Background of the Invention Technology Computer Aided Design (TCAD) tools are used for simulating electronic devices at transistor level. A typical electronic device is a logical gate. This device has three terminals: source, drain and gate. The voltage (V gate) at the gate terminal controls the source-drain current (I).

The purpose of the simulations is to predict the l-V gate characteristics of different designs and materials. Such simulations require a plurality of parameters that describe the material properties of the device. One of the most important material properties in a device is its band structure. The band structure describes the relation between the electron momenta and electron energy for the material properties of the device. A common approximation for the device band structure is the effective mass approximation, where the band structure is described by a second order polynomial around the band minimum. This may be written as: where s _nk is the electron energy of band number n, (k _x , k _y , k _z ) is the electron wave vector, h is the reduced Planck constant, ε is the energy of the band minimum, (k _x0 , k ₀ , k _z0 ) is the wave vector of the band minimum, and (tn _x , m _y , tn _z ) are the anisotropic effective masses. Different variants of the effective mass approximation include deviations from the second order polynomial approximation through non- parabolicity effects.

Commonly, the effective mass model for a device is obtained by fitting the parameters of the polynomial to the device material band structure when the material is in an infinite periodic crystal. This infinite periodic crystal is also called "bulk material". The band structure of a bulk material can either be calculated using electronic structure methods such as Density Functional Theory (DFT), or obtained experimentally e.g. using photoemission or inverse photoemission.

Summary of the Invention

The present disclosure relates to band structure calculations obtainable using electronic structure methods and to the extraction of approximate band structure models from the calculations.

In a nanoscale device, the electrons are confined by the device geometry and the confinement changes the band structure of the electrons; this is called a confined system or confined device, which is periodic at least in one direction. Therefore, the values of the effective mass and non-parabolicity parameters determined from a bulk system, cannot be used to construct an accurate description of the confined device. It can be very difficult, costly, and sometimes impossible, to obtain reliable experimental data for such parameters for relevant device geometries e.g. a confined nanodevice.

An alternative approach may be generating band structure data using atomic- scale modelling e.g. First Principles Methods (FPM). One could also use semi- empirical tight-binding methods. The FPM attempt to solve the basic Quantum Mechanical equations, i.e. the Schrodinger Equation, of the system without using any empirical parameters. However, it is not possible to solve the Quantum Mechanical equations exactly, and so a number of different approximations exist. The most popular approximation is DFT. The semi-empirical methods are a simplified version of the FPM, i.e. a number of empirical parameters are introduced. These parameters are determined such that the model gives a good description of the band structure in a number of reference systems.

For complex nanoscale devices, the calculated band structure data can be very complex, and it can be difficult to extract an effective mass description as needed for the TCAD tools. The use of atomic-scale modeling has become a standard tool for generating TCAD parameters, and consequently there is a need of robust and automatized procedures for extracting effective mass models from atomic-scale band structure data.

Extracting an effective mass model from the calculated band structure of a bulk system is a standard fitting procedure. However, extracting the effective mass model from a confined system may be complex or even not possible. In a confined device with at least one transport direction, the electron momentum has a continuum of values along the transport direction, i.e. the direction on which the electron must travel to move from the source to the drain electrode. In the remaining directions perpendicular to the transport direction, the system is said to be confined. In those perpendicular directions i.e. the confined directions, the electron momentum has discrete values. Each discrete value of the electron momentum in a confined direction gives rise to a plurality of sub-bands in the transport direction. Consequently, the effective mass band structure model may be defined by several sub-bands. These series of sub-bands are called band ladders. There may be several effective mass band ladders in the band structure of a confined system, each ladder related to a single electronic band of the corresponding bulk system and each described by a separate effective mass model. Put differently, band ladders may be interleaved, making the resulting device band structure very complex. There is currently no known system or method which can take the interleaved band structure of a confined device and construct an effective mass model that reproduces it.

In a first aspect, a computer-aided method for simulating a confined nanodevice is provided. Firstly, an atomic scale geometry specifying dimensions and material of the confined nanodevice may be generated. Secondly, a band structure comprising wave functions and Eigen energies may be calculated by FPM. Then, an effective mass model comprising wave functions and Eigen energies may be generated. Afterwards, the spatial wave functions of FPM and effective mass band structures may be mapped to each other. In response to such mapping, parameters may be fitted by curve fitting, i.e. a process of constructing a curve or mathematical function that has the best fit to a series of data points, to fit generated effective mass model energies to FPM energies so that an approximate electronic structure of the confined material may be modeled.

The implementation of methods according to this aspect enables the identification of the interleaved band ladders of confined nanodevices and therefore a more precise band structure for TCAD tools simulation may be obtained. Additionally, the method provides an accurate description of the confined nanodevice around the band minimum. By using this method a simulation as fast as one performed with TCAD tools and as accurate as one performed with DFT simulations may be achieved.

In some examples, the mapping step comprises providing Eigen energies and wave functions of both FPM and effective mass band structures and then, matching an unmatched FPM band with an effective mass band. The matching step may be repeated untill all relevant bands are matched. Finally, a list where each unmatched FPM band is matched with an effective mass band may be returned.

In some examples, the matching step may comprise identifying a lowest energy FPM band which is not matched and identifying also a lowest energy effective mass band. Then, an FPM wave function with the same local atomic behavior as the unmatched FPM wave function with lowest energy and global behavior of another effective mass wave function may be identified.

In a second aspect, a computer program for carrying out a method in accordance with any of the preceding examples is provided. The computer program may be embodied on a data medium (for example, a CD-ROM, a DVD, a USB drive, on a computer memory or on a read-only memory) or carried on a carrier signal (for example, on an electrical or optical carrier signal).

The computer program may be in the form of source code, object code, a code intermediate source and object code such as in partially compiled form, or in any other form suitable for use in the implementation of the method. The carrier may be any entity or device capable of carrying the computer program. For example, the carrier may comprise a storage medium, such as a ROM, for example a CD ROM or a semiconductor ROM, or a magnetic recording medium, for example a hard disk. Further, the carrier may be a transmissible carrier such as an electrical or optical signal, which may be conveyed via electrical or optical cable or by radio or other means.

When the computer program is embodied in a signal that may be conveyed directly by a cable or other device or means, the carrier may be constituted by such cable or other device or means.

Alternatively, the carrier may be an integrated circuit in which the computer program is embedded, the integrated circuit being adapted for performing, or for use in the performance of, the relevant methods.

In a third aspect, a computer system on which a computer program according to an example is loaded. The computing system may comprise a memory and a processor, embodying instructions stored in the memory and executable by the processor, the instructions comprising functionality to execute a method of simulating a confined nanodevice according to some examples disclosed herein.

The implementation of such systems provides robust and automatized procedures for extracting effective mass models from atomic-scale modeling, which may be used to simulate band structure data.

Brief Description of the Drawings

Various objects, features, and advantages of the disclosed subject matter can be more fully appreciated with reference to the following detailed description of the disclosed subject matter when considered in connection with the following drawings, in which like reference numerals identify like elements.

FIG. 1 is a flow diagram illustrating a method to obtain an approximate band structure model from an accurate atomic-scale band structure calculation according to an embodiment. FIG. 2 illustrates an atomic-scale structure of a 1 -dimensional nanodevice geometry according to an embodiment. The spheres represent atoms such as Hydrogen, Indium and Arsenic atoms, and the lines connecting the spheres represent chemical bonds.

FIG. 3a shows a DFT band structure calculation of the nanodevice geometry in FIG. 2.

FIGS. 3b and 3c show amplified portions of FIG. 3a for different electron momenta.

FIGS. 3d and 3e show projection function values for different effective mass bands and DFT bands where the highest value of each case is circled.

FIG. 4 illustrates the smoothed confinement potential used to generate the effective mass model of the structure in FIG. 2 according to an embodiment. The dots show the value of the confinement potential for a given x and y position. The potential is close to zero outside the atomic regions and negative inside the atomic regions.

FIG. 5 shows a flow diagram of matching DFT bands with bands in the effective mass model, according to an embodiment.

FIG. 6 illustrates real space wave functions calculated with the DFT model (602, 604, 606, 608) and with the effective mass model (610, 612, 614, 616) in accordance with an embodiment.

FIG. 7a illustrates an effective mass ladder. This ladder is formed by four DFT sub-bands around momentum k=0.

FIG. 7b illustrates another effective mass ladder. This second ladder is formed by three DFT sub-bands, again around momentum k=0.

FIG. 7c illustrates another effective mass ladder. This third ladder is formed by another three DFT sub-bands again around momentum k=0.

FIG. 8 illustrates the combined fit of the three effective mass models of FIGS. 7a-7c to the DFT band structure of FIG. 3 around momentum k=0. FIG. 9 illustrates a computer system having a processor with different processing modules and a memory according to an embodiment. The atomic-scale geometry resides in memory and the different processing modules handle the separate parts of the algorithm.

Detailed Description

In the following description, numerous specific details are set forth regarding the systems and methods of the disclosed subject matter and the environment in which such systems and methods may operate, etc., in order to provide a thorough understanding of the disclosed subject matter. It will be apparent to one skilled in the art, however, that the disclosed subject matter may be practiced without such specific details, and that certain features, which are well known in the art, are not described in detail in order to avoid complication of the disclosed subject matter. In addition, it will be understood that the examples provided below are exemplary, and that it is contemplated that there are other systems and methods that are within the scope of the disclosed subject matter.

The present invention is directed to a system and method for extracting an approximate band structure model from an atomic-scale band structure calculation of a confined system. The confined systems of most relevance to this invention are nanoscale devices, for instance, the Fin of a Fin Field Effect Transistor (FinFET) device. Such a Fin consists of a material, for instance, Silicon, with nanoscale dimensions in two directions, the confined directions. The aim is to construct an effective mass model (or similar approximate band structure model) for the electronic structure of the confined system e.g. the Fin of the FinFET, which can be used for a TCAD simulation of the confined system e.g. the Fin of the FinFET. The invention is a method for constructing such an approximate band structure in an automated way from an atomic-scale band structure calculation using first principles methods or semi-empirical methods.

The standard approach for generating an approximate band structure model for a confined nanosystem is to use parameters generated for a bulk system. The bulk system is periodic in all directions and therefor has a high symmetry. Due to the high symmetry, there are only a few bands. Also, it is relatively simple to fit an effective mass model to the band structure.

A confined system, as a consequence of the confinement, has a much more complex band structure, which consist of series of interleaved sub-bands which form ladders. The invention describes an automated method for mapping these sub- bands into e.g. different effective mass ladders.

FIG. 1 is a flow diagram illustrating method to obtain an approximate band structure model from an accurate atomic-scale band structure calculation. The objective is the generation of an approximate band structure model for a user defined device geometry using atomic-scale simulations. The steps in the algorithm will be illustrated for a particular 1 -dimensional confined system, a 3x3 nm Indium- Arsenic nanowire. The atomic-scale geometry of the nanowire cross-section is shown in FIG. 2. Items 202 are Hydrogen atoms, items 204 are Indium atoms and items 206 are Arsenic atoms. The Indium or Arsenic atoms positioned at the surface of the device do not have neighbor atoms in all directions. In those directions where a neighbor is missing, they are bonded with a Hydrogen atom. Such passivation of the atoms is a standard procedure for the atomic geometry to better describe the experimental situation where the nanowire will be capped with some material that will passivate the surface atoms. The invention is not limited to this system, but can be used for 1 -dimensional nanowires of any other suitable dimensions and materials, or 2-dimensional nanoscale systems which are confined in one or more directions, or any other suitable N-dimensional nanoscale systems.

In FIG. 1 , the method starts at step 102 where the user inputs the target device geometry. In step 104, the target geometry is converted into an atomic-scale structure. Typically, the target device geometry will specify dimensions and materials of the device, while in the atomic-scale structure the position and element type of each individual atom in the device are specified. The example of an Indium-Arsenic nanowire atomic-scale geometry is shown in FIG. 2. FIG. 2 shows a cross-section of an example nanoscale device, an Indium- Arsenic nanowire. The horizontal axis may be defined as x and the vertical axis may be defined as y, the device is confined in two directions, and as such the planar structure would be repeated along the z axis, perpendicular to the xy plane. The nanowire of Fig. 2 alternates eight horizontal rows, each row comprising six Indium atoms with eight horizontal rows each row comprising six Arsenic atoms. Each Arsenic atom is bonded to four Indium atoms and each Indium atom is bonded to four Arsenic atoms. The extremes of the nanowire are bonded to Hydrogen atoms, thus defining the confinement of the nanowire. Returning to Fig. 1 step 106 is to perform a band structure calculation of the device. For this purpose, the Hamiltonian of the confined device, e.g. an Indium- Arsenic nanowire, may be constructed using either FPM or semi-empirical methods. The effective potential of the confined device is also calculated. A Hamiltonian describes the coupling between different electron orbitals at different atoms. The confined device, e.g. the nanowire of FIG. 2, may be described as periodic in at least one direction, i.e. the perpendicular direction to the confinement direction in Fig. 2. In the periodic direction the Bloch theorem is used to Fourier transform the Hamiltonian, such that the electrons would be described in momentum space instead of real space in this direction. We can then write the Bloch equation as follows:

^H kWnk = > ( ² ) where H _k is the k component of the Fourier transformed Hamiltonian, and ψ _Λ ,ε _Λ are the wave function and energy of band n, respectively.

The energy-momentum relation is called the band structure. The band structure of ten of the conduction bands of the geometry of FIG. 2 is shown in FIG. 3a. The conduction bands are the bands that are not occupied by an electron, and these bands are most relevant for electron transport. The horizontal axis is the electron momentum (k) and the vertical axis is the electron energy. Each line represents a different DFT band and circles, triangles and squares represent bands matched with sub-bands m=0, m=1 and m=2 respectively, of an effective mass ladder. Areas of the graph have been amplified in FIGS. 3b and 3c and several DFT bands have been labeled in the amplified figures according to their index numbers n. The DFT index numbers n are ordered according to increased energy. Generally the bands used for the matching may be the ones having an energy over a number of eVs, e.g. up to 1.2-1.5 eV, above the lowest energy for any momentum. In FIGS. 3a- 3c the n indexing is sorted according to the energy values at each momentum.

The plot of FIG. 3a shows the ten lowest energy conduction bands from the band structure of the confined device but does not identify the ladders resulting from the confinement. The identification of such ladders will be explained later on in relation with FIGS. 7a-7c.

The band structure in FIG. 3a is calculated using Density Functional Theory (DFT), but may also be calculated with any other suitable approach or combination of approaches, e.g. a semi-empirical tight binding method. The DFT method is a FPM approach, i.e. without adjustable parameters, while a semi-empirical model has a number of adjustable parameters. FIG. 3a also shows six points which are matched to an effective mass band.

Step 108 is to generate an effective mass model of the confined device. This model does not need the atomic-scale geometry, but a confinement potential, which defines how the electrons are confined inside the device. In other words, said effective mass model may not be generated by the parameters of the atomic-scale geometry but by a confinement potential together with a generic effective mass model. Said confinement potential can, for example, be obtained from the effective potential of the FPM calculation, by smoothing out the atomic-scale details. The smoothing can, for example, be done by averaging the potential at a given grid point with the potential in the surrounding points. Such smoothed confinement potential of the atomic-scale geometry of FIG. 2 is illustrated in FIG. 4.

FIG. 4 shows the value of the smoothed confinement potential for different values of the spatial coordinates x and y. The potential is given on a set of grid points and the value at a given grid point is given as a dot. The smoothed confinement potential together with a generic effective mass model defines an effective mass model Hamiltonian, H _k . The exact value of the parameters in the generic effective mass model are not so important, the parameters generated for a bulk system may be used. The effective mass Hamiltonian H _k may now be diagonalized to generate effective mass wave functions y7 _mi and energies s _mk .

Step 1 10 of FIG. 1 is to map each band of the FPM, e.g. DFT, band structure model with a specific band of the effective mass band structure model.

A single effective mass model may not generate all the bands in a DFT band structure, but only a subset of the bands called a ladder.. In the current example, conduction bands of the confined system (the bands illustrated in FIG. 3a) may be mapped with the bands in the effective mass model.

The procedure for performing step 1 10 is detailed in the method shown in FIG. 5. Step 501 shows the input for the matching, the energies and spatial wave functions from the DFT band structure and the effective mass band structure. Step 502 is to select the lowest energy DFT band not yet matched; for the first iteration this will simply be the lowest energy conduction band. Step 504 is to select the lowest energy effective mass wave function wchich in the first iteration will correspond to the lowest energy band of all the effective mass bands. This lowest energy effective mass band is matched with the lowest energy DFT band selected in step 502. The objective of the algorithm is to match each of the remaining effective mass bands with a DFT band. For this purpose, a projector function, p™° , is introduced in step 506. The aim of the projector function is to find the DFT wave function, ψ _nk , which has the same relative changes to y/ ^ as y7 _mi has to ψ _Μ , thus, the aim is to find the DFT wave function which best fulfills the following equation: Ψ nk _ Ψ mk /rs\ For this purpose a trial wave function, _ψ ^tr = -^-ψ _k , may be constructed. Said trial wave function will have the same local atomic behavior as ψ , but the global n ₀ k

behavior of the effective mass wave function _mk . Said trial wave function may be projected onto all the DFT wave functions, ψ _Λ , to obtain the projector function ^p nk ° = {^n mkn ₀ )■ The result of the projector function is a dot product. The projector function, reports each overlap in a matrix form. The overlap may be a number between 1 and 0 that specifies how much two wave functions resemble each other. For example, two identical wave functions overlap would be 1. The DFT wave function with maximum p™° value is the one that most closely matches the trial wave function, and thus best fulfills Eq. (3).

FIGS. 3d and 3e show two tables with overlaps of nine different wave functions (n) with three different trial functions (m). The overlap for n=0 wave function and m=0 trial function, is not calculated because they are automatically matched. Each table corresponds to a different electron momentum (k), that of FIG. 3d corresponds to k=0 and that of FIG. 3e to k=0.22. In both figures the highest overlap value has been highlighted by a circle. Said overlaps are accepted as a match when the largest value is larger than the second largest value by a user defined factor, e.g. 1.5 times.

FIG. 3d shows the matrix of the projection values at k=0, for 9 DFT bands and 3 trial functions. The DFT bands correspond to the bands shown in FIG. 3a. The highlighted value 0.46, which is larger than the other printed values, shows a match between m=2 and n=3 represented in Fig. 3a by a square. This means that the effective mass band which best fits the DFT band n=3 would be m=2. The other highlighted value, 0.40, identifies the match of n=6 into m=1 represented by a triangle in FIG. 3a. This means that the effective mass band m=1 is the one which best fits the DFT band n=6.

FIG. 3e shows the matrix of the projection values for 9 DFT bands and 3 trial functions at k=0.15. The overlap values correspond to k=0.22 1/Ang after converting normalized units into 1/Ang. The marked overlap values 0.33 and 0.43 correspond to the triangle and the square highlighted in FIG 3a. Said values identify two matches, the first one fits n=1 into m=1 and the second one fits n=2 into m=2.

Indeed, the values used to do the matching are values of projections around minimum energy value for all k. In the case an Indium-Arsenic nanowire the matching is done around k=0 because this is where the minimum energy occurs. However, in case of other materials, e.g. Silicon, the matching may not be around k=0 but around a different value of k. The matching is done around where the band minimum occurs because the fitting e.g. a second polynomial function, will only be accurate in those areas around the band minimum energy. The values of FIG. 3e are within the k range limitations for an accurate effective mass band matching for this example.

The maximum projection values identified for both k values are highlighted in FIG. 3a, e.g. when k=0 (FIG. 3d) the best fit for n=3 is given by ™° = 0.46 and therefore m=2 is matched to n=3. The circles identify the overlap for effective mass band m=0, triangles identify the overlap for band m=1 and squares identify the overlap for band m=2.

In either cases, none of the DFT wave function are matched to the m=3 trial wave function, which means that m=3 wave function does not fit with any of the 10 considered DFT bands. Both tables illustrate how the matching number (n) of a trial function (m) may change for different k when the bands are interleaved in the band structure. In both cases the n ₀ =0 which means that these matched points are the same sub-band(s) of the same ladder at different electron momentum (k).

The process may be repeated for various values of k around the band minimum and the result may be the matching of all effective mass ladders into different DFT bands.

FIG. 6 shows four DFT wave functions 602, 604, 606, 608 and the four lowest energy effective mass wave functions 610, 612, 614, 616 in real space. The value of a wave function at each grid point is illustrated in FIG. 6 by a dot; for higher function values, larger dots are assigned. The sign of the wave functions is illustrated in FIG. 6 by the contrast of the dots; darker dots correspond to a positive sign while lighter dots correspond to a negative sign. The DFT wave functions 602, 604, 606, 608 are selected by the projector function to best match the effective mass wave functions 610, 612, 614 and 616. As shown in FIG. 6, 602 matches 610, 604 matches 612, 606 matches 614, and 608 matches 616. It is clear from the figure that the global behavior of the matched wave functions is similar.

Returning to FIG. 5, if all the DFT bands in the considered low energy range, i.e. over a number of eVs, e.g. 1 .2 eV, above the lowest energy, are not matched by an effective mass band, then the procedure is repeated in step 510. In the next iteration, at step 502, the procedure selects the lowest energy DFT band in the list of bands that is not matched. This unmatched lowest energy DFT band will be matched with the lowest level of another effective mass ladder and will be used to set up new trial wave functions. These new trial wave functions will be used to match remaining DFT bands. The procedure is continued until all considered DFT bands are matched. In this way, each effective mass ladder matches to a series of DFT bands.

FIGS. 7a-7c shows how the DFT bands from FIG. 3a have been matched with three different effective mass ladders. In other words, DFT bands of FIG. 3a have been split into three different series using the method described with reference to the method of FIG. 5 and each series corresponds to one effective mass ladder. When FIGS. 7a-7c are combined the result corresponds to all the bands shown in FIG. 3a.

FIG. 7a shows a ladder formed by the DFT bands n=0, n=3, n=6 and n=7. The FIG. 7b shows a ladder formed by the DFT bands n=1 , n=4 and n=8 wherein the lowest band of the ladder (n=1 ) correspond to the first unmatched band, that is, the next iteration of the method of FIG. 5. FIG. 7c shows a ladder formed by the DFT bands n=2, n=5 and n=9, wherein the lowest band (n=2) of the ladder corresponds to the second unmatched band and thus, the third iteration of the method of FIG. 5.

In step 1 12 of FIG. 1 the parameters e.g. effective mass and non-parabolicity parameter, in each effective mass model are readjusted to obtain the best possible fit of the effective mass band energies to the DFT band energies. This is illustrated, e.g. in FIG. 7a, wherein the first effective mass model equation is fitted to the DFT bands, e.g. by simple curve fitting, and the result is shown with a dashed line.

FIG. 8 illustrates the combined fit of the three effective mass ladders of FIGS. 7a-7c to the DFT band structure of FIG. 3a.

The fitted effective mass models are the output of the algorithm and the objective of the present disclosure. The fitted effective mass models may be then used in material or device simulations, e.g. TCAD simulations. In the example of the disclosure, a DFT calculation has been used for the reference band structure calculation; however, the calculation could also be done with a semi-empirical or another first principles method. The semi-empirical methods could, for instance, be a tight-binding method, empirical pseudo potential method, or any other suitable method that can provide eigen energies and wave functions for an atomic-scale geometry.

The material system of the example is, a 1 -dimensional nanowire of Indium- Arsenic. Nevertheless, the method works for all systems that can be described by an atomic-scale geometry, for instance a 2-dimensional, 3-dimensional, or any other suitable N-dimensional nanoscale system of any suitable material or combination of materials. The material system could also be an alloy of different elements where the elements are described using the virtual crystal approximation. In the virtual crystal approximation, an alloy between two elements is described by a new "virtual" element that has the average property of the two alloy elements.

For the approximate band structure model of the example an effective mass model has been used; however, this could also be another approximate model. The separation of the device bands into different series corresponding to different ladders allows for an automatized and less complex procedure to extract effective mass models from atomic-scale modeling that may be used in TCAD tools to simulate band structure data. Different models may now be used to approximate each series. One such example is an effective mass model with a non-parabolicity parameter and different effective masses in the different directions. In such a model: /?!^ 2m _y 2m _z where U _n and a are the fitting parameters, (k _x0 , k _y0 , k _z0 ) is the wave vector of the band minimum, and (m ^* , m ^* , m ^* ) are the anisotropic effective masses.

FIG. 9 shows an exemplary hardware implementation of the flowcharts in FIGS. 1 and 5. The atomic-scale geometry parameters may reside in a memory 902 and different processing units can access the data. There may be a specific processing unit 904 directed to generate or retrieve a DFT model. From said model the DFT band structure can be generated. Another processing unit 906 may generate an effective mass band structure. A third processing unit 908 may combine the DFT and effective mass wave functions to generate projector functions. Although these processing units are described as three separate units, they can be combined into one or two units, or further divided into additional processing units. Any other suitable processing units or combination of processing units can be used to access the atomic-scale geometry.

Although only a number of examples have been disclosed herein, other alternatives, modifications, uses and/or equivalents thereof are possible. Furthermore, all possible combinations of the described examples are also covered. Thus, the scope of the present disclosure should not be limited by particular examples, but should be determined only by a fair reading of the claims that follow.

Further, although the examples described with reference to the drawings comprise computing apparatus/systems and processes performed in computing apparatus/systems, the invention also extends to computer programs, particularly computer programs on or in a carrier, adapted for putting the system into practice.