Dear Readers,

This E-nano Newsletter issue contains two reportsproviding insights in two relevant areas: modellingat the nanoscale and scientific EU policy.

The nanoICT position paper provides an analy-sis of the current status of modelling at the na-noscale information processing and storagedevices in Europe and a comparison with thatin the rest of the world. The previous versionof the present paper (resulting from the 2008nanoICT Working Group meeting) has beenupdated reflecting the new issues that havebeen recognised as relevant and of significantcurrent interest.

The EU report summarises the discussionsthat took place end of October 2009 du-ring a "Future and Emerging Technologies"(FET) expert consultation workshop andgives a list of proposed challenges for the2011-2012 FET work programme.

We would like to thank all the authorswho contributed to this issue as well asthe European Commission for the finan-cial support (project nanoICT No.216165).

Dr. Antonio CorreiaEditor - Phantoms Foundation

Deadline for manuscriptsubmissionIssue No. 18: March 30, 2010Issue No. 19: June 30, 2010

Depósito legal/Legal Deposit:M-43078-2005

Impresión/Printing:Madripapel, S.A.

No. 16November 2009Published by PhantomsFoundation (Spain)

EditorDr. Antonio Correia antonio@phantomsnet.netAssistant EditorsJosé Luis Roldán, Maite Fernández and Carmen Chacón

1500 copies of this issue have been printed.Full color newsletter available at:www.phantomsnet/Foundation/newsletters.phpFor any question please contact the editor at:antonio@phantomsnet.net

Editorial BoardAdriana Gil (Nanotec S.L., Spain), Christian Joachim(CEMES-CNRS, France), Ron Reifenberger (PurdueUniversity, USA), Stephan Roche (CEA-INAC,France), Juan Jose Saenz (UAM, Spain), Pedro A. Se-rena (ICMM-CSIC, Spain), Didier Tonneau (CRMCN-CNRS, France) and Rainer Waser (Research CenterJulich, Germany)

E Editorial Information

5

29

32

nanoICT researchStatus of Modelling for Nanoscale In-formation Processing and Storage De-vices.

nanovacancies

EU reportAtomic & Molecular Scale Devices andSystems and Bio-Chemistry Based In-formation Systems.

ContentsC

3

2nd Position Paper for nanoICT Theory and Modelling Working Groupa5

nanoICT research

SSttaattuuss ooff mmooddeelllliinngg ffoorr nnaannoossccaallee iinnffoorrmmaattiioonnpprroocceessssiinngg aanndd ssttoorraaggee ddeevviicceessaa

M. Macucci1, S. Roche2, A. Correia3, J. Greer4, X.Bouju5, M. Brandbyge6, J. J. Saenz7, M. Bescond8, D.Rideau9, P. Blaise10, D. Sanchez-Portal11, J. Iñiguez12,G. Cuniberti13 and H. Sevincli13

1Università di Pisa, Italy; 2CEA-INAC, France; 3PhantomsFoundation, Spain; 4Tyndall National Institute, Ireland;5CNRS-CEMES, France; 6Technical University of Denmark,Denmark; 7Universidad Autónoma de Madrid, Spain;8CNRS-L2MP, France; 9STMicroelectronics, France; 10CEA-LETI-MINATEC, France; 11CSIC-UPV/EHU, Spain; 12Institutode Ciencia de Materiales de Barcelona (CSIC), Spain; 13Tech-nical University Dresden, Germany.

During the 2009 meeting of the Theory and Modellingworking group the current status of modelling for na-noscale information processing and storage devices hasbeen discussed and the main issues on which collabora-tion within the modelling community is needed havebeen pointed out. An analysis of the current situation inEurope in comparison with that in the rest of the worldhas been performed, too. The previous version of thepresent report (resulting from the 2008 meeting) hasbeen updated reflecting the new issues that have beenrecognized as relevant and of significant current inte-rest.

Introduction

We are presently witnessing the final phase of thedownscaling of MOS technology and, at the same time,the rise of a multiplicity of novel device concepts basedon properties of matter at the nanoscale and even atthe molecular scale.

Ultra-scaled MOS devices and nanodevices relying onnew physical principles share the reduced dimensionalityand, as a result, many of the modelling challenges. In ad-dition, new materials and process steps are being inclu-ded into MOS technology at each new node, to be ableto achieve the objectives of the Roadmap; these chan-ges make traditional simulation approaches inadequatefor reliable predictions. So far modelling at the nanos-cale has been mainly aimed at supporting research andat explaining the origin of observed phenomena.

In order to meet the needs of the MOS industry and tomake practical exploitation of new device and solid-stateor molecular material concepts possible, a new integra-ted approach to modelling at the nanoscale is needed,as we will detail in the following. A hierarchy of multi-scale tools must be set up, in analogy with what alreadyexists for microelectronics, although with a more com-plex structure resulting from the more intricate physicalnature of the devices.

A coordinated effort in the field of modelling is apparentin the United States, where significant funding has beenawarded to the Network for Computational Nano-technology, which is coordinating efforts for the deve-lopment of simulation tools for nanotechnology ofinterest both for the academia and for the industry.

Although the required integrated platforms need to bedeveloped, the efforts made in the last few years by themodelling community have yielded significant advancesin terms of quantitatively reliable simulation and of ab-initio capability, which represent a solid basis on whicha true multi-scale, multi-physics hierarchy can be built.The combination of these new advanced software toolsand the availability of an unprecedented and easily ac-cessible computational power (in particular consideringthe recent advances in terms of GPU-based general pur-pose computing) make the time ripe for a real leap for-ward in the scope and performance of computationalapproaches for nanotechnology and nanosciences.

Current status of MOS simulation and industrial needs

The continuous downscaling of MOSFET critical dimen-sions, such as the gate length and the gate oxide thick-ness, has been a very successful process in currentmanufacturing, as testified, e.g., by the ITRS require-ments. However, conventional scaling down of MOSFETchannel length is declining as the physical and economiclimits of such an approach are coming closer. Novel so-lutions are increasingly being used in MOSFET channelengineering within the industry.

Among the new technological features of very advan-ced devices, high-k dielectrics, the archetype of which ishafnium oxide, can significantly reduce gate leakage. Me-chanical strain applied in the channel and substrateorientation can also significantly improve carrier mobi-

lity. Moreover, alternative geometries, such as double-gated devices, in which the channel doping level is rela-tively low, must be evaluated within the perspective ofan industrial integration. In particular, the subsequenteffects of the high-k gate dielectric and of the double-gate geometry on channel mobility must be clearlyquantified.

Technology Computer-Aided Design (TCAD) refers tothe use of computer simulations to develop and opti-mize semiconductor devices. State-of-the-art commer-cial TCAD device simulators are currently working usingthe Drift-Diffusion (DD) approximation of the Boltz-mann transport equation. Quantum effects are accoun-ted for using the Density Gradient approximation, thatworks well for traditional bulk devices, but that can beunreliable for advanced devices such as the double-gated-MOS structure or for new materials. Moreover,emerging materials also significantly challenge the con-ventional DD-based tools, mostly due to a lack of ap-propriate models and parameters. It becomes urgent todevelop new physically-based models with a view of in-tegrating them into a standardized simulation platformthat can be efficiently used in an industrial environment.For this purpose, tight collaborations between world-class universities and research institutions, CAD vendorsand industrial partners must be established. Within theframework of these collaborations, there will be thebest chances of success, both in terms of academicmodel development and theoretical achievements, butalso in terms of concrete implementations and bench-marks of new models in TCAD tools. Innovative con-cepts based on nano-materials or molecular devices,new models and simulation tools would provide our ICTindustries a competitive advantage for device develop-ment and optimization in terms of time-cycle and wafer-costs.

Commercial v.s. academic quantum-transport solvers

In response to the industrial need of new simulationtools, a class of quantum and transport solvers is emer-ging. These commercial state-of-the-art solvers can bedivided into two categories. In the first category, onecan find the quantum-transport solvers, such as thosebased on the Non-Equilibrium Green Function Method,in which carrier transport is treated using the full quan-tum Green function formalism. In the second category

one can include the Monte Carlo Solvers, that modelcarrier transport via the Boltzmann equation. This equa-tion is solved in a stochastic way, using a classical des-cription of the free fly of the electrons but a quantumdescription of the interactions. The currently availablehigh-level NEGF [1] and MC solutions [2] are still in thedevelopment phase, and no ready-to-use industrial so-lutions are available so far to meet the requirements ofthe 32 nm node and beyond.

From the point of view of technology development sup-port, the Monte Carlo simulators should be able to pro-vide reliable electrical results on a regular basis for 32nm MOS devices. However the need for full-bandMonte-Carlo codes together with band structure sol-vers that account for strain and are capable of dealingwith new materials must be highlighted. Indeed, somecommercial 3D Schroedinger solvers [1] combined withNEGF solvers start being available. These solvers can beused to model ballistic quantum transport in advanceddevices with strong transverse confinement. However,they do not include any inelastic scattering mechanism,and thus are not suitable for the calculation of transportproperties in the 32 nm node devices and near-futurenodes.

On the other hand, high-level device simulation toolsare at an early stage of development in universities andresearch institutions. These codes generally include ad-vanced physical models, such as strain-dependent bandstructure and scattering mechanisms, and should pro-vide accurate predictions in complex nano-systems. Ho-wever, such simulation tools are in general difficult touse in an industrial environment, in particular because ofa lack of documentation, support and graphical user in-terface, although an increasing number of academiccodes are now including graphic tools [3,4]. Taking ad-vantage of these ongoing research projects, it should bepossible to integrate such high-level codes into industrialTCAD tools or to use them to obtain calibrated TCADmodels useful for the industry. Concerning this latterpoint, the quantum drift-diffusion-based solution mustbe “customized”, in order to make fast and accurate si-mulations of advanced devices possible. For instance,the effect of the high-k gate dielectric stack on deviceperformance must be addressed with a particular at-tention to its impact on carrier transport properties.This is definitely one of the most challenging issues insemiconductor industry at present.

nanoICT research

6

Efficient modelling tools, as well as accurate physics high-lights, would certainly bring a significant competitive ad-vantage for the development and the optimization ofthe 32 nm CMOS technology and for future technolo-gies, including molecular devices.

Importance of modelling variability

Near the end of the current edition of the InternationalTechnology Roadmap for Semiconductors (ITRS) in2018, transistors will reach sub-10 nm dimensions [5]. Inorder to maintain a good control of the electrical cha-racteristics, new transistor architectures have to be de-veloped. It is widely recognized that quantum effectsand intrinsic fluctuations introduced by the discretenessof electronic charge and atoms will be major factors af-fecting the scaling and integration of such devices as theyapproach few-nanometer dimensions [6-11].

For instance, in conventional one-gate nanotransistors,

variations in the number and position of dopant atomsin the active and source/drain regions make each nano-transistor microscopically different from its neighbours[12-16]. In nanowire MOS transistors the trapping of onesingle electron in the channel region can change the cur-rent by over 90% [17,18]. Interface roughness of theorder of 1-2 atomic layers introduces variations in gatetunnelling, quantum confinement and surface/bulk mo-bility from device to device. The inclusion of new ma-terials such as SiGe will induce additional sources offluctuations associated with random variations in thestructure, defects, strain and inelastic scattering [19,20].These intrinsic fluctuations will have an important im-pact on the functionality and reliability of the corres-ponding circuits at a time when fluctuation margins areshrinking due to continuous reductions in supply voltageand increased transistor count per chip [7,8].

The problem of fluctuations and disorder is actuallymore general and affects fundamental aspects of infor-

Advertisement

8

nanoICT research

mation storage and processing as device size is scaleddown. The presence of disorder limits the capability ofpatterning by introducing a spatial variance: when thepattern size approaches the spatial variance, patternsare unavoidably lost. An analogous problem exists as aresult of time fluctuations (shot noise) associated withthe granularity of charge: as current levels are reduced,the signal power decreases faster (quadratically with cu-rrent) than the shot noise power (linear with current),leading to a progressive degradation of the signal-to-noise power ratio.

Disorder has demonstrated all of its disruptive poweron nanodevices in the case of single-electron transistors:as a result of their extreme charge sensitivity, stray char-ges, randomly located in the substrate, are sufficient tocompletely disrupt their operation.

Fluctuations associated with the granularity of chargeand spatial disorder are fundamental roadblocks that af-

fect any effort towards handling information on an in-creasingly small scale.

It is thus of strategic importance to develop device si-mulation tools that are capable of efficiently exploringthe extremely large parameter space induced by suchvariability and evaluate the actual performance limits ofnew nanodevices. Strategies to decrease the amount ofnaturally occurring disorder or to cope with it need tobe devised as emerging devices are developed into newtechnologies aiming at the limits of the downscaling pro-cess.

Integration between material and device simulation

Both for decananometric MOSFETs and for most emer-ging devices, the distinction between material and de-vice simulation is getting increasingly blurred, becauseat low dimensional scales the properties of the material

Advertisement

9

nanoICT research

sharply diverge from those of the bulk or of a thin filmand become strongly dependent on the detailed devicegeometry. Such a convergence should start being re-flected also in research funding, because, at the dimen-sional scale on which research is currently focused, aproject cannot possibly take into consideration only oneof these two aspects. This was not the case until a fewyears ago, when a material could be investigated withinthe field of chemistry or material science and then pa-rameters were passed on to those active in the field ofdevice physics and design, who would include them intheir simulation tools.

Fortunately, the nanoelectronics simulation communityis not starting from scratch in terms of atomic scale ma-terials computations. Computational physics and quan-tum chemistry researchers have been developingsophisticated programs, some with on the order of mil-lions of lines of source code, to explicitly calculate thequantum mechanics of solids and molecules from firstprinciples. Since quantum mechanics determines thecharge distributions within materials, all electrical, opti-

cal, thermal and mechanical properties, in fact any physi-cal or chemical property can in principle be deducedfrom these calculations.

However, these programs have not been written withnanoelectronics TCAD needs in mind, and substantialtheoretical and computational problems remain beforetheir application in process and device modeling reachesmaturity. However, the coupling of electronic structuretheory programs to information technology simulationsis occurring now, and there is nothing to suggest thistrend will not continue unabated.

Quantum electronic structure codes come in essentiallytwo flavours: those using plane wave expansions (orreal-space grids) and those using basis sets of atomic or-bitals to span the electronic wave functions. Plane wavecodes are suitable for solid state calculations and havebeen mainly developed within the Solid State Physicscommunity. Codes using atomic orbitals were initiallydeveloped within the Quantum Chemistry community,although recently have also become popular within the

Fig. 1: Codes such as SIESTA can be used for the ab-initio simulation of relatively large structures.

11

nanoICT research

Materials Physics community. Quantum Chemistrycodes rely heavily on the expansion of the atomic orbi-tals in terms of Gaussian functions. This is mainly due tothe fact that, with the use of Gaussians, the four-cen-ter-integrals associated with Coulomb and exchangeinteractions become analytic and easy to calculate. Ho-wever, within the framework of Density Functional The-ory, due to non-linear dependence of the exchange -correlation energy and potential with the density, theevaluation of such contributions to the energy and Ha-miltonian has to be necessarily performed numericallyand the advantage of using Gaussian functions is mainlylost. This has opened the route for the use of othertypes of localized basis sets optimized to increase theefficiency of the calculations. For example, localizedbasis orbital can be defined to minimize the range ofthe interactions and, therefore, to increase the efficiencyof the calculation, the storage, and the solution of theelectronic Hamiltonian [21]. In particular, so-calledorder-N or quasi-order-N methodologies have been de-veloped over the last two decades that, using the ad-vantages of such local descriptions, allow for thecalculation of the electronic structure of very largesystems with a computational cost that scales linearlyor close to linearly with the size of the system [22,23].This is in contrast with traditional methods that typicallyshow a cubic scaling with the number of electrons in thesystem. Order-N schemes are particularly powerful androbust for insulators and large biomolecules. However,the design of efficient and reliable order-N methods formetallic systems is still a challenge.

The most widely used codes for ab initio simulations ofsolids and extended systems rely on the use of the Den-sity Functional Theory, rather than on Quantum Che-

mistry methods. Many of them have been developed inEurope, and some of them are commercial, althoughtheir use is mostly limited to the academic community.Among the commercial code, we can cite the plane-wave codes VASP [24] and CASTEP [25], and the Har-tree-Fock/DFT code CRYSTAL [26] that utilizesGaussian basis functions. Other widely used plane-wavecodes are the public domain CP2K [27] and Abinit [28].Abinit is distributed under GNU license and has becomea very complete code with a rapidly growing communityof developers. Among the most popular DFT codesusing local atomic orbitals as a basis set we can mentionthe order-N code SIESTA [21,29], which uses a basis setof numerical atomic orbitals, and Quickstep [30], thatuses a basis set of Gaussian functions.

One of the reasons why codes using basis sets of atomicorbitals have recently become very popular is that theyprovide the ideal language to couple with transportcodes based on Non-Equilibrium-Green’s functions(NEGFs) to study transport properties in molecularjunctions and similar systems. Using the local languageimplicit behind the use of atomic orbitals (with tight-bin-ding-like Hamiltonians) is trivial to partition the system indifferent regions that can be treated on different fotings.For example, it becomes relatively simple to combine in-formation from a bulk calculation to describe the elec-trodes with information obtained from a simulation thatexplicitly considers the active part of the device. AgainEurope has taken the lead along this path. Two of themost popular simulation tools for ballistic transportusing NEGFs combined with DFT have been developedbased on the SIESTA code: tranSIESTA [31] and Smea-gol [32]. In particular, tranSIESTA was developed by acollaboration of Danish and Spanish research groups

Fig. 2: With the help of ab-initio calculations, the effect of functionalization on graphene nanoribbon conductance can be evaluated.

12

nanoICT research

and, although there is a public domain version that willbe distributed with the latest version of the SIESTA code(siesta.3.0), it has also given rise to a commercial simu-lation package called QuantumWise [33].

Electronic structure theory represents the lowest andmost sophisticated level of computation in our simula-tion hierarchy. At this level there are many different de-grees of rigor and associated errors in thecomputations. The most accurate results are providedby post-Hartree-Fock Quantum Chemical methods. Un-fortunately, they are extremely demanding and not wellsuited for simulations of solids and condensed mattersystems. As already mentioned, the most popular ap-proach to study such systems is Density Functional The-ory, which provides a good balance between thecomputational cost and the accuracy of the results. Ho-wever, even at the DFT level, ab initio calculations arecomputationally too demanding to perform realistic si-mulations of devices. Therefore, it is necessary to de-velop more approximate methods and, finally, tocombine them in the so-called multi-scale approaches, inwhich different length scales are described with diffe-rent degrees of accuracy and detail. An interesting in-termediate stage between DFT calculations, which takeinto account the full complexity of the materials, andempirical models, which disregard most of the che-mistry and structural details of the system, is providedby the tight-binding approaches. Here a minimal basisof atomic orbitals is used to describe the electronicstructure of the system. As a consequence, only the va-lence and lowest lying conduction bands can be accura-tely treated. Traditionally, the hopping and overlapmatrix elements were considered empirical parametersthat were adjusted to describe the electronic band struc-ture of the material and its variation with strain. Thishas proven a quite powerful approach to describe com-plex system like, for example, quantum dots containingthousands or millions of atoms [34, 35]. A very interes-ting variation of these methods has been developed inrecent years: the so-called DFT-tight-binding. Here thetight-binding parameters (hoppings, overlaps and short-ranged interatomic repulsive potentials) are not obtai-ned by fitting to empirical information but they areobtained from DFT calculations for simple systems(mainly diatomic molecules) [36]. The parameters ob-tained in this way have proven to be transferable enoughto provide a reasonable description of systems like largemolecules and even solids.

The ability to treat varying length and time scales, withinvarying degrees of approximation, leads to the alreadymentioned requirement for a multi-scale approach tocoupled materials/device simulation. Although a multi-scale approach is more than ever needed at this stage,parameter extraction cannot be performed for a gene-ric material, but must be targeted for the particular de-vice structure being considered, especially forsingle-molecule transistors. There has to be a closedloop between the atomistic portion of the simulationand the higher-level parts, guaranteeing a seamless in-tegration. This convergence between material and de-vice studies also implies that a much moreinterdisciplinary approach than in the past is needed,with close integration between chemistry, physics, engi-neering, and, in a growing number of cases, biology. Tomake an example, let us consider the simulation of a si-licon nanowire transistor: atomistic calculations are ne-eded to determine the specific electronic structure forthe cross-section of the device being investigated, thenthis information can be used in a full-band solver fortransport or parameters can be extracted for a simplerand faster transport analysis neglecting interband tun-neling; then the obtained device characteristics can beused for the definition of a higher-level model useful forcircuit analysis. It is apparent that, for example, the ato-mistic simulation is directly dependent on the device ge-

Fig. 3: Ultra-high-vacuum STM image of a substrate after the fabrication withsingle-atom manipulation, of a complete interconnection circuit for an ORgate, obtained by removing hydrogen atoms from a silicon hydride surface.

13

nanoICT research

ometry, and that, therefore, work on the different partsof the simulation hierarchy has to be performed by thesame group or by groups that are in close collaboration.Indeed, the dependence between results at differentlength scales sometimes requires the combination of dif-ferent techniques in the same simulation, not just theuse of information from more microscopic descriptionsto provide parameters for mesoscopic or macroscopicmodels. Following the previous example, the electronicproperties of a silicon nanowire are largely determinedby its geometry. However, the geometry of the wire isdetermined by the strain present in the region where itgrows, which in turn is a function of the whole geo-metry of the device, its temperature and the mechani-cal constraints applied to it. Thus, a reliable simulationof such device can require, for example, a quantum me-chanical description of the nanowire itself, coupled to alarger portion of material described using empiricalinteratomic potentials, and everything embedded in aneven much larger region described using continuumelastic theory. Although still at its infancy, this kind ofmulti-scale simulations has already been applied verysuccessfully to the study of quite different systems, likebiological molecules, crack propagation in mechanicalengineering or combustion processes [37, 38]. This iscertainly one of the most promising routes towards thesimulation of realistic devices.

Another example where multi-scale and multi-physicssimulations become essential is represented by the ef-fort to merge electronics with nanophotonics. The in-tegration of CMOS circuits and nanophotonic deviceson the same chip opens new perspectives for optical in-terconnections (higher band widths, lower-latency linkscompare with copper wires) as well as for the possiblerole of photons in the data processing. These involvethe modelling of “standard” passive components, suchas waveguides, turning mirrors, splitters and input andoutput couplers, as well as active elements, such as mo-dulators and optical switches. Modelling of ultra-scaledoptoelectronic components involve the design of nano-meter scale optical architectures with new propertiesthat may differ considerably from those of their ma-croscopic counterparts.

This requires the development of new numerical toolsable to describe electromagnetic interactions and lightpropagation at different length scales. They should beable to describe the electromagnetic field from the scaleof a few light wavelengths (already of the order of thewhole micro-device) down to the nanometer scale ele-ments. These new tools should include a realistic des-cription of the optical properties including electro- andmagneto-optical active nanostructures and plasmonicelements which are expected to be key ingredients of anew generation of active optoelectronic components.

A mayor challenge of the “multi-physical” modelling willbe to simulate a full nanodevice where electronics, me-chanics and photonics meet at the nanoscale. For ins-tance, the coupling between mechanical vibrations andquantum conductance of single nanotubes has been re-cently observed. The interaction of an optical field witha device takes place not only through the electromag-netic properties, but also through the mechanical res-ponse (radiation pressure forces). The physicalmechanisms and possible applications of optical coolingof mechanical resonators are being explored. ModellingNano-Electro-Mechano-Optical (NEMO) devices isgoing to play a key (and fascinating) role in the deve-lopment and optimization of new transducers and devi-ces.

Thus, one of the main challenges for modelling in thenext few years is the creation of well organized colla-borations with a critical mass sufficient for the develop-

14

nanoICT research

Fig. 4: Functionalized graphene nanoribbons can be used to detect organic mo-lecules.

ment of integrated simulation platforms and with direct contacts with the industrial world.

Carbon-based electronics and spintronics

Amongst the most promising materials for the develop-ment of beyond CMOS nanoelectronics, Carbon Nano-tubes & Graphene-based materials and devices deservesome particular consideration. Indeed, first, their unu-sual electronic and structural physical properties pro-mote carbon nanomaterials as promising candidates fora wide range of nanoscience and nanotechnology appli-cations. Carbon is unique in possessing allotropes of eachpossible dimensionality and, thus, has the potential ver-satility of materials exhibiting different physical and che-mical properties. Diamond (3D), fullerenes (0D),nanotubes (1D-CNTs), 2D graphene and graphene rib-bons are selected examples. Because of their remarkableelectronic properties, CNTs or graphene-based materialsshould certainly play a key role in future nanoscale elec-tronics. Not only metallic nanotubes and graphene offerunprecedented ballistic transport ability, but they arealso mechanically very stable and strong, suggesting thatthey would make ideal interconnects in nanosized devi-ces. Further, the intrinsic semiconducting character ofeither nanotubes or graphene nanoribbons, as contro-lled by their topology, allows us to build logic devices atthe nanometer scale, as already demonstrated in manylaboratories. In particular, the combination of 2D gra-phene for interconnects (charge mobilities in graphenelayers as large as 400.000 cm2V-1s-1 have been repor-ted close to the charge neutrality point) together withgraphene nanoribbons for active field effect transistordevices could allow the implementation of completelycarbon-made nanoelectronics.

Besides the potential of 2D graphene and GNRs forelectronic device applications, transport functionalitiesinvolving the spin of the carriers have very recently re-ceived a particularly strong attention. First, althoughspin injection through ferromagnetic metal/semicon-ductor interfaces remains a great challenge, the specta-cular advances made in 2007 converting the spininformation into large electrical signals using carbon na-notubes [39] has opened a promising avenue for futurecarbon-based spintronic applications. The further de-monstration of spin injection in graphene [40] and theobservation of long spin relaxation times and lengths[41] have suggested that graphene could add some no-velty to carbon-based spintronics. For instance, takingadvantage of the long electronic mean free path and ne-gligible spin-orbit coupling, the concept of a spin field-ef-fect transistor with a 2D graphene channel has beenproposed with an expectation of near-ballistic spin trans-port operation [42]. A gate-tunable spin valve has beenexperimentally demonstrated [43]. Finally, a ferromag-netic insulator, such as EuO, may be used to induce fe-rromagnetic properties into graphene, through theproximity effect [44] and also to control the spin pola-rization of current by a gate voltage [45, 46]. This con-figuration does not make use of any ferromagneticmetallic contact to inject spin-polarized electrons. Thus,it could be a way to circumvent the problem of “con-ductivity mismatch” [47-49] which possibly limits the cu-rrent spin injection efficiency into a conventionalsemiconductor from a ferromagnetic metal. These phe-nomena and the corresponding devices need to be in-vestigated using the appropriate models ofrelativistic-like electron transport in 2D graphene struc-tures. Additionally, the presence of spin states at theedges of zigzag GNRS has also been demonstrated [50-52], and may be exploited for spintronics. Using first-principles calculations, very large values ofmagnetoresistance have been predicted in GNR-basedspin valves [53,54].

Additionally, the potential for 3D based organic spin-tronics has been recently suggested by experimental stu-dies [55]. Organic spin valves have shown spin relaxationtimes in the order of the microsecond and spin tunneljunctions with organic barriers have recently shownmagnetoresistance values in the same order of magni-tude as that of inorganic junctions based on Al2O3. Ho-wever, spin transport in these materials is basicallyFig. 5: A carbon nanotube consists of a rolled up sheet of graphene.

15

nanoICT research

unknown, and many groups are trying to decipher theimpressive experimental complexity of such devices. Or-ganic materials, either small molecules or polymers, willdefinitely allow large scale and low cost production of al-ternative non volatile memory technologies, with redu-ced power consumption. These new materials havetherefore the potential to create an entirely new gene-ration of spintronics devices, and the diverse forms ofcarbon-based materials open novel horizons for furtherhybridization strategies and all-carbon spintronics cir-cuits, including logic and memory devices.

The performance of these spintronic devices relies hea-vily on the efficient transfer of spin polarization acrossdifferent layers and interfaces. This complex transferprocess depends on individual material properties andalso, most importantly, on the structural and electronicproperties of the interfaces between the different ma-terials and defects that are common in real devices.Knowledge of these factors is especially important forthe relatively new field of carbon based spintronics,which is affected by a severe lack of suitable experi-mental techniques capable of yielding depth-resolved in-

formation about the spin polarization of charge carrierswithin buried layers of real devices.

In that perspective, it is noteworthy to remark that thefantastic development of first principles non-equilibriumtransport methods is progressively allowing for moreand more realistic assessment and anticipation on thetrue spintronics potential of carbon-based structures.This aspect also stands as an essential point for provi-ding guidance and interpretation schemes to experi-mental groups. As a matter of illustration, a few yearsago it has been theoretically shown that organic spin val-ves, obtained by sandwiching an organic molecule bet-ween magnetic contacts, could manifest largebias-dependent magnetoresistance, provided a suitedchoice of molecules and anchoring groups was made,which is now confirmed by experiments [56].

Finally one also notes that in addition to the potentialfor GMR in carbon-based materials, the spin manipula-tion and the realization of spin Qubits deserves a ge-nuine consideration. Recent theoretical proposals haveshown that spin Qubits in graphene could be coupled

Fig. 6: Functionalization of graphene nanoribbons

16

nanoICT research

over very long distances, as a direct consequence of theso-called Klein paradox inherent to the description ofcharge excitations in terms of massless Dirac fermions.The related challenges for device fabrication need to beobserved from different perspectives. Indeed, recent ex-periments in 13C nanotubes reveal surprisingly strongnuclear spin effects that, if properly harnessed, couldprovide a mechanism for manipulation and storage ofquantum information.

This may help to overcome the performance limitationsof conventional materials and of the conventional tech-nology for spin valve devices. The real potential of gra-phene-based materials for FET and related spintronicsapplications thus requires advanced modelling methods,including ab initio techniques, and a precise descriptionof spin degree of freedom.

To date, the development of nanotubes and graphenescience have been strongly driven by theory and quan-tum simulation [57,58].The great advantage of carbon-based materials and devices is that, in contrast to theirsilicon-based counterparts, their quantum simulationcan be handled up to a very high level of accuracy forrealistic device structures. The complete understandingand further versatile monitoring of novel forms of che-mically-modified nanotubes and graphene will howeverlead to an increasing demand for more sophisticatedcomputational approaches, combining first principles re-sults with advanced order N schemes to tackle materialcomplexity and device features, as developed in somerecent literature [59].

The molecular scale and the end of theroad

Molecular electronics research continues to explore theuse of single molecules as electron devices or for evenmore complex functions such as logic gates [60]. Expe-rimentally and theoretically the majority of researchwork focuses on single molecules between two metallicelectrodes or on molecular tunnel junctions.

Reproducibility of the measurements and accurate pre-dictions for currents across single molecule tunnel junc-tions remains a challenging task, although the results oftheory and experiment are converging [61]. To achievethe goal of using molecular components for computingor storage, or indeed for novel functions, requires refi-nement of the theoretical techniques to better mimic

the conditions under which most experiments are per-formed and to more accurately describe the electronicstructure of the molecules connected to the leads.

Most simulations of transport in molecular junctions todate are based on the combination of the non-equili-brium Green’s functions techniques with DFT calcula-tions [62,63]. Although this approach has proven quitepowerful, it also presents important shortcomings. Inparticular, the transport calculation is based on theKohn-Sham spectrum as calculated using standard localor semi-local exchange-correlation functionals. Thesefunctionals usually give a reasonable description of theelectronic spectrum for the normal metals that consti-tute the leads. However, they are known to be muchless reliable to predict the energy spectrum of small mo-lecules. This is extremely important, since the relativeposition of the molecular levels and the Fermi energyof the leads is crucial to determine the transport cha-racteristic. Some of the deficiencies of DFT to describelocalized levels can be corrected by including the so-ca-lled self-interaction corrections [64]. Still, the positionof the molecular levels is also strongly influenced by thedynamical screening induced by the metallic leads[65,66]. An accurate description of all these effects re-quires more elaborate theoretical treatments, beyondstandard DFT calculations. Fortunately, part of the ef-fects of screening can be included using a simple non-local self-energy model that basically contains image

Fig 7: Molecule connected to seven atomic wires on a ionic crystal. This setupis composed of 2025 atoms and more than 9700 atomic orbitals are neededto accurately calculate the function of the molecule (from “Picotechnologies :des techniques pour l'échelle atomique”, C. Joachim, A. Gourdon and X.Bouju in Techniques pour l'ingénieur NM130, 1-16 (2009).

18

nanoICT research

charge interactions that affect differently the HOMOand LUMO levels [68].

Another interesting issue is that of the coupling of theelectrons with structural deformations and vibrationalmodes. Within the framework of NEGF+DFT calcula-tions inelastic effects associated with the excitation oflocalized vibration within the contact region have beensuccessfully included in recent years. Most of the ap-proaches rely on different perturbative expansions oron the so-called self-consistent Born approximation [69].This allows for accurate simulations of the inelastic sig-nal in molecular junctions and, therefore, the characte-rization of the path followed by the electrons byidentifying which vibrations are excited during the con-duction process. Unfortunately, the self-consistent Bornapproximation does not allow accounting for polaroniceffects that are crucial to understand the electronictransport in flexible organic molecules [70]. In thesecases more sophisticated methods are necessary, thecoupling of which to ab initio DFT simulations is still anopen question.

Hence much of the work to date has focused on the un-derlying physical mechanisms of charge transport acrossmolecules, whereas very little is understood in terms ofthe use of molecular components in complex, or evensimple, circuits. This research area could also be cate-gorized as in its infancy in that very little is known abouttime-dependent or AC responses of molecules in tun-nel junctions or other circuit environments. To exploit

molecules in information processing, the use of multi-scale tools [71] as described previously are needed toembed molecular scale components between what areessentially classical objects: leads, drivers, and circuits.Further development of the simulations is needed to ac-curately describe the transport processes in molecularjunctions and its coupling to structural degrees of free-dom and, in particular, to molecular vibrations also inthe time-dependent response of the molecules to ex-ternal voltages and their interaction with light needs tobe studied.

Finally, to position a single molecule at the right place,experimental equipment has to be developed with ac-curate manipulation capabilities, as well as precisionelectrical probes for four terminal measurements [72].

Further development of the simulation techniques is ne-eded to describe the time-dependent response of mo-lecules to external voltages and their interaction withlight. As the size of molecules considered as a compo-nent remains quite large, adapted methods using a scat-tering approach seem to be more relevant than

high-level NEGF or other sophisticated methods [73,74].The concept of a molecular logic gate has emerged re-cently and specific approaches including quantum Ha-miltonian computing [75] will be used for a numericalintegration.

Fig. 8: Connection of a molecule to four atomic wires. Each wire has been fabricated (left panel) with f ive gold atoms deposited on the substrate; the mole-cule has then been placed at the center of the device (right panel), slightly modifying the structure of one of the wires.

19

nanoICT research

Nano-bio-electronics

A further fundamental research line which emerged inparallel to the development of molecular electronics isbio-electronics. It mainly aims at (i) mimicking of biolo-gical and biophysical processes via molecular electroniccircuits, and (ii) the exploitation of self-assembling andself-recognition properties of many biomolecules as scaf-folds in the integration of nano- and sub-nanoscale de-vices, and (iii) exploring the potential of arrays ofbiomolecules (like DNA and its artificial modifications)to serve as molecular wires interconnecting differentparts of molecular electronic devices. Obviously, greatchallenges from both the experimental and from thetheoretical side exist on the road to achieve such goals.The theoretical understanding of the bio/inorganic in-terface is in its infancy, due to the large complexity ofthe systems and the variety of different physical inter-actions playing a dominant role. Further, state of the artsimulation techniques for large biomolecular systems areto a large degree still based on classical physics approa-ches (classical molecular dynamics, classical statisticalphysics); while this can still provide valuable insight intomany thermodynamical and dynamical properties of

biomolecular systems, a crucial point is nevertheless mis-sing: the possibility to obtain information about the elec-tronic structure of the biomolecules, an issue which isessential in order to explore the efficiency of suchsystems to provide charge migration pathways. Moreo-ver, due to the highly dynamical character of biomole-cular systems -seen e.g. in the presence of multiple timescales in the atomic dynamics- the electronic structure isstrongly entangled with structural fluctuations. We arethus confronting the problem of dealing with the inter-action of strongly fluctuating complex molecules withinorganic systems (substrates, etc).

As a result, multi-scale simulation techniques are ur-gently required, which should be able to combine quan-tum-mechanical approaches to the electronic structurewith molecular dynamical simulation methodologies de-aling with the complex conformational dynamics of bio-logical objects. Conventional simulation tools ofsemiconducting microdevices can obviously not dealwith such situations.

Thermoelectric energy conversion

The importance of research on thermoelectric energyconversion is growing in parallel with the need for al-ternative sources of energy. With the recent develop-ments in the field, thermoelectric energy generatorshave become a commercial product in the market andtheir efficiencies are improving constantly, but the com-mercially available products did not take the advantageof nano-technology yet. In fact, thermoelectricity is oneof the areas in which nano-scale fabrication techniquesoffer a breakthrough in device performances.

It was predicted theoretically that, lowering the devicedimensions, it is possible to overcome the Wiedemann-Franz Law and enhance the device performances signi-ficantly [76].Quasi one-dimensional quantum wires [77], engineeredmolecular junctions [78,79], superlattices of quantumdots [80] are the other possible routes proposed forachieving a high thermoelectric figure of merit, ZT.

The thermoelectric figure of merit includes three pro-perties of the material, namely the electrical conducti-vity σ, Seebeck coefficient S, the thermal conductivityκ=κel+κph with electronic and phononic contributions as

Fig. 9: Diffusion processes are of fundamental importance in many disciplinesof physics, chemistry and biology. A low diffusion rate may hinder progress inmany issues related to technology and health improvement. For example highviscosity may prevent tailored chemical reactions or work as an undesired ba-rrier for targeted nano-engineered drug delivery in bio-chips. The classical dif-fusion of a small particle in a f luid can be greatly enhanced by the light f ieldof two interfering laser beams. Langevin Molecular Dynamics simulations showthat radiation pressure vortice, due to light interference, spin the particle outof the whirls sites leading to a giant acceleration of free diffusion. The effec-tive viscosity can then be notably reduced by simply increasing the laser in-tensity. [Albaladejo et al., Nano Letters 9, 3527 (2009)]

20

nanoICT research

well as the temperature T, ZT =σ S2 T/κIn order to optimize ZT, an electron-crystal togetherwith a phonon-glass behavior is required [81].

Indeed, it has recently been shown that Si nano-wireswith rough surfaces can serve as high performance ther-moelectric materials, since the edge roughness sup-presses phonon transport by a few orders of magnitudewhereas the electronic transport is weakly affected fromsurface roughness in these wires. Similar behavior is alsoreported for graphene nanoribbons where edge disor-der reduces lattice thermal conductivity while electronsin the first conduction plateau stay almost intact [82].At the sub-band edges, however, the electronic meanfree path is discontinuous which yields a large increaseof the Seebeck coefficient and ZT. Another criteria forhigh ZT is to have a narrow distribution of the energy ofthe electrons participating in the transport process [83].In molecular junctions, it is shown that the Seebeck co-efficient is maximized when the electrode-molecule cou-pling is weak and the HOMO level is close to the Fermienergy, these fulfill Mahan's criterion in the zero-di-mensional case [84]. The weak coupling condition is alsoin favor of the need of reduced vibrational thermal con-duction. Therefore self-assembled layers of moleculesbetween semiconducting surfaces are supposed to yieldhigh ZT values.

These developments offer new challenges for increasingdevice performances and also for the modelling of newdevices. An accurate modelling of the disordered struc-tures or of the layered structures of self-assembled mo-lecules between surfaces requires a lot of improvementsboth from the methodological and from the computa-tional point of view. Simulation of these systems entailsconsideration of very large number of atoms, of theorder of 106 to 108. Therefore fully parallelized order - N methods for both electronic and phononic compu-tations are needed. Though important progress hasbeen achieved in order-N electronic calculations, suchmethods are still lacking a proper description for thephonons. Adoption of the methods already developedfor electrons to phonons is a first goal to be achieved.

Multifunctional oxides

Multifunctional oxides, ranging from piezoelectrics tomagnetoelectric multiferroics, offer a wide range ofphysical effects that can be used to our advantage in thedesign of novel nanodevices. For example, these mate-

rials make it possible to implement a variety of tunableand/or switchable field effects at the nanoscale. Thus,a magnetoelectric multiferroic can be used to controlthe spin polarization of the current through a magnetictunnel junction by merely applying a voltage; or a pie-zoelectric layer can be used to exert very well contro-lled epitaxial-like pressures on the adjacent layers of amultilayered heterostructure, which e.g. can in turn trig-ger a magnetostructural response. These are just twoexamples of many novel applications that add up to themore traditional ones -- as sensors, actuators, memo-ries, highly-tunable dielectrics, etc. -- that can now bescaled down to nanometric sizes by means of moderndeposition techniques. In fact, nanostructuring of diffe-rent types, ranging from the construction of oxide na-notubes to the more traditional multi-layered systems,is opening the door to endless possibilities for the engi-neering/combination of the properties of this type ofcompounds, which are strongly dependent on the syste-m's size and (electrical, mechanical) boundary condi-tions.

The current challenges in the field are plenty, from themore technical to the more fundamental. At the levelof the applications, the outstanding problems includethe integration of these materials with silicon or theidentification of efficient and scalable growth techniquesfor complex oxide heterostructures. At a more funda-mental level, there is a pressing need to identify newsystems and/or physical mechanisms that can materia-lize some of the most promising concepts for the designof devices; for example, we still lack a robust room-tem-perature magnetoelectric multiferroic system that canbe integrated in real devices, we still have to understandthe main factors controlling the performance of a fe-rroelectric tunnel junction, etc. Finally, it should bestressed that the field is rich in opportunities for theemergence of novel effects and concepts that go beyondthe current prospects in the general area of electronics,the recent discovery of high electronic mobilities in all-oxide heterostructures (i.e., at the interfaces betweenLaAlO3 and SrTiO3) being a remarkable example.

Quantum-mechanical simulation is playing a key role inthe progress in multifunctional oxides. This has been his-torically the case, with many key contributions from thefirst-principles community to the understanding of fe-rroelectric, magnetic and magnetoelectric bulk oxides[85,86]. This trend is just getting stronger in this era of

21

nanoICT research

nanomaterias, for two reasons: (i) there is greater needfor theory to explain the novel physical mechanisms atwork in systems that are usually very difficult to cha-racterize experimentally, and (ii) modern depositiontechniques offer the unique chance to realize in the la-boratory the most promising theoretical predictions fornew materials. Thus, the importance of first-principlessimulation for fundamental research in functional oxi-des is beyond doubt, and very recent developments, e.g.for the first-principles study of the basic physics of mag-netoelectrics [87], ferroelectric tunnel junctions [88] ornovel oxide superlattices [89], clearly prove it. The con-tribution from simulations to resolve more applied pro-blems (e.g, that of the integration with silicon) is, on theother hand, just starting, and its progress will criticallydepend on our ability to develop novel multi-scale si-mulation methods that can tackle the kinetics of speci-fic growth processes (from pulse-laser deposition tosputtering methods under various oxygen atmospheres)directly. This is a major challenge that will certainly ge-nerate a lot of activity in the coming couple of decades.

Major current deficiencies of TCAD

A clear gap, which has in part been addressed in theprevious sections, has formed in the last few years bet-ween what is available in the TCAD market and whatwould actually be needed by those working on the de-velopment of advanced nanoscale devices. As alreadymentioned, classical TCAD tools have not been upgra-ded with realistic quantum transport models yet, suita-ble for the current 32 nm node in CMOS technology orfor emerging technologies, and, in addition, there aresome fields of increasing strategic importance, such asthe design of photovoltaic cells, for which no well-esta-blished TCAD platforms exist.Bottom-up approaches, which, if successful, could pro-vide a solution to one of the major bottlenecks on thehorizon, i.e. skyrocketing fabrication costs, are not sup-ported by any type of TCAD tools as of now. This maybe due to the fact that bottom-up approaches are still intheir infancy and have not been demonstrated in anylarge-scale application, but the existence of suitable pro-cess simulation tools could, nevertheless, facilitate theirdevelopment into actual production techniques. So-phisticated tools that have been developed within rese-arch projects are available on the web, mainly onacademic sites, but they are usually focused on specificproblems and with a complex and non-standardizeduser interface. An effort would be needed to coordi-

nate the research groups working on the developmentof the most advanced simulation approaches, the TCADcompanies and the final users, in order to define a com-mon platform and create the basis for multi-scale toolssuitable to support the development of nanoelectronicsin the next decade.

New computational approaches

The development of highly parallel and computationallyefficient graphic processors has recently provided a newand extremely powerful tool for numerical simulations.Modern Graphic Processing Units (GPU) approach apeak performance of a Teraflop (1012 floating pointoperations per second) thanks to a highly parallel struc-ture and to an architecture focusing specifically on dataprocessing rather than on caching or flow control. Thisis the reason why GPUs excel in applications for whichfloating point performance is paramount while memorybandwidth in not a primary issue. In particular, GPUhardware is specialized for matrix calculations (funda-mental for 3D graphic rendering), which do representalso the main computational burden in many types ofdevice simulations.

As a result, speed ups of the order of 30 - 40 have beenobserved, for tasks such as the simulation of nanoscaletransistors, with respect to state-of-the-art CPUs. Upto now the main disadvantage was represented by theavailability, in hardware, only of single-precision opera-tions, but last generation GPUs, such as the FireStream9170 by AMD, are advertised as capable of handlingdouble precision in hardware, although at a somewhatreduced rate (possibly by a factor of 5). Another im-pressive feature of GPU computation is the extremelyhigh energy efficiency, of the order of 5 Gigaflops/Win single precision or 1 Gigaflop/W in double precision,an aspect of growing importance considering the costsfor supplying power and air conditioning to computinginstallations.

The latest GPU hardware opens really new perspectivesfor simulation of nanodevices, also in "production envi-ronments," because GPU based systems could be easilystandardized and provided to end-users along with thesimulation software. Overall, this is a field that deservesinvesting some time and effort on the part of the de-

nanoICT research

22

vice modelling community, because it could result in areal breakthrough in the next few years.

Overview of networking for modelling inEurope and the United States

In the United States, the network for computational na-notechnology (NCN) is a six-university initiative esta-blished in 2002 to connect those who developsimulation tools with the potential users, including thosein academia, and in industries. The NCN has received afunding of several million dollars for 5 years of activity.One of the main tasks of NCN is the consolidation ofthe nanoHUB.org simulation gateway, which is currentlyproviding access to computational codes and resourcesto the academic community. According to NCN survey[23], the total number of users of nanoHUB.org rea-ched almost 70.000 in March 2008, with more than6.000 users having taken advantage of the online simu-lation materials. The growth of the NCN is likely to at-tract increasing attention to the US computationalnanotechnology platform from all over the world, fromstudents, as well as from academic and, more recently,industrials researchers. In Europe an initiative similar tothe nanoHUB, but on a much smaller scale, was star-ted within the EU funded “Phantoms network of exce-llence” and has been active for several years; it iscurrently being revived with some funding within thenanoICT coordination action. (www.europa.iet.unipi.it)

In a context in which the role of simulation might be-come strategically relevant for the development of nano-technologies, molecular nanosciences, nanoelectronics,nanomaterial science and nanobiotechnologies, it seemsurgent for Europe to set up a computational platform in-frastructure similar to NCN, in order to ensure its po-sitioning within the international competition. The needsare manifold. First, a detailed identification of Europeaninitiatives and networks must be performed, and de-fragmentation of such activities undertaken. A pioneerinitiative has been developed in Spain through theM4NANO database (www.m4nano.com) gathering allnanotechnology-related research activities in modellingat the national level. This Spanish initiative could serveas a starting point to extend the database to the Euro-pean level. Second, clear incentives need to be launchedwithin the European Framework programmes to en-courage and sustain networking and excellence in the

field of computational nanotechnology and nanoscien-ces. To date, no structure such as a Network of Exce-llence exists within the ICT programme, although theprogramme NMP supported a NANOQUANTA NoEin FP6, and infrastructural funding has been provided tothe newly established ETSF (European Theoretical Spec-troscopy Facility, www.etsf.eu). This network mainly ad-dresses optical characterization of nanomaterials, andprovides an open platform for European users, that canbenefit from the gathered excellence and expertise, aswell as standardized computational tools. There is alsoa coordinated initiative focused on the specific topic ofelectronic structure calculations, the Psi-k network(www.psi-k.org).

An initiative similar to the American NCN would be ne-eded in Europe, within the ICT programme that en-compasses the broad fields of devices and applicationsor, better, in conjunction between the ICT and the NMPprogramme, since the full scope from materials to devi-ces and circuits should be addressed.

Other initiatives such as the “Report on multiscale ap-proaches to modelling for nanotechnology” [90] intendsto provide an overview, albeit limited and certainly notexhaustive, of relevant aspects of modelling at the na-noscale, pointing out some important issues that are stillopen and affording the reader that is not yet active inthe field with an introduction to several widely usedtechniques and with a large body of references.

This review has been written by experts in the fields ofcomputational modelling; most of them have stronglycontributed to the development of European excellencein recent years, and have been leading EU-initiative overFP5, FP6 and FP7. Although more efforts will be neededto bridge different communities from ab initio develop-ment to device simulation, such initiatives need to besupported considering that contributors of this reportare overviewing promising methodologies to fill the gapbetween scientific communities, establishing some fra-mework for further promoting European-based net-working activities and coordination.

Past, present and future European advan-ces in computational approaches.

This novel initiative should be able to bridge advancedab-initio/atomistic computational approaches to ulti-

nanoICT research

23

mate high-level simulation tools such as TCAD modelsthat are of crucial importance in software companies.Many fields such as organic electronics, spintronics, be-yond CMOS nanoelectronics, nanoelectromechanicaldevices, nanosensors, nanophotonics devices definitelylack standardized and enabling tools that are howevermandatory to assess the potential of new concepts, orto adapt processes and architectures to achieve the des-ire functionalities. The European excellence in thesefields is well known and in many aspects overcomes thatof the US or of Asian countries. Within the frameworkof a new initiative, specific targets should be addressedin relation with the modelling needs reported by smalland medium sized software companies active in the de-velopment of commercial simulation tools, such asQUANTUM WISE (www.quantumwise.com)SYNOPSYS (www.synopsys.com)NANOTIMES (www.nanotimes-corp.com)SILVACO (www.silvaco.com)NEXTNANO3 (www.nextnano.de)TIBERCAD (www.tibercad.org)

Similarly, larger companies such as STMicroelectronics,Philips, THALES, IBM, INTEL make extensive usage ofcommercial simulation tools to design their technologi-cal processes, devices and packaging. The sustainabledevelopment of the computational simulation softwareindustry, including innovative materials (carbon nanotu-bes, graphene, semiconducting nanowires, molecular as-semblies, organics, magnetic material) and novelapplications (spintronics, nanophotonics, beyond CMOSnanoelectronics), could therefore be crucial to foster in-dustrial innovation in the next decade.

Conclusions

Recent advances in nanoscale device technology havemade traditional simulation approaches obsolete fromseveral points of view, requiring the urgent developmentof a new multiscale modelling hierarchy, to support thedesign of nanodevices and nanocircuits. This lack of ade-quate modeling tools is apparent not only for emergingdevices, but also for aggressively scaled traditionalCMOS technology, in which novel geometries and novelmaterials are being introduced. New approaches to si-mulation have been developed at the academic level,but they are usually focused on specific aspects and havea user interface that is not suitable for usage in an in-dustrial environment. There is therefore a need for in-tegration of advanced modelling tools into simulators

that can be proficiently used by device and circuit engi-neers: they will need to include advanced physical mo-dels and at the same time be able to cope withvariability and fluctuations, which are expected to beamong the greatest challenges to further device downscaling.

In addition, as dimensions are scaled down, the distinc-tion between material and device properties becomesincreasingly blurred, since bulk behavior is not obser-ved any more, and atomistic treatments are needed.There is therefore a convergence between material anddevice research, which should be reflected also in theformulation of research projects. Furthermore, newmaterials, such as carbon, bio-molecules, multifunctionaloxides, are emerging, with an impressive potential fordevice fabrication and with completely new require-ments for simulation.

A unique opportunity is now surfacing, with powerfulnew modelling approaches being developed and newlow-cost computational platforms (such as GPUs) withan unprecedented floating point performance. Thecombination of these two factors makes it clear that thetime is ripe for a new generation of software tools,whose development is of essential importance for thecompetitiveness and sustainability of European ICT in-dustry, and which requires a coordinated effort of allthe main players.

References

[1] Sparta is part of the Synopsys TCAD suite;http://193.204.76.120/ISETCADV8.0/PDFManual/data/Sparta.pdf

[2] Quantum3D is a Silvaco product; http://www.sil-vaco.com/products/vwf/atlas/3D/quantum3D/

[3] http://www.nextnano.de

[4] http://www.tibercad.org

[5] http://public.itrs.net

[6] H. P. Tuinhout, "Impact of parametric mismatch andfluctuations on performance and yield of deep-submi-cron CMOS technologies", Proc. ESSDERC, pp.95-101,Florence, Italy, 2002.

nanoICT research

24

[7] D. J. Frank and Y. Taur, "Design considerations forCMOS near the limits of scaling", Solid-State Electron.46, 315 (2002).

[8] K. Takeuchi, R. Koh and T. Mogami, "A study of thethreshold voltage variation for ultra-small bulk and SOICMOS", IEEE Trans. Electron Dev 48, 1995 (2001).

[9] T. Mizuno, J. Okamura and A. Toriumi, "Experi-mental study of threshold voltage fluctuation due to sta-tistical variation of channel dopant number inMOSFET’s", IEEE Trans. Electron Devices 41, 2216(1994).

[10] A. Asenov, A. R. Brown J. H. Davies, S. Kaya†, andG. Slavcheva, "Simulation of Intrinsic Parameter Fluc-tuations in Decananometre and Nanometre scale MOS-FETs", IEEE Trans. Electron Devices 50, 1837 (2003).

[11] P. A. Stolk, F. P. Widdershoven, D. B. M. Klaassen,"Device modeling of statistical dopant fluctuations inMOS transistors," Proc. SISPAD’97, p. 153, 1997.

[12] H.-S. Wong and Y. Taur "Three dimensional ‘ato-mistic’ simulation of discrete random dopant distribu-tion effects in sub-0.1 mm MOSFETs", Proc. IEDM Dig.Tech. Papers., p. 705, 1993.

[13] D. J. Frank, Y. Taur, M. Ieong and H.-S. P. Wong,"Monte Carlo modeling of threshold variation due todopant fluctuations", 1999 Symposium on VLSI Techno-logy Dig. Techn. Papers, p, 169, 1999.

[14] D. Vasileska, W. J. Gross and D. K. Ferry, "Mode-ling of deep-submicrometer MOSFETs: random impu-rity effects, threshold voltage shifts and gate capacitanceattenuation", Extended Abstracts IWEC-6, Osaka 1998,IEEE Cat. No. 98EX116, p. 259.

[15] A. Asenov, S. Kaya and A. R. Brown, "Intrinsic Pa-rameter Fluctuations in Decananometre MOSFETs In-troduced by Gate Line Edge Roughness", IEEE Trans.Electron Dev. 50, 1254, (2003).

[16] M. Bescond, N. Cavassilas, K. Nehari, J. L. Autran,M. Lannoo and A. Asenov, "Impact of Point Defect Lo-cation and Type in Nanowire Silicon MOSFETs", Proc.35th European Solid-State Device Research Conference(ESSDERC), 221, Grenoble (France), September 2005.

[17] A. Martinez, M. Bescond, J. R. Barker, A. Svizhen-kov, A. Anantram, C. Millar, A. Asenov, "Self-consistentfull 3D real-space NEGF simulator for studying of non-perturbative effects in nano-MOSFET", IEEE Trans. Elec-tron Dev. 54, 2213, (2007).

[18] M. Ono at al. "Effect of metal concentration nonu-niformity in gate dielectric silicates on propagation delaytime of CMIS invertors", Proc. SSDM 2002 Nagoya,Japan, 710 (2002).

[19] J.R. Barker, J.R. Watling, "Non-equilibrium dielec-tric response of High-k gate stacks in Si MOSFETs: Ap-plication to SO interface phonon scattering", J. Phys.:Conference Series 35, 255 (2006).

[20] J.-C. Charlier, X. Blase, and S. Roche, “Electronicand Transport Properties of Nanotubes”, Rev. Mod.Phys. 79, 677-732 (2007) [21] A.K. Geim and K. S. No-voselov, “The rise of Graphene”, Nature Materials 6,183 (2007).

[21] J. M. Soler, E. Artacho, J. D. Gale, A. García, J. Jun-quera, P. Ordejón, D. Sánchez-Portal, “The SIESTA me-thod for ab initio order-N materials simulations”,Journal of Physics: Condensed Matter 14, 2745(2002).

[22] S. Goedecker, ”Linear scaling electronic structuremethods”, Rev. Mod. Phys. 71, 1085 (1999).

[23] Ordejon P, “Order-N tight-binding methods forelectronic-structure and molecular dynamics”, Comp.Mat. Sci. 12, 157 (1998).

[24] http://cms.mpi.univie.ac.at/vasp/

[25] http://www.castep.org

[26] http://www.crystal.unito.it

[27] http://cp2k.berlios.de

[28] http://www.abinit.org

[29] http://www.icmab.es/siesta/

[30] http://cp2k.berlios.de/quickstep.html

[31] M. Brandbyge, J. L. Mozos, P. Ordejon, J. Taylor, K.Stokbro, “Density-functional method for nonequilibriumelectron transport”, Phys. Rev. B 65 , 165401 (2002).

nanoICT research

25

[32] A. R. Rocha, V. M. Garcia-Suarez, S. W. Bailey, C.J. Lambert, V. Ferrer, S. Sanvito, “Spin and molecularelectronics in atomically generated orbital landscapes”,Phys. Rev. B 73, 085414 (2006).

[33] http://www.quantumwise.com

[34] C. M. Goringe, D. R. Bowler, E. Hernandez,”Tight-binding modelling of materials”, Reports on Progress inPhysics 60, 1447 (1997).

[35] W. Jaskólski, M. Zielinski, G. W. Bryant, and J. Aiz-purua,"Strain effects on the electronic structure ofstrongly coupled self-assembled InAs/GaAs quantumdots: Tight-binding approach", Phys. Rev. B. 74, 195339(2006).

[36] D. Porezag, Th. Frauenheim, Th. Köhler, G. Seifert,R. Kaschner, Phys. Rev. B 51, 12947 (1995).

[37] G. Lu and E. Kaxiras, “Overview of Multiscale Si-mulations of Materials", In: M. Rieth and W. Schommers(eds.), Handbook of Theoretical and ComputationalNanotechnology, vol. X, pp.1-33, American ScientificPublishers (2005).

[38] P. Vashishta, R. Kalia, University A. Nakano, B.Homan, K. L. McNesby, “Multimillion Atom Reactive Si-mulations of Nanostructured Energetic Materials”,Journal of Propulsion and Power 23, 688 (2007).

[39] L. E. Hueso, J. M. Pruneda, V. Ferrari, G. Burnell,J. P. Valdes-Herrera, B. D. Simons, P.B. Littlewood, E.Artacho, A. Fert & N. D. Mathur, “Transformation ofspin information into large electrical signals using car-bon nanotube”, Nature 445, 410 (2007).

[40] M. Ohishi, M. Shiraishi, R. Nouchi, T. Nozaki, T.Shinjo, Y. Suzuki, “Spin injection into a graphene thinfilm at room temperature”, Jpn. J. Appl. Phys. 46, L605-L607 (2007).

[41] N. Tombros, C. Jozsa, M. Popinciuc, H. T. Jonkman,B. J. van Wees, “Electronic spin transport and spin pre-cession in single graphene layers at room temperature”,Nature 448, 571 (2007).

[42] Y. G. Semenov, K. W. Kim, J. M. Zavada, ”Spin fieldeffect transistor with a graphene channel”, Appl. Phys.Lett. 91, 153105 (2007).

[43] S. Cho, Y.F. Chen, M.S. Fuhrer, ”Gate-tunable spinvalve”, Appl. Phys. Lett. 91, 123105 (2007).

[44] H. Haugen, D. Huertas, A. Brataas, “Spin trans-port in proximity-induced ferromagnetic graphene”,Phys. Rev. B 77, 115406 (2008).

[45] T. Yokoyama, “Controllable spin transport in fe-rromagnetic graphene junctions”, Phys. Rev. B 77,073413 (2008).

[46] V. Nam Do, V. Hung Nguyen, P. Dollfus, A. Bour-nel, “Electronic transport and spin-polarized effects ofrelativistic-like particles in graphene structures”, J. Appl.Phys. 104, 063708 (2008).

[47] G. Schmidt, D. Ferrand, L. W. Molenkamp, A. T.Filip, B. J. van Wees “Fundamental obstacle for electri-cal spin-injection from a ferromagnetic metal into a dif-fusive semiconductor”, Phys. Rev. B 62, 4790-4793(2000).

[48] E. I. Rashba “Theory of electrical spin-injection:tunnel contact as a solution of the conductivity mis-match problem”, Physical Review B 62, 16267-16270(2000).

[49] A. Fert, H. Jaffrès “Conditions for efficient spin in-jection from a ferromagnetic metal into a semiconduc-tor”, Phys. Rev. B 64, 184420 (2001).

[50] H. Lee, Y. W. Son, N. Park, S. W. Han, J. J. Yu,“Magnetic ordering at the edges of graphitic fragments:Magnetic tail interactions between the edge-localizedstates”, Phys. Rev. B 72, 174431 (2005).

[51] L. Pisani, J. A. Chan, B. Montanari, N. M. Harrison,“Electronic structure and magnetic properties of gra-phitic ribbons”, Phys. Rev. B 75, 064418 (2007).

[52] M. Topsakal, H. Sevinçli, S. Ciraci, “Spin confine-ment in the superlattices of graphene ribbons”, Appl.Phys. Lett. 92, 173118 (2008).

[53] W. Y. Kim, K. S. Kim, “Prediction of very large va-lues of magnetoresistance in a graphene nanoribbon de-vice”, Nature Nanotechnology 3, 408 (2008).

[54] W. F. Munoz-Rojas, J. Fernandez-Rossier and J. J.Palacios, “Giant Magnetoresistance in Ultrasmall Gra-phene Based Devices”, Phys. Rev. Lett. 102, 136810(2009).

[55] V. Dediu, L. Hueso, I. Bergenti and C. Taliani, ”Spinroutes in organic semiconductors”, Nature Materials 8,707 (2009).

nanoICT research

26

[56] G. Szulczewski, S. Sanvito and M. Coey, ”A spin oftheir own”, Nature Materials 8, 693 (2009).

[57] A.K. Geim and K. S. Novoselov, “The rise of Gra-phene”, Nature Materials 6, 183 (2007).

[58] A. Lherbier, X. Blase, F. Triozon, Y-M Niquet andS. Roche, “Charge Transport in Chemically Doped 2Dgraphene”, Physical Review Letters 101, 036808 (2008).

[59] M. Lundstrom, G. Klimeck, G.B. Adams, M. McLen-nan, "HUB is Where the Heart is”, IEEE Nanotechno-logy Magazine 28, March 2008.

[60] C. Joachim, J. K. Gimzewski and A. Aviram, “Elec-tronics using hybrid-molecular and mono-molecular de-vices”, Nature 408, 541 (2000).

[61] M. Galperin, M. A. Ratner, A. Nitzan, and A. Troisi,"Nuclear Coupling and Polarization in Molecular Trans-port Junctions: Beyond Tunneling to Function”, Science,319, 1056 (2008).

[62] Brandbyge M, Mozos JL, Ordejon P, Taylor J, Stok-bro K, “Density-functional method for nonequilibriumelectron transport”, Phys. Rev. B 65 , 165401 (2002).

[63] Rocha AR, Garcia-Suarez VM, Bailey SW, LambertCJ, Ferrer J, Sanvito S, “Spin and molecular electronicsin atomically generated orbital landscapes”, Phys. Rev. B73, 085414 (2006).

[64] C. Toher and S. Sanvito, “Efficient Atomic Self-Interaction Correction Scheme for NonequilibriumQuantum Transport”, Phys. Rev. Lett. 99, 056801(2007).

[65] J. B. Neaton, M. S. Hybertsen, and S. G. Louie,“Renormalization of Molecular Electronic Levels atMetal-Molecule Interfaces”, Phys. Rev. Lett. 97, 216405(2006).

[66] K. S. Thygesen and A. Rubio, “Renormalization ofMolecular Quasiparticle Levels at Metal-Molecule Inter-faces: Trends across Binding Regimes”, Phys. Rev. Lett.102, 046802 (2009).

[68] S. Y. Quek, L. Venkataraman, H. J. Choi, S. G.Louie, M. S. Hybertsen, and J. B. Neaton, “Amine−GoldLinked Single-Molecule Circuits: Experiment and The-ory”, Nano Lett. 7, 3477 (2007).

[69] T. Frederiksen, M. Brandbyge N. Lorenteand Antti-Pekka Jauho, “Inelastic Scattering and Local Heating in

Atomic Gold Wires”, Phys. Rev. Lett. 93, 256601(2004).

[70] H. Ness, S. A. Shevlin, and A. J. Fisher, Phys. Rev.B 63, 125422 (2001).

[71] H. Guo, D. Martrou, T. Martrou, J. Polesel-Maris,A. Piednoir, E. Dujardin, S. Gauthier, M. A. F. van denBoogaart, L. M. Dieswijk and J. Brügger, Appl. Phys.Lett. 90, 093113 (2007).

[72] J. S. Yang, D. Jie, N. Chandrasekhar and C. Joa-chim, “UHV-STM manipulation of single flat gold nano-islands for constructing interconnection nanopads onMoS2”, J. Phys.: Conf. Series 61, 1288 (2007).

[73] S. Ami and C. Joachim, “Intramolecular circuitsconnected to N electrodes using a scattering matrix ap-proach”, Phys. Rev. B 65, 155419 (2002).

[74] N. Jlidat, M. Hliwa and C. Joachim, “A moleculeOR logic gate with no molecular rectifier”, Chem. Phys.Lett. 470, 275 (2009).

[75] N. Renaud, M. Ito, W. Shangguan, M. Saeys, M.Hliwa and C. Joachim, “A NOR-AND quantum runninggate molecule”, Chem. Phys. Lett. 472, 74 (2009).

[76] L. D. Hicks and M. S. Dresselhaus, Phys. Rev. B 47,16631 (1993).

[77] A. I. Hochbaum, R. Chen, R. D. Delgado, W.Liang, E. C. Garnett, M. Najarian, A. Majumdar and P.Yang, Nature 451, 163 (2008).A. I. Boukai, Y. Bunimovich, J. Tahir-Kheli, J.-K. Yu, W.A. G. Iii and J. R. Heath, Nature 451, 168 (2008).

[78] P. Reddy, S.-Y. Jang, R. A. Segalman and A. Ma-jumdar, Sicence 315, 1568 (2007).

[79] S.-H. Ke, W. Yang, S. Curtarolo and H. U. Baran-ger, Nano Lett. 9, 1011 (2009).

[80] T. C. Harman, P. J. Taylor, D. L. Spears, and M. P.Walsh, Electron. Mater. 29, L1 (2000).T. C. Harman, M. P. Walsh, B. E. Laforge, and G. W. J.Turner, Electron. Mater. 34, L19 (2005).R. Venkatasubramanian, E. Silvola, T. Colpitts, and B.O’Quinn, Nature 413, 597 (2001).

nanoICT research

27

[81] G. A. Slack "CRC Handbook of Thermoelectrics",ed. D. M. Rowe, Boca Raton, FL: CRC Press (1995).

[82] H. Sevinçli and G. Cuniberti, arXiv: 0908.3207.

[83] G. D. Mahan and J. O. Sofo, “The best thermoe-lectric”, PNAS 93, 7436 (1993).

[84] D. Nozaki, H. Sevinçli, W. Li, R. Gutierrez, G. Cu-niberti, arXiv: 0908.0438.

[85] K. Rabe, Ch. H. Ahn, and J.-M. Triscone, Eds.,"Physics of Ferroelectrics, A Modern Perspective",(Springer-Verlag Berlin Heidelberg 2007).

[86] M. Imada, A. Fujimori, and Y. Tokura, "Metal-insu-lator transitions", Reviews of Modern Physics 70, 1039(1998).

[87] N. A. Spaldin and R. Ramesh, "Electric-field con-trol of magnetism in complex oxide thin films", MRS Bu-lletin 33, 1047 (2008).

[88] J. P. Velev, P. A. Dowben, E. Y. Tsymbal, S. J. Jen-kins, and A. N. Caruso, “Interface effects in spin-pola-rized metal/insulator layered structures", SurfaceScience Reports 63, 400 (2008).

[89] E. Bousquet, M. Dawber, N. Stucki, C. Lichtenstei-ger, P. Hermet, S. Gariglio, Jean-Marc Triscone, and P.Ghosez, "Improper ferroelectricity in perovskite oxideartificial superlattices", Nature 452, 732 (2008).

[90] M. Macucci, S. Roche and A. Correia, “Report onmultiscale approaches to modelling for nanotechno-logy”, ed. Phantoms Foundation (2008).

Advertisement

28

nanoICT research

o PhD Position (Centre Investigacions Nano-ciencia Nanotecnologia (CSIC-ICN)(CIN2),Spain): "Quantum Transport in Carbon Nanotubes"

The research will be focused on the electrical and me-chanical properties of structures with dimensions of afew nanometers, such as carbon nanotubes or gra-phene. These structures are so tiny that quantum effectsstart to play a dominant role. For e.g. the energy levelsare quantized, just like in atoms and molecules. Interes-tingly, these structures are large and robust enough tobe implemented in a variety of different microfabrica-ted devices, which allow the tuning of their quantumproperties.This work will be carried within the Quantum Nano-Electronics group, which is part of the Centre Investi-gacions Nanociencia Nanotecnologia (CSIC-ICN). Thegroup is located at the Campus Universitat Autonomade Barcelona. The funding for the salaries comes froman EURYI award.

The deadline for submitting applications isApril 08, 2010For further information about the position, please con-tact: Adrian Bachtold (adrian.bachtold@cin2.es)

o PhD Position (CIC nanoGUNE, Spain): "Na-nomagnetism"

The Nanomagnetism Group at the CIC nanoGUNE Re-search Center in San Sebastian (Spain) is currently loo-king for a motivated PhD Student (Physicists/MaterialsScientists/Engineers) who is interested in working in aninspiring international and interdisciplinary environmentin the following research areas:

• Time resolved magneto-optic Kerr microscopy withsubnanosecond time resolution for studying the mag-netization dynamics (e.g., fast magnetic switching).

• Laterally patterned magnetic nanostructures that arefabricated by means of electron beam lithographyand thin film and multilayer deposition techniques.

The deadline for submitting applications isMarch 31, 2010For further information about the position, please con-tact: Paolo Vavassori (p.vavassori@nanogune.eu)

o Job Vacancy (CIC nanoGUNE, Spain): "Elec-tron Microscopy Technician"

CIC nanoGUNE Consolider, located in San Sebastian,Basque Country (Spain), is a R&D center created re-

cently with the mission of conducting basic and appliedworld-class research in nanoscience and nanotechno-logy, fostering training and education excellence, andsupporting the growth of a nanotechnology-based in-dustry. An electron microscopy facility is currently being esta-blished to accompany nanoGUNE´s already ongoing re-search activities by means of a high-level structuralcharacterization laboratory, including Cs correctedTEM, FIB and ESEM. In order to provide qualified assis-tance to the Staff Microscopist that is managing thesefacilities, we have an immediate opening for an ElectronMicroscopy Technician. We are searching for a highlymotivated and creative person with a solid technical/en-gineering background, mechanical and electronic designand construction skills, and experience in operating andmaintaining TEM sample preparation facilities.

The deadline for submitting applications isMarch 01, 2010For further information about the position, please con-tact: nano@nanogune.eu

o PhD Position (University of Castilla — LaMancha, Spain): "Ultrafast spectroscopy of cyclodextrin-based drug nanocarriers"

The Laboratory of Ultrafast Spectroscopy at the Uni-versity of Castilla — La Mancha, Toledo, Spain under thesupervision of Prof. Abderrazzak Douhal is seeking anoutstanding PhD candidate to carry out studies oninteraction of drugs with cyclodextrin-basednanocarriers using steady state and fs- to ms-time re-solved spectroscopic techniques of absorption and emis-sion, as well as time- and space resolved single moleculespectroscopy.

The deadline for submitting applications is Fe-bruary 28, 2010For further information about the position, please con-tact: Abderrazzak Douhal (Abderrazzak.Douhal@uclm.es)

o PostDoctoral Position: "Ultrafast spectroscopy ofcyclodextrin-based drug nanocarriers"

A postdoctoral position for a period of 18 months isavailable at the University of Castilla — La Mancha, To-ledo, Spain. The Laboratory of Ultrafast Spectroscopy at the Uni-versity of Castilla-La Mancha (Toledo, Spain) is lookingfor an outstanding postdoctoral candidate for a Euro-pean-funded research project (Cyclon, FP7, Marie CurieITN Network) based on the use of ultrafast techniques

29

nanovacancies

to study the interaction of drugs with cyclodex-trin-based nanocarriers. The Laboratory of Ultra-fast Spectroscopy is equipped with femtosecondfluorescence up-conversion, transient-absorption andtime- and spectrally-resolved single-molecule fluores-cence microscopy techniques.

The deadline for submitting applications is Fe-bruary 28, 2010For further information about the position, please con-tact: Abderrazzak Douhal (Abderrazzak.Douhal@uclm.es)

o Scholarships (National Science Council ofTaiwan (NSC): "10 scholarships will be allocated to Spa-nish graduate students to undergo 2-month summer trainingsin Taiwanese research institutes"

In order to promote a further collaboration with Spa-nish Education and Research authorities, the NationalScience Council of Taiwan (NSC) intends to provide 10Spanish graduate students - PhD or Master students inscience and social sciences and humanities - of first-handresearch experience in Taiwan, an introduction to itsscience and technology policy infrastructure, and orien-tation to the culture and language. The goals of the program are to introduce students toTaiwanese scientific and technological development inthe context of a research laboratory, and to initiate per-sonal relationships that will better enable them to co-llaborate with foreign counterparts in the future.

The deadline for submitting applications is Fe-bruary 15, 2010For further information about the position, please con-tact: Charles Peng (fran01@nsc.gov.tw)

o Scientist Positions (Madrid Institute for Ad-vanced Studies of Materials, Spain):

"Virtual Processing of Structural Composites" Virtual Processing of Structural Composites, includingnumerical simulation of curing and infiltration and designof optimum preforms (braiding, stitching, 3D weaving,knitting).

"Computational Design of Metallic Materials"Computational Design of Metallic Materials to predictthe influence of alloy composition and processing routeon the microstructure (phases, precipitation, segrega-tion, grain size and shape, texture, etc.) through advan-ced numerical modeling (computational thermodyna-mics, phase-field modeling, etc.).

"Advanced Processing of Structural Composites"Advanced Processing of Structural Composites with par-ticular emphasis on out-ofautoclave consolidation of pre-pegs, pultrusion, VARTM, RFI, and processing ofthermoplastic composites (glass-mat strands and longfiber reinforced thermoplastics direct-process).

"Processing of Metallic Materials"Processing of Metallic Materials to develop innovativestrategies and to improve current methods of manufac-turing near net-shape components for high temperaturestructural applications (Ni-based superalloys, intermeta-llics) and light weight alloys (Mg, Al) for the aerospaceand automotive sectors. Techniques of particular inte-rest include casting, secondary processing (forging, ex-trusion), and novel methodologies (rapid prototyping, electron-beam melting, direct-metal laser-sintering, etc.).

The deadline for submitting applications is Fe-bruary 04, 2010For further information about the position, please con-tact: IMDEA-Materials (https://www.imdea.org/internationalcall)

o PostDoctoral Position (Groupe Nanoscien-ces - CEMES, France): "Investigation of molecular ad-sorption on SiC surfaces by UHV-NC-AFM and STM"

The successful candidate will participate in a projectwhose main objective is to use silicon carbide recons-tructed surfaces as templates for adsorbed molecules, ina overage ranging from isolated molecules to one mo-nolayer. He will be in charge of experimental studiesbased on Scanning tunneling microscopy (STM) and ato-mic force microscopy (AFM) in the non-contact mode.Two STM/AFM microscopes will be used. A room tem-perature STM/cantilever-based AFM and a low tempe-rature (6 K) STM/tuning fork-based AFM, dependingon the experimental requirements. The Nanosciencesgroup in CEMES comprises experimental physicists, che-mists specialized in the synthesis of molecules for na-noscience and theoreticians with a strong background inSTM and AFM image calculations. The candidate will clo-sely collaborate with the chemists and the theoreticiansto optimize the molecules and for image interpretationand calculation.

The deadline for submitting applications is Ja-nuary 31, 2010For further information about the position, please con-tact: Sébastien Gauthier (Gauthier@cemes.fr)

More vacancies available at: http://www.phantomsnet.net/Foundation/jobs.php?project=1

30

nanovacancies

32

EU report

AAttoommiicc && MMoolleeccuullaarr SSccaallee DDeevviicceessaanndd SSyysstteemmss aanndd BBiioo--CChheemmiissttrryyBBaasseedd IInnffoorrmmaattiioonn SSyysstteemmss

Expert Consultation Workshop: 22 & 23 October2009, Brussels (Belgium).

Executive Summary

Future and Emerging Technologies − Proactive Initiati-ves (FET − Proactive) is preparing the next work pro-gramme for 2011—2012 and later. As part of the processto identify new grand research challenges, FET organi-sed a workshop on 22 and 23 October 2009, in Brus-sels, on the following two themes: Atomic & MolecularScale Devices and Systems, and Bio-Chemistry Based In-formation Systems. Around 20 researchers covering thedifferent relevant areas of research attended the meeting.The objective was to find new topics for future challengingand promising FET proactive initiatives for Information andCommunication Technologies (ICT). The meeting cove-red also the possible affinities between its two themes,such as the emergent field of quantum effects in biolo-gical systems. This report summarises the discussionsthat took place during the workshop and gives a list ofproposed challenges for the work programme.

1. Introduction

FET (Future Emerging Technologies) structures researchin a number of proactive initiatives, which typically con-sist of a group of projects funded around a commontheme. The themes are shaped through interaction withthe research community, and focus on novel approa-ches, foundational research and initial developments onlong-term research and technological innovation. Thisworkshop discussed how both atomic & molecular scalesystems, and bio-chemistry based information systemscan be investigated to push for a new generation, andeven paradigm, of information and communication tech-nologies at a new and smaller scale, exploring both thequantum and the biological effects offered by nature.The expected outcomes from the meeting were:

• To obtain visionary scenarios for atomic scale andbio-chemical ICT systems and what is needed to de-velop, test and implement them.

• To identify the core challenges, their impact and timescale.

• To identify the adequate target research communities.

• To produce an outline of funding schemes and coo-peration strategies to be deployed.

2. Synthesis of the discussions

The workshop included discussions amongst the mem-bers of both the Atomic & Molecular Scale Devices andSystems group and the Bio-chemistry Based InformationSystems group, as well as joint discussions, namely toidentify potential joint challenges. The following aspectswere covered:

• The challenges faced and identified in these areas, na-mely as potential FET research topics

• The impact of successful projects on both society andresearch communities

• The identification of the core communities thatshould be involved

• The timescales for the uptake

• The instruments of funding most relevant

• International cooperation

Each of these points is discussed below giving a sum-mary of the views expressed during the discussions.

In particular, the first four points are presented for bothworkshop themes: Atomic & Molecular Scale Devicesand Systems, in section 3, and Bio-Chemistry Based In-formation Systems, in section 4. In section 5, the possi-ble affinities and overlaps between these two themesare discussed. Finally the funding instruments and in-ternational cooperation are discussed in section 6. As additional resources, the list of participants in theworkshop is presented in annexe I, the terms of refe-rence of the meeting are presented in annexe II, thewritten contributions of the experts consulted are avai-lable at: ftp://ftp.cordis.europa.eu/pub/fp7/ict/docs/fet-proactive/wsconsult09-amolbio-02_en.pdf

3. Atomic & Molecular Scale Devices andSystems

The technological revolution brought about by the de-velopment of miniaturized information processing devi-

ces over the last 50 years has had an extraordinary im-pact on almost every aspect of our daily lives. At presentessential properties of these technologies may still beunderstood semi-classically whereby quantum mecha-nical properties enter at the level of global material pro-perties, but not as a basic functional resource. However,miniaturization is now reaching the quantum frontierwhere quantum properties are playing an essential rolefor the behaviour of systems that are either built, bot-tom-up, from individual atomic or molecular compo-nents or, top-down, from bio-molecular systems createdthrough evolution. Nanofabricated devices can now bebuilt to operate at this quantum frontier and bio-mole-cular systems can now be shown to exhibit quantummechanical features in their dynamics.

Hence, quantum mechanics, the theory of the verysmall, is widely appreciated to be at the core of the nexttechnological revolution and the development of me-thods for the creation, manipulation and control of ato-mic, molecular and bio-molecular devices at thequantum-classical frontier holds the promise of turningthe quantum behaviour of such structures to our ad-vantage, thus providing the foundations of atomic scaleinformation technologies for the 21st century.

3.1. Research topics addressed (based on theparticipants' written contributions and pre-sentations)

A large number of topics were addressed and discussedduring the workshop, namely in the Atomic & Molecu-lar Scale Devices and Systems break-out sessions. Takingalso into account the written contributions, there emer-ged some common themes and topics, which the ma-jority of contributors addressed and discussed. Thesemay be summarised as being:

• Fabrication at the atomic scale

• Controlling and probing single atoms and molecularscale systems

• Embedding and connecting atomic and molecularquantum systems

• Understanding and utilizing the local nanoscale envi-ronment

• Experimental exploration of the quantum-classical in-terface

• Theoretical and numerical methods for modelling atthe quantum-classical border

• Exploiting new functionalities in sensing, memory,logic and energy harvesting

• New information exchange mechanisms (spin, mag-netism, light, etc.)

• Scalability and production methods

• Controlled doping at the nanoscale

• Efficient information extraction

• Thermodynamics at the molecular scale

• Quantum effects in biology

• Bio-inspired new paradigms of information codingand processing

It was also suggested that a future call on this themewould be better named Atomic Scale Technologies.

3.2. Challenges

Several important challenges were identified. Some ad-dressed fundamental issues, while other focused on fu-ture applications of atomic-scale systems. Fourinterconnected challenges were mentioned:

1. ICT functionality from atomic and molecular designusing top-down, bottom-up or a combination of bothapproaches

• Designing functionality from the guided assembly ofatoms and molecules or by exploiting bio-moleculesrealized by nature

• Improving existing ICT technologies by pushing thelimits of miniaturization to the atomic scale

• Characterizing atomic and molecular scale compo-nents

• Providing the routes towards cost effective fabrica-tion at the atomic or molecular scale

2. Embedding and interfacing of atomic and moleculardevices within a technological environment

• Embedding atoms and molecules in a controlled en-vironment: fabrication and connection

• Engineering stability, uniformity and robustness

• Interconnecting atomic and molecular devices

33

EU report

• Interfacing atomic and molecular systems with theoutside world, e.g. light-matter connectivity, electro-nic transport, spin resonance, mechanical response,etc.

• Investigating the effect of operation on device beha-vior

3. Exploit coherence, decoherence and the quantum-classical boundary for ICT

• Enhancing coherence to take advantage of the syste-m’s quantum behaviour

• Enhancing decoherence to turn the system classical inthe shortest possible time

• Exploiting the interplay between decoherence andcoherence for optimized performance in atomic, mo-lecular and bio-molecular systems

• Exploiting dynamical control of molecular and ato-mic systems

• Exploiting coherence for sensing and metrology

• Developing techniques to probe and extract infor-mation efficiently from such systems

4. Develop experimental models and theoretical me-thods for atom, molecular and bio-molecular systemsat the quantum-classical boundary

• Developing proof-of-concept experimental modelsystems for component functionalities

• Developing theoretical methods to describe multi-component connected quantum systems interactingwith their environment

3.3. Impact

Developing atomic and molecular scale ICT devicestakes us to the ultimate limit of miniaturization, wherequantum physics can give us improved speed, precisionand efficiency, as well as new functionalities, therebycontributing to the ongoing information revolutionwhich benefits virtually all areas of society.

In particular, a profound impact is expected in the fo-llowing two areas:

Science and industry: namely with the new ability tobuild and control multiscale systems, new measurement

systems, new standards, new tools and ICT devices. Aprogramme in this area would help define and give a le-ading role to the EU in molecular scale engineering andelectronics.

Society: namely in increased and new areas of em-ployment, improved man-machine interactions − in par-ticular sensors (health, safety, environment and qualityof life), improved energy efficiency and energy harves-ting/scavenging from new sources.

3.4. Core Communities

These were identified as being the communities that canhave an input in the project calls, in order to achieve themaximum benefit to the wider community and to en-sure that all relevant technical areas are able to contri-bute to the projects in the most effective manner. The core communities addressed by this call belong tothe following fields:

• Physics

• Electronics and nano-science

• Chemistry

• Molecular biology

• Surface science

3.5. Timescale for uptake

The following short, medium and longer-term timesca-les were identified to meet the main challenge of pro-gressing from concept to marketable molecular scaleinformation systems:

• A shorter term of around five years is expected to bethe timescale for sensors (but not single moleculesensors), basic proofs-of-concept and possibly somesimpler molecular based devices.

• In the medium term (5-10 years) one could expectthe development of new measurement systems andmeasurement references/standards to support thedevelopment of molecular electronics and their cha-racterisation to emerge.

• In the longer term (10+ years) ICT using moleculartechnology which complements and integrates withCMOS would emerge.

34

EU report

4. Bio-chemistry Based Information Systems

Living cells and organisms input, manipulate and pro-cess information, and they do this in a way very differentfrom our conventional digital computers. The informa-tion is embodied in complex biochemicals, and infor-mation processing includes the modification,construction, and replication of these chemicals. Ourabilities to exploit these processes are still at an earlystage. But once we have an understanding and controlof biochemical information processing, we will be ableto affect and change processes in living cells, with ob-vious medical applications. Furthermore, we will be ableto develop analogous processes in non-living systemsand engineer new materials with an enormous range ofnovel physical and information processing capabilities. Developing fundamental understanding, theories, tech-niques, and tools of embodied biochemical informationprocessing, requiring novel inputs from computerscience and from biochemistry, is crucial to developingthese capabilities.

4.1. Research topics addressed (based on theparticipants' written contributions and pre-sentations)

The Bio-chemistry Based Information Systems break-outgroup discussed a large and diverse range of topics,which are summarised below, taking also into accountthe written contributions:

• Processes and applications targeting living cells andorganisms

• Self construction, self assembly, self organisation, selfregulation, self destruction, feedback (internal, andvia the environment) and emergence in bio-chemicalsystems, from the molecular scale upwards

• Making matter (biological or non-biological) “smar-ter”

• The importance of geometric issues (space, surface,scale, structure)

• The source of the information processing power:construction, replication, self-assembly, auto-cataly-tic networks

• The “self-consistent” (non-arbitrary) relationship bet-ween information and function

• Theoretical underpinnings will be more physical thanlogico-mathematical

• Abstraction of the functionality of molecules, andreimplementation in a different framework

• Developing appropriate computational and informa-tion processing models for these novel paradigms andabstractions, that are inherently stochastic and “soft”

• Interfacing with other microscopic and with macros-copic systems

• Safety as an architectural systems issue, not (just) acomponent issue

4.2. Challenges

The following 20-year vision was proposed for Bio-che-mistry Based Information Systems: a mainstream tech-nology based on programmable self-organising systemsin which the generic components are physical construc-tors.

The overall challenge is to make substantial and signifi-cant steps in the directions of the vision: to design andprogram self-organising construction systems down tothe molecular scale, that are able to perform informa-tion processing. This is broken down into the followingrequirements:

1. Theories, methods, tools, languages, architecturesand simulations allowing self-organised systems to beprogrammed

2. Implementations of programmable physical systemsNamely at levels including physical self-organisingsystems, synthetic chemistry, supra-molecularsystems, reaction networks, sub-cellular/cellularsystems, organism (single robots), and swarm robots

3. Nature-inspired design and fabrication processes forrobust self-modifying systems

4. Environment and interface:

• Flexible reaction containers, flexible geometry, finegrained control of reactor topology, interplay bet-ween reactions and topology, for information pro-cessing, content sensitive systems

• Biology inspired strategies of information coding andexchanging

4.3. Impact

Bio-chemistry Based Information Systems is a rather

35

EU report

novel area and its impact is not easy to predict, but it of-fers the potential for ICT disruptive technology. It willprovide novel sensors, interfaces, recognition systems,and actuators, with clear applications for health, the en-vironment, and security. It offers new paradigms and ar-chitectures for computer science, of embodieddecentralised pervasive information processing, as wellas for robotics and its applications, which play such an im-portant role in our society. It promises a range of com-pletely novel “smart” materials that combine and exploitboth physical properties and information processing ca-pabilities. It provides a radical new outlook on sustaina-bility of products that can be locally constructed and thensustainably dismantled when no longer needed.

4.4. Core Communities

These were identified as being those communities thatwill need to have an input in to the project calls so as toachieve the maximum benefit to the wider communityand to ensure that all relevant technical areas are ableto contribute to the projects in the most effective man-ner. The core communities need to be:

• Chemistry

• Biochemistry

• Molecular biology

• Physics

• Computer science

• Robotics

• Electronics

• Surface science

4.5. Timescale for uptake

An overarching 20-year vision was proposed for Bio-che-mistry Based Information Systems, with four challengesthat are four parallel strands that all have to be achievedto realise the vision.

Each strand has steps along the way to maturity, and thefollowing global timescale was envisioned:

• After 5+ years, there should be laboratory prototy-pes (that solve most of the scientific problems)

• After 10+ years, there should be commercial pro-totypes (that solve most of the scaling-up and engi-neering problems)

• After 20 years, there should be ICT consumer pro-ducts.

5. Atomic & Molecular vs. Bio-chemicalbased ICT

During the meeting were also discussed potential topicsaddressing both atomic & molecular and bio-Chemistrybased information systems. The discussion has shownthat we are still at a very early stage of a path that couldreveal itself promising and technologically disruptive.This workshop has triggered a process by identifying po-tential directions exploring the affinities between thesetwo areas, at an intermediate scale still little explored. Inparticular, the following topics were suggested:

• Guided self-assembly in a practical environment

• Connectivity and interfacing

• Advanced classical and quantum simulation of com-plex non-modular systems

• Nature-based optimization of information coding andprocessing

• Bio-molecular behaviour at the quantum-classicalboundary

It was noted that the meeting could have benefited fromthe presence of more biology experts, including for la-ying out further perspectives for the emergent area ofquantum effects in biological systems and its conse-quences for biology. Altogether it was suggested itwould be important to further explore this path, pos-sibly through an FP7 Coordination and Support Action,to help build the respective research community and es-tablish its challenges.

6. Instruments and International Coope-ration

6.1. Instruments

The various types of instruments open for considera-tion were briefly discussed, as well as possible modifi-cations to those mechanisms. Both for Atomic &Molecular Scale Devices and Systems (which, it was sug-gested, should be renamed Atomic Scale Technologies)and for Bio-chemistry Based Information Systems, it wasgenerally thought that smaller projects were the mostappropriate for these novel areas, but it was noted that

36

EU report

their three-year lifetime may be too short for the typeof challenges envisioned. The possibility of having callsthat would accept projects of any size, competing forthe same budget, was also suggested. Regarding the emergent area overlapping between thetwo main themes of the workshop, namely quantum ef-fects in biological systems, it was pointed out that ins-truments such as training network would be a veryimportant step to train young European researchers andinterweave the community in these emerging researchfields. Such a structure could be the basis for subsequentSTREP-like projects.

Finally, it was noted that Coordination and Support Ac-tions could be explored for two purposes: to help de-velop the more embryonic topics, and to buildconnections towards related FET initiatives.

6.2. International Cooperation

The members of the workshop briefly discussed therole of international cooperation. It was considered thatit is in general beneficial for EU research groups to co-operate with top research groups/centres from othercountries, namely from the USA, Japan and Australia,amongst others, specially in the novel and emergentareas, where there are still not that many groups in theworld. No national programs that could complementFET schemes were identified.

Annex I: List of Participants

Robert Baptist CEA Leti Erik Dujardin CEMES Jerzy Gorecki Polish Academy of Sciences Thomas JungPaul Sherrer Institute Zoran KonkoliChalmers University of TechnologyJohn McCaskillRuhr University, Bochum Barry McMullinDublin City UniversityYasser OmarTechnical University of Lisbon (rapporteur)

Vittorio PellegriniScuola Normale Superiore, PisaMartin PlenioTechnical University of Ulm Alfonso Rodriguez-PatonUniversidad Politécnica de Madrid Jörg Schmiedmayer Tech. University of Vienna (AMS moderator)Pierre Seneor THALES Michele Simmons Centre for Quantum Computer TechnologySusan Stepney University of York (Bio-Chem moderator) Herre Van der Zant TU Delft Ian Walmsley University of OxfordLukas Worschech University of Würzburg Klaus-Peter Zauner University of Southampton

Annex II: Terms of Reference

Atomic & Molecular Scale Devices andSystems and Bio-Chemistry Based Informa-tion Systems (22-23 October 2009, Brussels)

1. Context

The ICT Future and Emerging Technologies (FET) Pro-active scheme fosters frontier research that opens upnew avenues across the full breadth of future infor-mation technologies. FET acts as a pathfinder promo-ting the exploration of radically new ideas and trendsfor future research and innovation and provides sustai-ned support to emerging areas that require fundamen-tal research to be carried out over a long period. It aimsto go beyond the conventional boundaries of ICT andventures into uncharted areas, often inspired by and inclose collaboration with other scientific disciplines. Ra-dical breakthroughs in ICT increasingly rely on freshsynergies, cross-pollination and convergence betweendifferent scientific disciplines.

One research area pursued within FET is frontier rese-arch at the atomic and molecular scale which re-quires both operating in a world dominated by quantum

37

EU report

effects and linking it to the environment, be it electro-nic, mechanical, optical, organic or of another type. Thisconvergence of scale and function calls for interdiscipli-nary collaborations and combination of knowledge froma wide range of sciences.

FET also embraced the emerging research field of bio-chemical and bio-inspired information proces-sing. This denominator covers research challenges, suchas e.g.; how to exploit nature's information processingcapabilities, how to address evolution, (bio-) immersivecomputing, interfacing traditional silicon-based IT and(immersed) wetware, how to deal with formalisms, etc.

Fundamental research in the area of nanoelectronics,quantum information and bio-chemistry has been pur-sued in FET since FP6, indicating the sustained and gro-wing importance of these areas (for more details seewebsite http://cordis.europa.eu/fp7/ict/fet-proac-tive/areas_en.html):

• FP6, call 3: Emerging Nanoelectronics (E-Nano) aimed at advancing research in hybrid andmolecular electronics, and prepared the bases for anextension of integrated circuit technology beyond thelimits of CMOS scaling. The selected projects startedin September 2005.

• FP7, call 1: Nano-scale ICT Devices andSystems (Nano-ICT) aimed at demonstrating un-conventional solutions to increase computing per-formance, functionality or communication speed, orto reduce cost, size and power consumption of ICTcomponents beyond the expected limits of CMOStechnology. The selected projects started in January2008.

• FP7, call 4: Bio-Chemistry-based informa-tion technologies (Chem-IT) aimed at develo-ping the foundations for a radically new kind ofinformation processing technology, inspired by che-mical processes in living systems. The selected pro-jects are under negotiation and are supposed to startin February 2010.

• FP7, call 4: Quantum Information Founda-tions and Technologies (QI-FT) aimed at rese-arch in quantum information theory, entanglement-enabled quantum technologies, scalability of quan-tum processing systems and long distance quantumcommunication. The selected projects are under ne-gotiation and are supposed to start in February 2010.

2. Objectives of the Consultation Workshop

This consultation is targeted to scientists interested inatomic & molecular scale devices and systems, in quan-

tum information foundations and technologies, in bio-chemistry based information systems, in the move to-wards zero-power ICT and in the interconnection oftera-scale devices to shape monolithic informationsystems. The purpose is to identify new topics for fu-ture challenging and promising FET proactive initiatives.This meeting will concentrate on topics related to Ato-mic & Molecular Scale Devices and Systemsand Bio-Chemistry Based Information systemswith a view on those grand challenges that should beincluded in the FET work programme 2011 — 2012 orlater to tackle them. The contributions and discussionare expected to serve at the level of research program-mes rather than at the level of single project ideas. Sug-gestions regarding the means to implement the researchare also welcome — whether the projects should be bigor small, or whether networking or coordination bet-ween researchers should be fostered.

38

EU report

http://cordis.europa.eu/fp7/ict/fetproactive/calls_en.html