanoparticles are generated in an Arc Cluster Ion Source (ACIS) [5] with a typical size between 5 and 15 nm. It consists of a hollow cathode which is made from Iron0.5 Nickel0.5 alloy. Argon is injected in the back of the cathode. An arc discharge strikes the inner wall and removes metal. The Argon gas leaves the cathode through a small nozzle and is taking the metal vapor with it. The gas is cooled in a supersonic expansion. Two skimmer stages with a roots and turbomolecular pump respectively remove argon gas and shape a narrow beam of nanoparticles. The additional injection of Helium can improve the cooling process. A high amount of singly charged particles in the beam and the narrow velocity distribution of the super-sonic expansion facilitate mass selection by a static electric field. This field is created in a quadrupole deflector with a maximum voltage of ±2 kV. The best possible mass resolution is M/M = 10%. The first pumping stage can be throttled to widen the velocity distribution. This allows for a wider mass distribution. Particle flux can be monitored at a grid after deflection. The subsequent deposition fulfills soft-landing condition at a kinetic energy below 0.1 eV per atom. The pressure
http://dx.doi.org/10.1016/j.apsusc.2016.04.161 0169-4332/© 2016 Elsevier B.V. All rights reserved.
Please cite this article in press as: H. Bettermann, et al., Temperature-induced processes for size-selected metallic nanoparticles on surfaces, Appl. Surf. Sci. (2016), http://dx.doi.org/10.1016/j.apsusc.2016.04.161
G Model APSUSC-33162; No. of Pages 4
ARTICLE IN PRESS
2
H. Bettermann et al. / Applied Surface Science xxx (2016) xxx–xxx
in the deposition chamber is in the 10−7 mbar regime during ACIS operation. 2.2. Substrate preparation The substrate for nanoparticle deposition is a tungsten single crystal with a (110) surface. Well known cleaning procedures are outlined by e.g. Bode et al. [6]. The W(110) crystal is cleaned before deposition in a separate chamber by electron bombardment heating (flashing) at T > 2000 K and p0 < 5 ×10−10 mbar. Carbon impurities are removed by oxygen annealing at T = 1500 K and poxygen = 10−7 mbar. Temperature is determined by narrow band pyrometric measurement at wavelength of 960 nm and emissivity of = 0.3. A resistive heater is integrated into the manipulator of the STM chamber. This allows tempering of samples between 350 K and 1100 K. 2.3. Scanning tunneling microscopy The nanoparticles are investigated in situ by means of scanning tunneling microscopy (STM). The microscope is a MicroSPM by Omicron. Base pressure is below 3 × 10−10 mbar. All measurements are done at room temperature. STM gives easy access to the height of nanoparticles in the size regime of ACIS. Additional information can be obtained about the structure of the top-facets. Determination of lateral size is limited due to tip-surface convolution. XY-calibration and imagedistortion were checked on carburized W(110). The R(15 × 3) reconstruction exhibits thin lines with a separation of 1.37 nm that are aligned at ±35◦ with respect to the W[001] axis [7]. Axis orientation was cross-checked by LEED while x-calibration is in perfect agreement with measurements of HOPG. Slow scanning speeds (typically 1000 nm/s at 1000 nm image size) were employed to keep the tip from crashing into nanoparticles and for providing low-drift conditions. Z-calibration is based on the well-known step height of W(110). 3. Theory The melting properties of nano-scaled systems were previously investigated. A behavior that is often observed is known as unrolling carpet. I.e. a layer of atoms forms around the nanoparticle or dot during melting. Reuter et al. [8] have shown this behavior for tempering of iron, nickel and cobalt dots that were grown on W(110). Similar behavior has been found for deposited iron nanoparticles. Other systems, like Pd [9] exhibit similar properties.
(a)
(b)
Fig. 1. Hard ball model for surface diffusion of islands. Unrolling carpet mechanism: (a) Atoms from the nanoparticle are moving on top of the monolayer until the reach it’s edge. Adapted from [8]. Anisotropic diffusion on bcc(110): (b) The free atom (1) can move along the 111 axes. Atom (2) may detach from the island. Atom (3) may move along the island until it reaches position (4). Adapted from [10].
The effect is driven by a strong interaction between substrate and the first monolayer whereas the interaction between first and second monolayer is comparably weak. The first monolayer is therefore stable. Single atoms from the central nanoparticle are transported by diffusion until they reach the edge of the first monolayer (see Fig. 1(a)). The monolayer is spreading because these atoms are subsequently trapped at the step edges. One factor that contributes to the shape of islands is the crystallographic structure of the substrate. Köhler et al. [10] describe a model for an anisotropic spreading of islands on a bcc(110) surface. The 111 axes are the easiest directions for hopping between adsorption sites (see Fig. 1(b), atom (1)). Hopping along the 001 direction is rather unlikely because it leads over an on-top position. I.e. an atom (2) that is attached at a 110 edge of an island will probably detach from the island while atoms (3) at 111 edges can move easily along the edge until they reach a relatively stable site with multiple nearest neighbors (4). Therefore islands are elongated along the [001] axis. 4. Results Fe0.5 Ni0.5 nanoparticles were deposited on W(110) and subsequently heated. This section will discuss the properties of these nanoparticles after deposition and their response to elevated temperatures. 4.1. Softlanding Fig. 2(a) shows the size distribution for different source parameters. Both the first stage pumping speed and the deflection voltage were varied. The nanoparticles considered for further treatment
Fig. 2. FeNi nanoparticles after deposition on W(110). (a) Height distribution for different source parameters. (b) STM image after deposition at reduced pumping speed. No preferred landing sites (+1.0 V, 0.8 nA).
Please cite this article in press as: H. Bettermann, et al., Temperature-induced processes for size-selected metallic nanoparticles on surfaces, Appl. Surf. Sci. (2016), http://dx.doi.org/10.1016/j.apsusc.2016.04.161
G Model APSUSC-33162; No. of Pages 4
ARTICLE IN PRESS H. Bettermann et al. / Applied Surface Science xxx (2016) xxx–xxx
3
Fig. 3. STM images after heating at 458 K (+1.0 V, 0.8 nA). (a) After 30 min. Unrolling carpet has created irregular shaped islands around these two nanoparticles. (b) After 120 min. Islands have developed into a compact form. Some material has agglomerated along the substrate’s step edge.
have an average height of 10.4±1.3 nm. The nanoparticles cover the whole substrate without any preference toward specific sites (Fig. 2(b)). No agglomeration at step edges occurs. They do not form larger clusters. Surface diffusion after deposition can therefore be excluded. 4.2. Tempering One hour of tempering to 423 K does not affect the size or shape of the nanoparticles. Islands with monoatomic height around some nanoparticles can be observed after 30 min at 458 K. This is due to unrolling carpet behavior. These islands have an irregular shape (see Fig. 3(a)). Longer tempering leads to diffusion of material away from the islands toward adjacent terrace steps. A clear separation between the tungsten step edge and the accumulated FeNi layer is visible. The islands attain elliptic shape with elongation along the substrate’s [001] direction. Some nanoparticles remain intact. This elongation is independent from the distance to tungsten step edges. This behavior is accelerated at higher temperatures. The increased surface diffusion can be observed in the formation of
small-sized islands which are spread evenly across the substrate (see Fig. 4(a)). A distinct separation between the step-edges of the substrate and of iron-nickel is visible. All nanoparticles disappear by tempering at 1010 K for 30 min (see Fig. 4(b)). The material from the nanoparticles has formed monolayer high islands and terraces. Vacancy-islands within terraces can be observed (not shown). The islands are evenly spread across the surface. Larger islands may indicate the site of molten nanoparticles. Furthermore no separation between step-edges of tungsten and iron-nickel monolayer is visible. The material contained in the nanoparticles is sufficient to cover the surface with roughly 1.6 monolayers. The observed coverage of islands is 23%. The observed coverage in the second layer might be lower than expected due to the creation of a closed layer adjacent to the terrace steps. 5. Discussion We have examined the response of iron-nickel nanoparticles to elevated temperature. These particles remain unaffected at 423 K. Unrolling carpet behavior was observed above 485 K. This carpet
Fig. 4. STM images after 30 min of heating. (a) At 581 K. Islands have spread along the step edges. New dots have spread along the surface (+1.0 V, 0.8 nA). (b) At 1010 K. Monolayer high terraces and islands (+1.0 V, 1.8 nA).
Please cite this article in press as: H. Bettermann, et al., Temperature-induced processes for size-selected metallic nanoparticles on surfaces, Appl. Surf. Sci. (2016), http://dx.doi.org/10.1016/j.apsusc.2016.04.161
G Model APSUSC-33162; No. of Pages 4
ARTICLE IN PRESS
4
H. Bettermann et al. / Applied Surface Science xxx (2016) xxx–xxx
Temperature induced surface diffusion leads to compact shapes. This effect is enhanced with increasing temperature. The melting behavior of iron-nickel nanoparticles on W(110) is in strong contrast to previous results concerning iron and cobalt. Cobalt and Iron0.5 Cobalt0.5 alloy particles do not exhibit unrolling carpet behavior at any temperature. Height reduction is only due to direct evaporation. A change in shape occurs only at very high temperatures (>1000 K), the particles are elongated along the tungsten’s [001] axis (see Fig. 6, [3]). This behavior can also be observed for iron particles. In addition iron has a critical temperature Tcrit = 640 K. Below that temperature evaporation without surface spreading is taking place. Tcrit is sufficient to cause a realignment from random to epitaxial orientation [1]. Unrolling carpet behavior can be observed above Tcrit [2]. Fig. 5. Development of nanoparticles and islands at different temperatures. 30 min of tempering each. Inset: Tempering at 458 K for different time intervals.
6. Conclusions We have shown that the melting point for iron-nickel nanoparticles is below 485 K. This temperature is close to the melting point of pure iron nanoparticles [1]. We could not observe any evaporation of material below that temperature. Unrolling carpet mechanism and surface diffusion are the driving processes to form the different shapes above the melting point. A slight anisotropy occurs during melting. References
Fig. 6. STM image of FeCo nanoparticles after heating for 1 h at 1000 K. The nanoparticles are spread along the [001] direction (+1.0 V, 0.1 nA).
is one monolayer high and can unroll even across terrace steps (see Fig. 3(a)). The time and temperature dependent development from nanoparticles to islands is shown in Fig. 5. The development to nanoparticles with a surrounding layer (unrolling nanoparticles) and subsequently to islands depends on the temperature and the duration of annealing. The islands initially formed have irregular shapes. This is in agreement with measurements of iron nanoparticles of similar size by Rosellen et al. [11]. Longer heating leads to compact island shapes. Accumulation of atoms at nearby step edges indicate that diffusion on substrate away from the nanoparticles is taking place. The remaining islands are elongated along the [001] axis. This is in good agreement with theoretical models for diffusion on bcc(110) (see Section 3).
[1] A. Kleibert, A. Voitkans, K.-H. Meiwes-Broer, Size-dependent alignment of Fe nanoparticles upon deposition onto W(110), Phys. Rev. B 81 (7) (2010), http:// dx.doi.org/10.1103/physrevb.81.073412. [2] W. Rosellen, C. Kleinhans, V. Hückelkamp, F. Bulut, A. Kleibert, J. Bansmann, M. Getzlaff, Influence of substrate and temperature on the shape of deposited Fe, Co, and FeCo nanoparticles, Phys. Status Solidi B 247 (5) (2010) 1032–1038, http://dx.doi.org/10.1002/pssb.200945569. [3] F. Bulut, W. Rosellen, M. Getzlaff, Structural properties of size-selected FeCo nanoparticles deposited on W(110), Appl. Phys. A 97 (1) (2009) 185–189, http://dx.doi.org/10.1007/s00339-009-5230-2. [4] W. Rosellen, H. Bettermann, T. Veltum, M. Getzlaff, Low energy impact of size selected FeCo nanoparticles with a W(110) surface, Physica E 44 (7-8) (2012) 1683–1686, http://dx.doi.org/10.1016/j.physe.2012.04.020. [5] R.-P. Methling, V. Senz, E.-D. Klinkenberg, T. Diederich, J. Tiggesbäumker, G. Holzhüter, J. Bansmann, K. Meiwes-Broer, Magnetic studies on mass-selected iron particles, Eur. Phys. J. D 16 (1) (2001) 173–176, http://dx.doi.org/10. 1007/s100530170085. [6] M. Bode, S. Krause, L. Berbil-Bautista, S. Heinze, R. Wiesendanger, On the preparation and electronic properties of clean W(110) surfaces, Surf. Sci. 601 (16) (2007) 3308–3314, http://dx.doi.org/10.1016/j.susc.2007.06.017. [7] M. Bode, R. Pascal, R. Wiesendanger, STM study of carbon-induced reconstructions on W(110): strong evidence for a surface lattice deformation, Surf. Sci. 344 (3) (1995) 185–191, http://dx.doi.org/10.1016/00396028(95)00873-x. [8] D. Reuter, G. Gerth, J. Kirschner, Anisotropic diffusion of 3 d metals on W(110): Competition between crystalline structure and surface steps, Phys. Rev. B 57 (4) (1998) 2520–2529, http://dx.doi.org/10.1103/physrevb.57.2520. [9] R. Butz, H. Wagner, Surface diffusion of Pd and Au on W single crystal planes, Surf. Sci. 87 (1) (1979) 69–84, http://dx.doi.org/10.1016/00396028(79)90170-5. [10] U. Köhler, C. Jensen, C. Wolf, A. Schindler, L. Brendel, D. Wolf, Investigation of homoepitaxial growth on bcc surfaces with STM and kinetic Monte Carlo simulation, Surf. Sci. 454-456 (2000) 676–680, http://dx.doi.org/10.1016/ s0039-6028(00)00151-5. [11] W. Rosellen, Strukturelle Eigenschaften massenselektierter 3d-Metall-Partikel, Ph.D. thesis, Universitäts-und Landesbibliothek der Heinrich-Heine-Universität Düsseldorf, 2013.
Please cite this article in press as: H. Bettermann, et al., Temperature-induced processes for size-selected metallic nanoparticles on surfaces, Appl. Surf. Sci. (2016), http://dx.doi.org/10.1016/j.apsusc.2016.04.161