Single-crystal metallic nanowires （from Nature Journal）

Nature 430, 61-65 (1 July 2004) | doi: 10.1038/nature02674

Single-crystal metallic nanowires and metal/semiconductor nanowire heterostructures

Yue Wu^1,3, Jie Xiang^1,3, Chen Yang¹, Wei Lu¹ and Charles M. Lieber^1,2

Top of page

Abstract

Substantial effort has been placed on developing semiconducting carbon nanotubes^{1, 2, 3} and nanowires⁴ as building blocks for electronic devices—such as field-effect transistors—that could replace conventional silicon transistors in hybrid electronics or lead to stand-alone nanosystems^{4, 5}. Attaching electric contacts to individual devices is a first step towards integration, and this step has been addressed using lithographically defined metal electrodes^{1, 2, 3, 4, 6, 7, 8}. Yet, these metal contacts define a size scale that is much larger than the nanometre-scale building blocks, thus limiting many potential advantages. Here we report an integrated contact and interconnection solution that overcomes this size constraint through selective transformation of silicon nanowires into metallic nickel silicide (NiSi) nanowires. Electrical measurements show that the single crystal nickel silicide nanowires have ideal resistivities of about 10 µ cm and remarkably high failure-current densities, >10⁸ A cm^-2. In addition, we demonstrate the fabrication of nickel silicide/silicon (NiSi/Si) nanowire heterostructures with atomically sharp metal−semiconductor interfaces. We produce field-effect transistors based on those heterostructures in which the source−drain contacts are defined by the metallic NiSi nanowire regions. Our approach is fully compatible with conventional planar silicon electronics and extendable to the 10-nm scale using a crossed-nanowire architecture.

Our focus on NiSi nanowires is motivated by previous investigations of metal silicides, which can exhibit low resistivity, compatibility with conventional silicon manufacturing, and the ability to form ohmic contacts to both p- and n-type silicon⁹. We prepared NiSi nanowires using an approach (Fig. 1a) involving deposition of nickel metal onto single-crystal Si nanowires^{10, 11}, solid-state reaction at 550 °C to form NiSi, followed by removal of remaining metal by wet etching (see Methods). Low-resolution transmission electron microscopy (TEM) studies (Fig. 1b) of materials prepared in this way show nanowires that have uniform diameters and contrast, which are properties indicative of single-crystal structures, and lengths of tens of micrometres. Analysis of samples prepared from Si nanowires with diameters of 20.3 2.3 nm yielded materials with diameters of 22.8 3.4 nm. The increase in average diameter agrees with the value of 22.4 nm, expected for conversion of Si to NiSi on the basis of different unit-cell volumes¹². In addition, energy dispersive X-ray spectroscopy (EDS) measurements show that the Ni:Si atom ratio in the nanowires is 1.03:1, as expected for NiSi.

Figure 1: Preparation and structural characterization of single-crystal NiSi nanowires.

a, Preparation of single-crystal NiSi nanowires. (1) Si nanowires (blue) of uniform diameter are (2) coated with Ni metal (green) to a total thickness comparable to the Si nanowire diameter, (3) reacted at 550 °C to form NiSi nanowires (brown), and then (4) etched to remove any excess Ni metal. b, TEM image of three NiSi nanowires prepared using Si nanowires with an average diameter of about 20 nm. The NiSi nanowires are highlighted by white arrows; scale bar is 1 µm. c and d, High-resolution TEM images of single-crystal NiSi nanowires. c, TEM image of a 20-nm NiSi nanowire prepared using Si nanowires of average diameter 20 nm. The white arrow indicates the growth front of the nanowire. Inset, two-dimensional Fourier transform of the image showing the [10] zone axis of NiSi. d, Image of a 32-nm NiSi nanowire prepared using Si nanowires of average diameter 30 nm. The white arrow indicates the growth front of the nanowire. Inset, two-dimensional Fourier transform of the image depicting the [20] zone axis of NiSi. The scale bars in c and d are 5 nm.

High resolution image and legend (98K)

High-resolution TEM studies (Figs 1c and d) demonstrate clearly that the nanowires are single-crystal NiSi. The reciprocal lattice peaks, which were obtained from two-dimensional Fourier transforms (2DFT) of the lattice-resolved images (insets to Figs 1c and d) can be indexed to the orthorhombic structure of NiSi¹² with the zone axes along the [10] and [20] directions, respectively. This analysis enables assignment of the growth front of NiSi to be the (11) and (001) planes for the 20- and 30-nm-diameter nanowires in Figs 1c and d, respectively. The different NiSi nanowire growth directions may be explained by different growth directions of the starting silicon nanowires¹¹, although future work will be required to quantify this relationship. These high-resolution data also show that the NiSi nanowires have smooth surfaces with very little (<>

Representative electrical transport data recorded on a 29-nm-diameter NiSi nanowire (Fig. 2a) show linear current (I) versus voltage (V) behaviour with two- and four-terminal resistances of 886 and 184 , respectively. The current levels observed from 30-nm NiSi nanowires typically exceed 100 µA at 100 mV bias, are quite remarkable in comparison to the starting Si nanowires, and are clearly indicative of the expected metallic nature of NiSi. Indeed, calculation of the resistivity on the basis of the four-terminal data yields a value of 9.5 µ cm, close to the value of 10 µ cm for NiSi single crystals¹³. These low-resistivity values have been observed for NiSi nanowires with diameters from about 15 to 45 nm, and thus demonstrate that these single-crystal wires can be scaled to ultrasmall dimensions without degradation of properties. Temperature-dependent measurements (inset to Fig. 2a) further show that the nanowire resistance decreases monotonically down to about 30 K and then saturates as expected for a metal. The similarity of NiSi nanowire resistivities to the bulk value, the temperature dependence of the resistance, and the scaling of resistance with length (to 6 µm) are indicative of diffusive transport. Using a previously reported carrier density value for NiSi¹³ we estimate the scattering mean free path to be of the order of 5 nm. Hence, it should be possible to retain the attractive metallic transport in NiSi nanowires down to the sub-10-nm-diameter scale, and also to study the effects of fundamental properties, such as dephasing, in a pure metal regime that should be accessible from recently reported molecular-scale silicon nanowires¹¹ using our synthetic approach.

Figure 2: Transport measurements on individual single-crystal NiSi nanowires.

a, Current versus voltage curves recorded on a 29-nm NiSi nanowire, with line (1) and line (2) corresponding to four- and two-terminal measurements, respectively, taken from the device shown in the SEM image (inset, upper left). The scale bar in the image is 1 µm. Lower inset, temperature-dependent normalized resistance obtained from a four-terminal device; the resistance, R, is normalized by the value at 200 K, R_o. b, I−V data recorded at large applied voltages. The rapid drop at about 1.8 V corresponds to the failure point of this nanowire. Inset, SEM image of the nanowire after breakdown. The image highlights the break close to the middle of the nanowires. Scale bar, 500 nm.

High resolution image and legend (56K)

We have also characterized the maximum transport current for the metallic NiSi nanowires. For example, the 29-nm-diameter nanowire discussed above could carry a current of 1.84 mA before failure (Fig. 2b). Notably, this yields a maximum current density, J_max, of 3 10⁸ A cm^-2. This large value of J_max is reproducible, and is approximately independent of nanowire diameter. For example, a nanowire of diameter 40−45 nm exhibited a failure current of 5.5 mA, and J_max > 3 10⁸ A cm^-2. In general, failure occurs in the middle of the nanowires, which is where the peak temperature is expected to occur, owing to dissipative self-heating; this suggests the breakdown mechanism is due to melting. The high J_max values can be attributed to the single-crystal structures, which preclude energy dissipation and void diffusion¹⁴ at the grain boundary and defect sites that are common in lithographically defined wires.

The J_max for the NiSi nanowires is comparable to the best values—10⁹ A cm ²— reported for single-walled carbon nanotubes¹⁵. We believe this finding is important because the high J_max values are achieved in essentially every NiSi nanowire, without the need to find and connect the metallic versus semiconducting carbon nanotubes. In addition, the maximum current can be scaled through the growth of specific-diameter NiSi nanowires. The J_max for the NiSi nanowires is also about two orders of magnitude larger than that observed in lithographically defined noble metal lines¹⁶. This suggests an immediate advantage of NiSi nanowires over lithographically defined nanoscale metal lines: for example, they could be used as local- and intermediate-level interconnects to nanowire field-effect transistors that might be key components in hybrid nanoelectronics.

Importantly, the approach outlined above can also be used to transform selectively single-crystal silicon nanowires to produce NiSi/Si heterostructures and superlattices. To demonstrate this concept (Fig. 3a) we patterned nickel metal onto silicon nanowires and then transform the nickel-containing regions to NiSi (see Methods). A dark-field optical image of a nanowire patterned in this way using 1-µm-wide nickel regions on a 2-µm pitch (Fig. 3b) exhibits periodic variations in contrast extending over the full length of the 65-µm-long nanowire. Analysis of the image shows that the average lengths of the Si and NiSi regions are both 1 µm, and in good agreement with the width and pitch of nickel metal deposited on the nanowire during fabrication.

Figure 3: Fabrication and structural characterizations of NiSi/Si nanowire heterostructures and superlattices.

a, Fabrication of NiSi/Si nanowire heterostructures and superlattices. (1) Si nanowires (blue) dispersed on a substrate are (2) coated with photoresist (grey) and lithographically patterned, (3) selectively coated with Ni metal (green) to a total thickness comparable to the Si nanowire diameter, and (4) reacted at 550 °C to form NiSi nanowires. b, Dark-field optical image of a single NiSi/Si nanowire heterostructure. The bright green segments correspond to silicon and the dark segments to NiSi. Scale bar is 10 µm. c, TEM image of a NiSi/Si heterostructured nanowire. The bright segments of the nanowire correspond to silicon and the dark segments, which are highlighted with arrows, correspond to NiSi. Scale bar is 1 µm. d, High-resolution TEM image of the junction between NiSi and Si showing an atomically abrupt interface. Insets, two-dimensional Fourier transforms of the image depicting the [110] and [111] zone axes of NiSi and Si, respectively, where the arrows highlight the growth fronts of the NiSi (221) and Si (112). Scale bar is 5 nm.

High resolution image and legend (103K)

TEM images of similar NiSi/Si nanowire heterostructures (Fig. 3c) show a similar periodic variation in contrast that is consistent with NiSi (dark) and Si (light) materials within the heterostructure. This assignment was confirmed by EDS analysis: the dark and light regions corresponded to an Ni:Si ratio of 1.02:1.00 and pure Si, respectively. Analysis of the images also shows that the average lengths of the Si and NiSi regions are 0.93 0.10 and 1.03 0.11 µm, respectively. The good agreement with the expected pattern demonstrates clearly that our approach enables spatially controlled transformation of silicon to metallic NiSi nanowire heterostructures. Notably, detailed examination of NiSi/Si heterostructure by high-resolution TEM (Fig. 3d) shows that the transformation yields an atomically abrupt interface (irrespective of any axial diffusion). The reciprocal lattice peaks, which were obtained from two-dimensional Fourier transforms of the lattice-resolved image (insets to Fig. 3d) correspond to the [110] and [111] zone axes of NiSi and Si, with (221) and (112) nanowire growth fronts, respectively.

The atomically sharp metal−semiconductor interfaces produced in these nanowire heterostructures have the potential to yield a range of precisely defined electronic devices and device arrays on individual nanowires. To explore this opportunity we have prepared field-effect transistor (FET) devices in which the critical source−drain regions are defined by metallic NiSi nanowire sections on p-type Si nanowire. Current versus source−drain voltage (V_sd) data (Fig. 4a) are linear to |V_sd| 1 V, which suggests that the NiSi/Si contacts behave for practical purposes as ohmic contacts at room temperature, although preliminary temperature-dependent measurements show a barrier at low temperature. The room-temperature behaviour can be explained by the reported segregation of dopant to the NiSi/Si interface during NiSi formation⁹, although other factors may also contribute to the ohmic response at room temperature. Notably, I−V_sd data recorded at different back gate voltages (V_g) exhibit the behaviour expected¹⁷ of a depletion mode p-FET with a high hole mobility of 325 cm² V s (see Methods).

To demonstrate that the active channel of this device was defined by the separation between NiSi nanowire regions, and not the much larger lithographically defined metal contacts, we carried out scanning gate microscopy (SGM) measurements. SGM images recorded with the scanning gate voltages of -9 V (Fig. 4c) and +9 V (Fig. 4d) respectively show enhanced (accumulation) and reduced (depletion) conductance only in the silicon region over the overall device (Fig. 4b). These data thus confirm that our approach yields spatially and electronically well-defined metal−semiconductor devices.

Figure 4: Transport properties of a NiSi/p-Si/NiSi heterojunction field-effect transistor.

a, I versus V_sd, exhibiting the typical characteristics of a NiSi/p-Si/NiSi heterojunction transistor. Inset, gate sweep obtained from the same device in the saturation regime with V_sd = -3 V. The device was fabricated using a 30-nm diameter p-Si nanowire (dopant density = 1 10¹⁸ cm^-3) on a heavily doped silicon substrate with 600 nm of thermal oxide; the channel length is 3 µm. b, SEM image of a field-effect transistor device fabricated using a NiSi/p-Si/NiSi nanowire heterojunction. Inset, dark-field optical image of the same device, where the bright green segment corresponds to silicon and the dark segments to NiSi. Scale bars, 3 µm. c and d, SGM images show reduced conductance with +9 V (Fig. 4c) and enhanced conductance with -9 V gate voltage (Fig. 4d) on the atomic force microscope tip. The dotted lines mark the interface between the NiSi and p-Si regions. Scale bars, 3 µm.

High resolution image and legend (107K)

Lastly, we have investigated scaling of our approach by assembling¹⁸ crossed-nanowire arrays in which the crossed nanowires function as masks defining active channels in NiSi/Si/NiSi heterostructures (see Methods) and could also function as local gates (Fig. 5a). TEM images recorded on a device following removal of the crossed-nanowire mask (Fig. 5b) show a well-defined silicon channel of 20 nm in this 10-nm-diameter NiSi/Si/NiSi heterostructure. This channel length is comparable to the best that can be achieved in state-of-the-art planar devices¹⁹. The TEM results indicate lateral diffusion of several nanometres during the formation of NiSi, and suggest that it should be possible to prepare shorter channel devices in a well-defined manner simply by varying the diameter of the nanowire mask. More generally, the capability of transforming Si to NiSi in a spatially well-defined manner to form NiSi/Si nanowire heterostructures and superlattices with atomically sharp metal−semiconductor interfaces opens up the possibility of integrating both active devices and high-performance interconnects from a single nanoscale building block. By extending our approach to crossed nanowires as shown above it should become possible to assemble large and dense arrays, perhaps using Langmuir−Blodgett assembly techniques²⁰, of transistors and other devices that could enable hybrid integrated circuits and could represent a key step towards stand-alone integrated nanosystems.

Figure 5: Self-aligned nanoscale NiSi/Si/NiSi devices.

a, Schematic illustrating Si nanowire (blue) crossed with three Si/SiO₂ core (blue)−shell (grey) nanowires²¹. Deposition, annealing and removal of excess Ni yields NiSi (brown) regions separated by Si in the nanowire. b, TEM image of the NiSi/Si/NiSi nanowire heterostructure. The dark regions correspond to NiSi and the light region to Si with NiSi/Si interfaces highlighted by black arrows. Scale bar is 10 nm. Inset, TEM image of the same nanowire before silicidation. The crossed Si/SiO₂ core−shell nanowire (approximately vertical in image) was used as a mask to define the Si region and removed after silicidation. Scale bar is 20 nm. The sample was prepared and imaged on a 50-nm-thick Si₃N₄ membrane.

High resolution image and legend (62K)

Top of page

Methods

Preparation of single-crystal metallic NiSi nanowires

Si nanowires were synthesized via chemical vapour deposition using monodisperse gold nanoclusters (Ted Pella) as catalysts, silane (SiH₄) as the vapour-phase reactant, diborane as the dopant and hydrogen as the carrier gas^{10, 11}. The silicon nanowires produced in this manner have clean surfaces with no visible amorphous oxide¹¹. Immediately following synthesis, the growth substrate with uniform diameter, free-standing Si nanowires was loaded into a metal deposition system, and then Ni metal was deposited to a thickness comparable to the average Si nanowire diameter. NiSi nanowires were produced by annealing the Ni-metal-coated Si nanowires at 550 °C. Excess Ni was completely removed by etching (TFG, Transene) for one hour at 50 °C, followed by post-annealing at 600 °C. Each of the annealing steps was carried out in forming gas (N₂:H₂ ratio, 90:10) for 5 min in a rapid thermal annealer (Heatpulse 610, Metron Technology).

Preparation of NiSi/Si nanowire heterostructures

Si nanowires dispersed in ethanol were deposited on a Si wafer with 600 nm of thermal oxide, and then the substrate was coated with photoresist (Shipley 1813, Rohm and Haas Electronics Materials). The photoresist was exposed for about 2 s on an ABM photoaligner using a simple striped pattern with a 2-µm pitch: 1 µm linewidth and 1 µm spacing. After developing for 1 min, the wafer was transferred to a thermal evaporator and Ni was evaporated with a thickness equal to the average nanowire diameter. After lift-off, the samples were annealed and etched as described above. Ultrasmall NiSi/Si/NiSi nanowire heterostructures were fabricated using crossed Si/SiO₂ core-shell nanowires²¹ as masks to define the lengths of the unreacted Si regions. The crossed-nanowire structures were assembled on Si₃N₄ membrane window grids (Structure Probe) by fluidic assembly¹⁸, and then Ni was evaporated and annealed as described above. The Si/SiO₂ core−shell nanowires were removed with hydrogen fluoride solution (Transene) to enable direct TEM imaging of the NiSi/Si/NiSi heterostructure on the Si₃N₄ membranes.

Calculation of device characteristics

The hole mobility, µ, is computed using a standard model¹⁷. In the small bias linear transport region, the mobility is expressed in terms of the transconductance, g_m, as g_m = µCV_sd/L², where V_sd is the source−drain voltage, L is the device channel length, and C is the gate capacitance, which was estimated using C = 2₀L/ln(2t_ox/r) (where is the effective dielectric constant, t_ox is the SiO₂ dielectric thickness, and r is the nanowire radius). For the device in Fig. 3a, g_m was evaluated for V_g from -3 to + 3 V at V_sd = −1 V, and has a value of 275 nS.

Top of page

Acknowledgments

We thank M. C. McAlpine, C. J. Barrelet and D. C. Bell for discussions. C.M.L. thanks the Defense Advanced Research Projects Agency and Intel for support of this work.

Competing interests statement:

The authors declared no competing interests.

Top of page

References

1. Yao, Z., Dekker, C. & Avouris, Ph. Electrical transport through single-wall carbon nanotubes. Top. Appl. Phys. 80, 147−171 (2001) | ISI | ChemPort |
2. McEuen, P. L., Fuhrer, M. S. & Park, H. Single-walled carbon nanotube electronics. IEEE Trans. Nanotechnol. 1, 78−85 (2002) | Article | ISI |
3. Dai, H. Carbon nanotubes: synthesis, integration, and properties. Acc. Chem. Res. 35, 1035−1044 (2002) | Article | PubMed | ISI | ChemPort |
4. Lieber, C. M. Nanoscale science and technology: Building a big future from small things. MRS Bull. 28, 486−491 (2003) | ISI | ChemPort |
5. Lieber, C. M. The incredible shrinking circuit. Sci. Am. 285, 58−64 (2001) | PubMed | ISI | ChemPort |
6. Cui, Y., Zhong, Z., Wang, D., Wang, W. U. & Lieber, C. M. High performance silicon nanowire field effect transistors. Nano Lett. 3, 149−152 (2003) | Article | ISI | ChemPort |
7. Javey, A., Guo, J., Wang, Q., Lundstrom, M. & Dai, H. Ballistic carbon nanotube field-effect transistors. Nature 424, 654−657 (2003) | Article | PubMed | ISI | ChemPort |
8. Heinze, S. et al. Carbon nanotubes as Schottky barrier transistors. Phys. Rev. Lett. 89, 106801 (2002) | Article | PubMed | ChemPort |
9. Morimoto, T. et al. Self-aligned nickel-mono-silicide technology for high-speed deep submicrometer logic CMOS ULSI. IEEE Trans. Electron Devices 42, 915−922 (1995) | Article | ISI | ChemPort |
10. Cui, Y., Lauhon, L. J., Gudiksen, M. S., Wang, J. & Lieber, C. M. Diameter-controlled synthesis of single-crystal silicon nanowires. Appl. Phys. Lett. 78, 2214−2216 (2001) | Article | ISI | ChemPort |
11. Wu, Y. et al. Controlled growth and structures of molecular-scale silicon nanowires. Nano Lett. 4, 433−436 (2004) | Article | ISI | ChemPort |
12. Toman, K. The structure of NiSi. Acta Crystallogr. 4, 462−464 (1951) | Article | ISI | ChemPort |
13. Meyer, B. et al. Intrinsic properties of NiSi. J. Alloys Compounds 262/263, 235−237 (1997) | Article |
14. Pierce, D. G. & Brusius, P. G. Electromigration: a review. Microelectron. Reliab. 37, 1053−1072 (1997) | Article | ISI |
15. Yao, Z., Kane, C. L. & Dekker, C. High-field electrical transport in single-wall carbon nanotubes. Phys. Rev. Lett. 84, 2941−2944 (2000) | Article | PubMed | ISI | ChemPort |
16. International Technology Roadmap for Semiconductors 2003 edn http://public.itrs.net/Files/2003ITRS/Interconnect2003.pdf (2003).
17. Sze, S. M. Physics of Semicondutor Devices 438−445 (John Wiley & Sons, New York, 1981)
18. Huang, Y., Duan, X., Wei, Q. & Lieber, C. M. Directed assembly of one-dimensional nanostructures into functional networks. Science 291, 630−633 (2001) | Article | PubMed | ISI | ChemPort |
19. Chau, R. et al. Silicon nano-transistors for logic applications. Physica E 19, 1−5 (2003) | Article | ISI |
20. Whang, D., Jin, S., Wu, Y. & Lieber, C. M. Large-scale hierarchical organization of nanowire arrays for integrated nanosystems. Nano Lett. 3, 1255−1259 (2003) | Article | ISI | ChemPort |
21. Lauhon, L. J., Gudiksen, M. S., Wang, D. & Lieber, C. M. Epitaxial core−shell and core−multishell nanowire heterostructures. Nature 420, 57−61 (2002) | Article | PubMed | ISI | ChemPort |

Silicon nanoparticles come in various sizes and colours

PROCESS for creating silicon nanoparticles, developed at the University of Illinois, has now been shown to produce a family of discrete particle sizes useful for microelectronics, optoelectronics and biomedical applications.

As reported in the journal Applied Physics Letters, researchers demonstrated that the electrochemically etched particles come in particular sizes. They also fluoresce in distinct colours. The smallest four sizes are blue, green, yellow and red luminescent particles.

"The availability of specific particle size and emission in the red, green and blue range makes the particles useful for electronic displays and flash memories," said Munir Nayfeh, a University of Illnois (UI) professor of physics .

``The benign nature of silicon also makes the particles useful as ultra-bright fluorescent markers for tagging biologically sensitive materials."

Current medical and biological fluorescent imaging is limited by the use of dye markers, which are not photo stable, Nayfeh said.

The dyes can break down under photo excitation, room light or higher temperatures.

Not only are the new silicon particles photo stable, they are also bright.

The light from a single nanoparticle can be readily detected.

To convert bulk silicon into nanoparticles, Nayfeh and his colleagues use an electrochemical treatment that involves gradually immersing a silicon wafer into an etchant bath of hydrofluoric acid and hydrogen peroxide while applying an electrical current.

The process erodes the surface layer of the material, leaving behind a delicate network of weakly interconnected nanostructures.

The wafer is then removed from the etchant and immersed briefly in an ultrasound bath.

Under the ultrasound treatment, the fragile nanostructure network crumbles into individual particles, which may be easily separated into the different size groups.

"The availability of different coloured markers is very important for biomedical applications," said Nayfeh.

``By placing particles of different colours in strategic locations, you could study such phenomena as growth factors in cancer cells or how proteins fold."

The silicon particles fluoresce when struck with ultraviolet light. They also can fluoresce when struck with two photons of infrared light — a technique that can non-invasively penetrate human tissue.

In a separate paper the researchers also demonstrated laser oscillation in small aggregates of the silicon nanoparticles.

``At 6 microns in diametre, these clusters of particles are one of the smallest lasers in the world," pointed out Sahraoui Chaieb, a UI professor of theoretical and applied mechanics.

This microlasing is an important step towards the realisation of a laser on a chip, which could ultimately replace wires with optical interconnects.

The emission was dominated by a deep-red colour, said Chaieb, who also is a researcher at the Beckman Institute. The clusters are currently stimulated by green light from a mercury lamp.

Star-Shaped Gold Nanoparticles Synthesized

By modifying the synthesis protocol for gold nanorods, scientists at Rice University in Houston have produced gold particles 100 nm across with multiple tips that display multidirectional polarized scattering characteristics and a high dielectric sensitivity. Reported in the March 28 online issue of Nano Letters, the star-shaped nanoparticles may have applications in sensors and as microscopic labels in biological research. Courtesy of Jason H. Hafner. ©2006, American Chemical Society.

Spectroscopy Bares Nanoparticle Structure

by Daniel C. McCarthy
Sometimes you need a bigger laser to study smaller particles. Specifically, a group of Dutch researchers applied a free-electron laser to study nanoparticles representing matter between the atomic and bulk states

Infrared spectroscopy demonstrated a marked change in the vibrational spectra of nanoclusters, whose cubic layers measured less than three atoms.

---

Conventional materials research has focused either on single atoms or on the bulk limit. But material properties vary dramatically between these two extremes, where a few atoms bond together to form clusters. The particles within this range are of increasing technological importance as semiconductors, biotechnology and other fields employ nanotechnology. However, clusters are not made easily and often dissolve quickly, making them difficult to study by conventional methods.
Researchers from the FOM Institute for Plasma Physics Rijnhuizen in Nieuwegein and the University of Nijmegen presented vibrational spectra for binary compounds containing between eight and 100 atoms, specifically niobium or tantalum carbide. They derived these spectra from resonance-enhanced multiphoton ionization, a technique that earlier yielded the first direct infrared spectra of gas-phase metal carbide clusters [see "IR Exposes Gas-Phase Metal Carbides," May 2000, pp. 42-43].
Although this is a well-established spectroscopic method, what sets the Dutch study apart is its application of a free-electron laser, which was tuned from 370 to 1650 cm²¹ to correspond to the vibrational frequency of the clusters. The spectrum this produced revealed a lot about the particles' structure.
The researchers found, for instance, that vibrational spectra of particles comprising more than 20 atoms corresponded to spectra of the bulk compound. This suggested that these nanocrystal structures resemble those of the bulk material and are just a cubic alternating arrangement of atoms. However, the spectra of species that comprised fewer than 20 atoms varied dramatically from the spectra of bulk material.
The scientists speculated that this discrepancy could be a finite size effect that occurs in systems whose cubic dimensions include three or more atoms. Whatever the cause, the work demonstrates the transition of vibrational spectra from the atomic to the bulk level.

Scanning Probe Microscopy and Atomic Force Microscopy, Recent Updates and an Overview of The Technology

Topics Covered

Background Brief History Basic AFM Components Scanning System Probe Probe Motion Sensor Controller Electronics Noise Isolation Computer Applications/Scanning Techniques Scanning Tunneling Microscopy Contact Mode AFM TappingMode AFM Non-contact Mode AFM LiftMode PhaseImaging Lateral Force Microscopy Magnetic Force Microscopy Force Modulation Electric Force Microscopy Surface Potential Imaging Electrochemical SPM Scanning Electrochemical Potential Microscopy Scanning Capacitance Microscopy and Scanning Spreading Resistance Microscopy Scanning Thermal Microscopy Tunneling AFM and Conductive AFM TRmode Nanoindenting Environmental Controls Recent Technology Advances Combined Environmental Controls Higher lateral resolution “Q”– Control Summary

Background

Scanning probe microscopes (SPMs) are instruments that measure properties of surfaces. They include atomic force microscopes (AFMs) and scanning tunnelling microscopes (STMs). In their first applications, SPMs were used mainly for measuring 3D surface topography and, although they can now be used to measure many other surface properties, that is still their primary application. SPMs are the most powerful tools of our time for surface metrology, measuring surface features whose dimensions range from interatomic spacing to a tenth of a millimeter. The main feature that all SPMs have in common is that the measurements are performed with a sharp probe operating in the near field, that is, scanning over the surface while maintaining a very close spacing to the surface. These instruments, specifically STMs, were the first to produce real-space images of atomic arrangements on flat surfaces. SPMs are now most commonly used to perform very precise, threedimensional measurements on the Ångstrom-to-micrometer scale. Table 1. Comparison and the characteristics of common microscopes. Optical Microscope SEM SPM Sample Operating Environment Ambient air, liquid or vacuum Vacuum* Ambient air, liquid or vacuum Depth of Field Small Large Medium Depth of Focus Medium Large Small Resolution: X, Y 1.0μm 5nm 2-10nm for AFM 0.1nm for STM Resolution: Z n/a n/a 0.05nm Effective magnification 1X – 2x103X 10X – 106X 5x102X – 108X Sample preparation requirement Little Little to substantial Little or none Characteristics required for sample Sample must not be completely transparent to light wavelength used Surface must not build up charge and must be vacuum compatible Sample must not have local variations in surface height >10 μm *Environmental SEMs operate at higher pressure and low eV, but resolution is sacrificed. Until the 1980s, researchers had relied upon other instruments for imaging and measuring the morphology of surfaces. Now in existence for over two decades, SPMs are the newest entry into the surface metrology field. As opposed to optical microscopes and electron microscopes (SEMs, TEMs), SPMs measure surfaces in all three dimensions: X, Y, and Z. Like SEMs, SPMs image and measure the surface of the sample. X and Y topographic resolution for most SPMs, including AFMs, is typically 2 to 10 nanometers (STM resolution can be as good as 0.1nm). Z resolution is about 0.1nm for a well-designed AFM or STM. Optical microscopes and SPMs are the easiest to use, with little or no sample preparation and no vacuum required. Optical microscopes and SEMs can have larger fields of view, but SPMs provide the highest magnifications and resolution in 3D. Furthermore only SPMs work on most samples with minimal sample preparation. Brief History The scanning tunneling microscope (STM) was the first SPM technology and was recognized as having atomic resolution capability in 1981. STM, in fact, still provides the best resolution available (Figure 1). The STM uses the tunneling current between tip and sample to image the sample surface. Unfortunately, there are some limitations, the most significant of which is that the surface of the sample must be conductors or semiconductors. This limits the materials that can be studied. Figure 1. STM image showing single-atom defect in iodine adsorbate lattice on platinum. 2.5nm scan. This limitation led to the invention in 1986 of the first atomic force microscope. The first commercially available AFM, the Digital Instruments NanoScope® was introduced in 1989. Like the STM, the AFM also uses a very sharp tip to probe and map the morphology of a surface. However, in AFM there is no requirement to measure a current between tip and sample. In this case, the tip is at the end of a micro-fabricated cantilever with a low spring constant. In contact mode AFM, the first AFM technique, the tip-sample force is held fixed by maintaining a constant and very low deflection of the cantilever, pushing the tip against the sample. This force can be in the range of interatomic forces in solids. Next, we describe the basics of AFM, including how the vertical motion of the tip is detected and transformed into topographic data. Basic AFM Components The basic AFM is relatively simple in concept (Figure 2a). Its closest predecessor is the stylus profiler. Figure 2. (a) Simplified diagram of a generic AFM. Photos show examples of (b) MultiMode SPM, (c) Dimension 3100 SPM, and (d) fully automated robotic Dimension X3D system for semiconductor applications. AFM technology uses sharper probes and lower forces than stylus profilers to provide higher resolution information without sample damage. A generic AFM comprises the following components: · Scanning System · Probe · Probe Motion Sensor · Controller Electronics · Noise Isolation · Computer Scanning System The most fundamental component of the AFM and the heart of the microscope is the scanner. Depending on the individual design, the scanner may scan (move) the sample (Figure 2b, MultiMode™ SPM) if the sample is small enough, or it may scan the probe over a larger sample (Figure 2c, Dimension™ 3100 SPM). To accomplish the precision required, a piezoelectric tube scanner is typically used in order to provide sub-Ångstrom motion control. Probe Another key component in the system is the probe. As mentioned above, the probe can be stationary and the sample can be scanned under it or the probe can be scanned over the sample. With today’s sophisticated technology, tip/cantilever assemblies that make up the probe (Figure 3) can be mass-produced with consistently shaped, very sharp tips. These tips are integrated into the end of cantilevers, which have a wide range of properties designed for a variety of applications. Figure 3. SEM micrograph of an etched single-crystal silicon AFM tip and tip/cantilever assembly Probe Motion Sensor This unit senses the force between the probe and the sample and provides a correction signal to the Z portion of the piezoelectric scanner (Figure 2a) to keep the force constant. The most common design for this function is called the optical beam deflection system, which is the lowest noise, most stable, and most versatile system available. This design uses a laser beam shining onto and reflecting off the back of the cantilever and onto a segmented photodiode to measure the probe motion. Controller Electronics This unit provides interfacing between the computer, the scanning system and the probe motion sensor. It supplies the voltages that control the piezoelectric scanner, accepts the signal from the probe motion sensor and contains the feedback control electronics for keeping the force between sample and tip constant. Noise Isolation To achieve the highest resolution, the microscope must be isolated from noise in its surroundings. There are very effective, yet simple systems for isolating AFMs from floor vibrations and from acoustic, electrical and optical noise sources. Computer Finally, scanning probe microscopy and AFM would not be feasible without the availability of powerful, high-speed PCs to drive the system and to process, display, and analyse the wealth of data produced. Applications/Scanning Techniques In its short lifetime, SPM has already added many variations to the fundamental scanning tunnelling theme. Once the AFM overcame the severe application limit of STM (the sample conductivity requirement), the variety of techniques and the range of applications began to mushroom. Although topographic mapping is still the dominant application for AFM (Figure 4), commercially available SPMs now provide some or all of the following techniques: Scanning Tunneling Microscopy Scanning Tunnelling Microscopy (STM) measures topography using the tunneling current between the probe tip and a conductive sample surface. Figure 4. Detailed topography of three defects – two protrusions and a depression – in a phase-shift photolithography mask. A cross section measures the smaller of the two protrusions (A) ~140nm across in the plane of the image. The depression defect (B) measures less than 6nm deep. 1.5μm scan. Contact Mode AFM Contact Mode AFM measures topography with the probe perpetually in contact with the sample. TappingMode AFM TappingMode AFM (patented) measures topography by lightly tapping the surface with an oscillating probe tip. Eliminates shear forces (present in contact mode). TappingMode is now the scanning mode of choice for most applications, particularly for softer surfaces like polymers. Non-contact Mode AFM Non-contact Mode AFM measures topography by sensing Van der Waals attraction between surface and probe tip. It is less stable than either contact or TappingMode. LiftMode LiftMode (patented) is a two-pass technique that separately measures topography and another selected property (magnetic force, electric force, etc.) using topographic information to track the probe tip at a constant distance above the surface. PhaseImaging PhaseImaging (patented) maps surface composition based on differences in local mechanical or adhesive properties of the sample. Lateral Force Microscopy Lateral Force Microscopy (LFM) maps frictional forces between the probe tip and the sample surface. The tip can be functionalized with chemical species for chemical force microscopy. Magnetic Force Microscopy Magnetic Force Microscopy (MFM) maps magnetic force gradient and distribution above the sample surface using LiftMode (Figure 5). Figure 5. AFM (a) and LiftMode MFM (b) images of pole tip region on magnetoresistive (MR) read/write head used in computer hard drives. MFM image shows domain structure and MR sensor that cannot be seen in the AFM topography. 12μm scan. Force Modulation Force Modulation (patented) maps relative stiffness of surface features. Electric Force Microscopy Electric Force Microscopy (EFM) maps electric field gradient and distribution above the sample surface using LiftMode. Surface Potential Imaging Surface Potential Imaging is one of the few AFM techniques that makes quantifiable maps of a quantity other than surface topography. Using LiftMode, it maps the distribution of surface electric potential of the sample. Recent applications include corrosion studies of alloys. Electrochemical SPM Electrochemical SPM maps topographic changes in-situ as induced by electrochemical reactions in electrolyte solutions simultaneously with electrochemical cell potential control (e.g. voltammetry). Can be performed with AFM or STM. Scanning Electrochemical Potential Microscopy Scanning Electrochemical Potential Microscopy (SECPM) (patented) is in-situ imaging or potential mapping of the electrode surface by measuring the potential difference between the potentiometric probe and the sample immersed in an electrolyte solution or a polar liquid (Figure 6). Figure 6. Scanning electrochemical potential microscope (SECPM). Scanning Capacitance Microscopy and Scanning Spreading Resistance Microscopy Scanning Capacitance Microscopy (SCM) and Scanning Spreading Resistance Microscopy (SSRM) both map 2D carrier (dopant) concentration profiles in semiconductor materials. Scanning Thermal Microscopy Scanning Thermal Microscopy (SThM) maps surface temperature distribution. Tunneling AFM and Conductive AFM Tunneling AFM and Conductive AFM measure tip-sample current for characterization of electrical conductivity and evaluation of thin film integrity. TRmode TRmode maps lateral forces and force gradients. Interleaves with TappingMode for complementary lateral and vertical characterization (Figure 7). Figure 7. TRmode is a technique that uses torsional oscillations of an AFM probe. Nanoindenting Nanoindenting measures mechanical properties and wear characteristics (hardness, adhesion, durability) of thin films, polymers, etc. (e.g. dielectrics, DLC). These techniques are being applied to a wide array of application areas, from biology to semiconductors, from data storage devices to polymers, and from integrated optics to measurement of forces between particles and surfaces. Other applications include MEMS fabrication, paints and coatings, metals/alloys/platings, plastics/polymers, biomaterials, biotechnology, food and food packaging, optics/optical films, optical disks, ceramics, thin films, liquid crystals, cosmetics, and geological and environmental studies. In addition, AFM systems have already been developed for highly specific applications, including automated robotic systems for handling semiconductor wafers (Figure 2d). They have also been developed with analysis routines designed for specific applications such as CD and DVD bump/pit measurements, as well as pole tip recession measurements for data storage read/write head manufacturing. These applications continue to expand. Environmental Controls AFM applications are carried out in a variety of environments. AFMs can be operated in ambient air, in vacuum, and in liquids (Figure 8). Biological measurements, in particular, are often carried out in-vitro in liquids. Electrochemical experiments are performed in liquid cells, allowing atomic-scale observation of electrochemical processes. In some cases, surface cleaning studies are done at atmospheric pressure in the controlled environment of a dry glove box. Figure 8. Condensed deoxyribonucleic acid (DNA) has been proposed as a gene delivery mechanism for biotechnology applications. Here, unfixed molecules were imaged in salt solution. 20μm scan. Recent new products include heating systems for biological and polymer applications up to 250°C (Figure 9), complete with sophisticated sample and environmental sensing. Systems are also now available for controlling the gaseous environment of the sample under study (Figures 10a and b). Figure 9. Successive phase images of poly(hexacyclodimethyl)siloxane at (a) 85ºC and (b) 90ºC. Heating induces formation of liquid islands within the amorphous polymers (a), which convert into arrays of small dots on additional heating (b). 10μm scans. Figure 10a. The Atmospheric Hood for the MultiMode SPM allows control of the gaseous imaging environment to vary humidity or image under inert gases. Figure 10b. The EnviroScope offers high vacuum, heating, electrochemical cell potential control, and purged gas environment. Recent Technology Advances New hardware and software have extended the utility of high-end SPM systems beyond measurement and characterization to include nanomanipulation and nanolithography. Examples of in-plane and out-of-plane nanomanipulation are shown in Figures 11a and 11b. An example of point-and-click nanolithography is seen in Figure 11c. Figure 11a. AFM in-plane nanomanipulation uses the AFM probe to image, manipulate nanometer-scale objects (carbon nanotubes), and image again to see the results. Figure 11b. AFM out-of-plane nanomanipulation uses the AFM probe to image, pull a single biomolecule out of the plane of the sample while measuring the unfolding of the molecule, and image again to see the results (in this case, the removal of one molecule from an array). Figure 11c. AFM nanolithography. New controllers and electronics (e.g., the NanoScope IV and IVa SPM controllers) have been designed to enhance performance relative to traditional designs. Some of the recent developments in AFM technology include: Combined Environmental Controls The latest generation of SPMs offer combinations of environmental controls, including vacuum and high temperature (Figure 12). Figure 12. Poly-sbs at room temperature in air (a) and at 180°C in 10-5 Torr pressure (b). Images captured with the Enviroscope (Figure 10b). Higher lateral resolution AFM systems now provide higher data density to allow zooming into the finest details, even on large scans. This provides the resolution required to characterize sidewalls on such samples as DVD bumps/pits and semiconductors. It also allows observation and measurement of nanoscale details on large scans — without the need to spend additional time re-scanning the sample with a smaller scan area (Figure 13). Figure 13. TappingMode+ height image and zoom of a copolymer. The square image is a zoom into the boxed area in the original rectangular image. This detail is revealed by simply zooming in with software and without the need for time-consuming, repetitive smaller scans. Without this higher resolution scanning, the zoomed image would not have the pixel resolution required to view nanoscale details. 10μm x 1.24μm scan and 1μm x 1μm zoom. “Q”– Control Controlling the quality factor, or Q, of the oscillating AFM probe allows better control of the forces between tip and sample and improves the sensitivity of measurements such as with PhaseImaging and MFM (Figure 14). Figure 14. Images of the same area on magnetic recording tape scanned with and without Q-control. Phase detection MFM images and average cross-section measurements of the probe phase shift illustrate nearly 4x enhanced signal-to-noise ratio for the Q-controlled image. 15μm scans. Summary Scanning tunneling microscopy produced dramatic images of atomic lattices and atomic force microscopy broadened the technology to non-conductive surfaces. Development of atomic force microscopes has allowed scientists and engineers to see structure and detail with unprecedented resolution and without the need for rigorous sample preparation. Several advances have further extended this technique’s utility to a wide range of applications. TappingMode permits imaging of soft materials without damage to the sample and LiftMode allows separate but simultaneous imaging of topography and other parameters, such as magnetic or electric forces, without cross-contamination. PhaseImaging has opened up the capability for mapping of surface compositional variations. New scanning and measurement technologies have expanded the range of measurements and thus further increased the utility of AFM for a broad variety of applications. These developments have taken AFM, in a few short years, from a laboratory curiosity to one of the most powerful, flexible, and widely used technologies for surface characterization.

Source: Veeco Instruments