Atoms of gold enter the circuit

Discoveries on nanowires drive research on chips for computers that work on a molecular scale

Theory and experience have rarely integrated themselves so well. On December 17th last year, three Brazilian scientists signed the cover story of Physical Review Letters, one of the most important magazines specialized in physics in the world. In four pages, they describe the discoveries on the behavior of gold nanowire atoms, structures that measure some billionth parts of a meter. They represent a strategic material for making components for the next generation of computers, which should take the place of the current, silicon-based, ones.

The results are based on data accumulated since October 20th, 1999, when researchers from the National Synchrotron Light Laboratory, in Campinas, adjusted the focus of a high resolution electronic microscope, capable of magnifying 1.2 million times, and observed for the first time the rupture of gold nanowires. Very probably, it is the first time that the magazine gives its most prominent space to the work of Brazilian researchers. In the four pages of the article, entitled How Do Gold Nanowires Break?, Edison Zacarias da Silva, from the University of Campinas (Unicamp), and Adalberto Fazzio and Antônio José Roque da Silva, both from the University of São Paulo, use computer simulation to show the formation and evolution of the structures that appear in the gold wire before and after breaking.

Described step by step, the sequence displays a detailing impossible to be achieved with the electronic microscope, given the very proportions of the atom, even one of gold, mid-size, with 79 electrons around the nucleus, and the limitations of the equipment. “We wanted to contribute towards the interpretation of the experiments and to understand the mechanisms involved inthe dynamic process that leads to the formation ofthe lines of atoms and, finally, to their rupture”, says Zacarias.

Getting to know these processes is important for a basic reason: due to its characteristics, above all its capacity for not reacting with oxygen and its being able to be stretched out without breaking, the so-called ductility, gold is regarded as the best material for forming the electrical contacts among the new devices to be created to replace the silicon chips that is today the basic material in computers at present. The prospects are based on the discoveries made in the 90’s that molecules are capable of conducting electricity in the same ways as wires and semiconductors themselves.

Today’s machines are now regarded as beings in extinction. “Silicon physics, which provides all the current computers, has its days numbered”, Fazzio comments. “Perhaps miniaturization based on silicon may last for another 10 or 15 years, but it will hardly go beyond that”. In an article published, also in December, in the Nanotechnology magazine, Ramón Campañó, the director-general of the European program for the development of nanocircuits, ponders that the so-called Moore Law, according to which the capacity of microprocessors doubles in periods of from 18 to 24 months, has been valid for 30 years, but has no way of carrying on.

Quickness and speed
One of the ways most adopted to increase quickness and speed in computers has been to shrink the size of the transistors, the units that process the information. Today, some 40 million transistors can be squeezed into a chip the size of a postage stamp. Each one of them is 60,000 times larger than a molecule – an indication that it will not be easy to go from microelectronics to nanoelectronics, in which information runs in a dimension of billionths of a meter, from one atoms to another (one nanometer corresponds to one billionth of a meter). Smaller is better, it is imagined. The chips from the next generation will have to be hundreds of times smaller than the present ones. The calculation is that their dimensions may range from 10 to 1,000 angstroms, at the most – an angstrom is one ten billionth of a meter, equivalent to the diameter of an average atom.

In contacts for the transmission of electric current, it is probable that the computer of the next few decades will adopt at least one of the forms of nanowire tips, which represents one of the principal scientific finds of the work of the researchers from USP and Unicamp: the French hat, as it was baptized by the researchers. It is an arrangement of atoms in a trapezium shape, with two hexagons (each point corresponds to one atom) at the sides and one at the base, which is reminiscent of a hat made from a folded sheet of a newspaper, used in past by children.

The physicists concluded that this is one of the most stable shapes for the gold nanowire, which is formed moments after rupture. This work puts into evidence the scientific value of computer simulation, the technique that has made it possible to analyze the behavior of the atoms. From the number crunching carried out on the computers of the National Center for High Performance Processing (Cenapad), in Campinas, and analyzed and visualized at the workstations in Unicamp and USP, the physicists were able, not only to reproduce and explain with admirable precision, the results that emerge from the direct observation of the atoms under the microscope, but they were also able to obtain new information on unprecedented forms of organization of atoms, like the French hat itself, which escapes experimental observation.

With these results, the physicists have opened up, simultaneously, a wide door for new computer simulation activities, which can also precede the experimental stage itself, and allow, among other things, to save in time and money: just like the chemists and biologists, who already avail themselves of computers for planning medicines, molecule by molecule, in the area of research that has been nicknamed virtual alchemy, physicists may soon feel comfortable to plan materials on the atomic scale.

Manipulating atoms and constructing molecules is the essence of nanotechnology, a vanguard area that is mobilizing the world. In physics, one of the main fronts for nanotechnology is precisely the replacement of silicon. In its place, according to research under way, carbon nanotubes and organic molecules like fullerenes or buckyballs can be used, molecules of a geodesic shape formed by 60 carbon atoms. The discoveries are picking up speed. In August last year, a researcher with IBM announced a semiconductor circuit made with carbon nanotubes. In November, Science magazine announced the development of transistors assembled with organic molecules united by gold tips.

The December article on nanowires was not the only one by the researchers from USP and Unicamp in Physical Review Letters, although it was the most prominent one. In 2001, the group led by Fazzio published another two works in the same magazine about electronic and structural properties of materials, besides almost 20 in international magazines, with the collaboration of researchers from Unicamp and the federal universities of Santa Maria (UFSM), in Rio Grande do Sul, and Uberlândia (UFU).

Other Brazilian discoveries about gold nanowires – built in ultrahigh vacuum chambers – will pop up in the next few months, once again in Physical Review Letters. In an article already accepted for publication, On the Origin of Anomalous Long Interatomic Distances in Suspended Gold Chains, Daniel Ugarte and Varlei Rodrigues, from the LNLS, in collaboration with Sérgio Legoas and Douglas Galvão, from Unicamp, explain why the distance between the atoms in gold nanowires is greater than in gold used, for example, to make jewelry.

According to the group’s calculations, carbon atoms can exist between the gold atoms in nanowires, so that the distance is greater than in common gold. The intruders would explain even extremely long distances, of 5 angstroms, while in distances between 3 and 3.6 angstroms, situations with carbon and without carbon may coexist, being clean nanowires and contaminated ones. In this case, everything is figured out mathematically, because there is no way of identifying carbon: as its mass is much lower than gold (its has only six electrons, and, in the nucleus, six protons), it is transparent for the electronic transmission microscope.

Since 1998, an explanation has been sought for interatomic distances that are larger than expected. That was the year when researchers from Japan discovered that the distance between gold atoms in nanowires, when one is lined up next to the other, moments before freeing themselves, may reach 3.6 angstroms – while in normal gold, as used by jewelers, it is 2.9 angstroms. In the following year, the team from the Center for High Resolution Electronic Microscopy of the LNLS, which has been working with nanowires of gold, silver and platinum since 1995, successfully reproduced the Japanese findings, and went further.

In an article published in 2000 in Physical Review Letters, the group from the LNLS showed for the first time that, before they break, gold nanowires take on only three shapes, each one of them with different mechanical components. Two of these shapes are ductile: the nanowires are easily stretched, like chewing gum. And one is brittle: the nanowire breaks easily when stretched. “These results were successfully related to properties of electrical resistance, not understood until then”, Rodrigues comments.

Mysteries are being cleared up, but the nanometric world remains intriguing, entirely unforeseeable. On this scale, nanowires no longer obey the classic Ohm’s Law, one of the basic pillars of microelectronics, according to which the intensity of an electric current varies in a linear manner, on a regular scale, in accordance with the diameter of the wire. “In the case of the nanowires”, Ugarte explains, “the current shows levels separated by leaps, that is, while we vary the diameter of the wire, the current stays at a fixed value (level), and suddenly leaps to another level”.

Cathedral Square
Using an analogy, it is as if the electrons were converted into people who have to cross a space and deliver a package, the electric current. In the first moment, which is equivalent to the microelectronic scale, and, more broadly, to any electric wire, the electron people move with difficulty amongst a multitude in Cathedral Square, in downtown São Paulo. The number of electron persons that manage to reach the other side depends on the size of the square: if it is larger, the number of packages carried will also grow, in a continuous manner. The entry of more package carriers is not even noticed – and there is always room for one more. In the second, on the nanometric scale, the electron-persons have to go down corridors, where they can go ahead freely.

One goes first, and then another, and so on. More than one electron-person can only go ahead at the same time if there are more corridors – and it is only possible to note that the quantity of packages carried has altered when the space grows exactly in the proportion of one corridor. The quantity of packages (electric current) increases in an discontinuous way, by leaps, as the electron-persons reach the corridors. The physicists call this going forward by leaps, already observed in experiments, quantization.

In the last few years other groups have tried to understand the behavior of atoms already organized in a line, moments before rupture, but they have not prospered. The physicists from USP and Unicamp preferred a full hand strategy, if you can call it that, in the case of such tiny objects. “We looked for a more realistic form, the nanowire itself, imagining that the atoms would somehow keep some similarity with the structure to which they belong”, says Roque da Silva.

The atoms, which are shown in reality – in the form of black pin-points – under the microscopes of the LNLS, are represented by tables of figures that come out of the computers and indicate their relative positions in the course of time. Only after all calculations have been done, when the work has been done, is it that they take on the undeniably more comprehensible form of colored dots.

With computer simulation, the physicists set off from a structure analogous to a gold crystal, with one atom at each vertex of a regular hexagon and another in the center – is the most compact configuration possible, with the greatest number possible of elements per unit of space. Each hexagon forms a plane. The packaging of these figures makes up the crystal, represented in the computer simulation by a global structure of ten planes, with 70 atoms.

Next, the researchers let the structure be reconstructed in accordance with the equations of quantum mechanics, with the atoms seeking spontaneously the positions with less energy. The redistribution of the atoms originates the nanowire, and extremely thin tubular structure, the surface of which, dense and compact, and made up of cells with seven atoms, constitutes the greatest atomic packaging possible – there is no other geometric form that makes it possible to put more atoms in the same space.

After giving the coordinates for the formation of the crystal, the only order that the researchers gave the computer was to apply a tractive force to the two ends, as if two hands were pulling the wire, at a temperature of roughly 300° C. This is how the stretching of the nanowire was simulated, which in the first moment avoids breaking at the expenses of its own filling: the atoms at the center jump to the surface. As a result, the length of the wire increases. “No one before had seen the wire become hollow”, Roque da Silva comments.

Lost symmetry
When all the atoms from inside have been used up, the nanowire finds itself at dead end: either it breaks, or it gets thin. It prefers to get thin and ceases to be a hexagon at one unspecified point, although it still tries to preserve triangular structures between the atoms. At this point, a new plane emerges, the eleventh, with only five atoms. A defect is formed there, a region of instability: it is where the wire is going to get thin and, later, to break. In the non-defective portion of the wire, symmetry ensured that the structure is maintained, in spite of the traction.

In the region of instability, though, deformation continues. According to the analyses carried out, the structure goes through various stages, until reaching the configuration with just one atom linking the two ends. At this stage, wires of metals like sodium would break, but not gold wire: the noble metal’s ductility gives extra life to the wire, with the critical region incorporating new atoms, pulled from the ends, in a linear sequence.

This atomic line, in which the atoms come to be 15% more distant than in the gold crystal, is sustained until they incorporate from 4 to 5 atoms. The distances have been checked, and, once again, they agreed with the experimental results. But the lined up atoms do not uphold themselves: the tensions become unbearable and the wire breaks. For the researchers, one of the concerns was to understand precisely the structures that form in the vicinity of the breaking point – they are the ones that make possible contact between the gold wire and the nanodevices and confer them the prospects for technological application.

This is the moment, right after the wire breaks, that the two resulting ends, with very symmetrical and stable configurations, form the structure called the French hat. According to Roque da Silva, this shape is so stable that no atom can, any longer, manage to leave the tip to enter the wire that is still growing. The finding is a suggestive one. “It may be that every time the wire is pulled, the French hat is formed at the ends”, he imagines.

Although the emphasis is currently on the study of phenomena on the atomic scale, computer simulation generally has wider applications. In one of the articles published last year, Roque da Silva, Fazzio, doctorate student Gustavo Dalpian and former doctorate student at USP, Anderson Janotti, simulate an experiment carried out on an electronic tunneling microscope, to understand how germanium atoms build up on the surface of silicon – once again, practice alone has not made it possible to understand some structures that form after this deposition. The results reached benefited both current microelectronics and the perfecting of solar cells, used as a source of electricity.

At the moment, the researchers are studying the carrying of an electric current between, for example, gold nanowires and carbon nanowires. Besides, of course, the discoveries that they may make on the behavior of the atoms, of a purely scientific nature, the application of the results also disquiets them. “We have algorithms to solve equations in quantum mechanics that give us confidence in the design of new materials on the computer and in planning cheaper experiments, with very reliable results”, Fazzio comments.

“Computer simulation can be decisive in the study of nanomaterials, not only in nanoelectronics, but also in nanochemistry and in nanobiology. Brazil lost the opportunity to become independent in microelectronics, produces almost no chips, but we still have some chance in nanotechnology, because nobody knows exactly what material is going to replace silicon. This is the moment to choose. “Evidently, fundamental science is not enough. This is the reason why Fazzio regards it as indispensable for engineers and companies to commit themselves to the nanoworld, as the scientific screenplay is ready.

In theory, using detailed knowledge of the geometry of the gold nanowire tips, one can think about applying a difference in electrical potential (more intense charges on one side and less on the other) between the wires. The device works like a one molecule transistor, in which just one electron passes (the smallest electrical charge) at a time. If the passage of the electron is associated with the number 1 and non-passage to number 0, the minimum conditions for the system to work a binary language will be created. Although understanding all the physics involved in the process is a long way off, its practical outcome can be imagined: a set of these devices can make up the future nanochip.

The four centers of the Brazilian nanonetwork

This years sees the beginning of operations for four research centers selected in December by the National Program of Nanosciences and Nanotechnology, created by the National Council for Scientific and Technological Development (CNPq) and launched in November 2000 to define the country’s course of activities in an area regarded as strategic. The first of them is to be found at the Federal University of Rio Grande do Sul (UFRGS), will be coordinated by Israel Baumvol, and incorporates the proposals of the LNLS and of the group from the Federal University of Minas Gerais; another is centered on Unicamp and led by Nélson Durán; and the other two are in the Federal University of Pernambuco (UFPE), coordinated by Eronides Felisberto da Silva Junior and Oscar Loureiro Malta.

The centers will work in a network, in conjunction with about 40 research institutes from Brazil and abroad, besides two companies (France Telecom and PQSD – Ponto Quântico Sensores e Densímetros), in activities that include the development of carbon nanotubes, ceramics, semiconductor materials, molecular sieves and medicines. They are objectives as wide-ranging as nanotechnology itself, an area that, to take an example, is forecasting the creation of adhesives that can stick point by point one surface onto another, and drugs that act on the organism with unimaginable precision.

“At least, we will not be spectators”, ponders Celso Pinto de Melo, the coordinator of the national program and a researcher at UFPE. “We have the conditions for going into the game.” According to him, Brazil should invest about R$ 3 million this year (a little more than US$ 1 million) and adopt a strategy similar to the German model, which has a budget in the order of dozens of millions of dollars and decided to create or to consolidate centers of excellence in the major lines of nanotechnology. The United States is now seeking leadership, with growing resources for research in nanotechnology: US$ 270 million in 2000, US$ 422 million in 2001, and US$ 520 million (still to be approved) in 2002.

Another Brazilian venture is the implementation of the National Reference Center for Nanotechnology. Cylon Gonçalves da Silva, who until July last year ran the National Laboratory of Synchrotron Light, maintained by the Ministry of Science and Technology, is the person in charge of the planning, which he intends to conclude before the end of this half year. According to him, the center will be acting in few areas, where research institutions and industry can be linked, so as to benefit the economic development of the country. “We have ability in uniting theoretical and experimental research”, he says. “The challenge is to make fundamental research and innovation go hand in hand.” With funds in the order of US$ 3 million this year for setting it up, the center should start to operate in 2003.

What physicists wish

Every physicist in the world wishes to see his research published in Physical Review Letters, or simply PRL, one of the most – if not the most – important scientific magazine in this area. But it is far from easy. Published by the American Physical Society (APS), of the United States, this magazine’s prestige rests on a rigorous selection of the articles. “The Physical Review Letters only accepts work that really represents significant advances”, reiterates José Roberto Leite, the of the Brazilian Physics Society (SBF). Since 1988, the rejection rate of works for publication oscillates between 37% and 41% a year. In 2001, 71 articles by researchers from Brazilian institutions have come out in the PRL.

Out of this total, 30 were from São Paulo.The magazine began to circulate on July 1st, 1958, with the objective of publicizing the results of research of general interest for physicists from any area – from atomistics to cosmology. It has not stopped growing ever since, it reached its record for articles published in one year in 1999 (2,800, chosen from the 7,650 received), and is getting ready to surpass itself this year, with a forecast of publishing 3,100 articles and 11,800 pages in the course of its 52 issues. Its impact rate – which gives the dimensions of the importance of a magazine in scientific circles, measured by the number of citations divided by the number of articles published – is 6.10, one of the highest in the area of physics.

It is better not to call the publication just the Physical Review, because of the other publications by the American Physical Society, which also publishes the Physical Reviews A, B, C, D and E, aimed at making public the results of research in even more specific areas, Physical Review Special Topics and the Review of Modern Physics.

The projects
High Resolution Electric Microscopy Center (96/04241-5); Modality: Infrastructure Program – Multi-user Equipment; Coordinator: Daniel Mário Ugarte – National Synchrotron Light Laboratory; Investment: R$ 5,075,635.07.
2. Studies of Materials: Electronic and Structural Properties (98/16536-5); Modality: Regular research benefit line; Coordinator: Adalberto Fazzio – Institute of Physics of USP; Investment: R$ 143,061.48.