3

6.730 PHYSICS FOR SOLID STATE APPLICATIONS

Department of Ele tr! al En"!neer!n" an# Comp$ter S !en e %a&&a '$&ett& In&t!t$te of Te 'nolo"(

)REAL* SE%ICOND+CTOR PRO,ECT

Intro#$ t!on This is an extensive project intended to introduce you to something approximating the real world of solid state calculations. Because this is a real problem, you are encouraged to collaborate with classmates and work in teams if you so desire. owever, everyone must turn in their own project. !ach team will be assigned a different semiconductor material to use throughout the project. The project consists of the following three parts" #irst, you will be asked to describe fully a material$s crystal structure, and to determine x%ray diffraction patterns for it. The second section of the project involves development of a model for the material$s vibrational properties &phonon dispersion relationships, elastic constants, specific heat, etc.'. (n the final part of the project, you will use an )*+, model to calculate the material$s electronic band structure &energy gap, whether the material is direct or indirect, effective densities of states for the valence and conduction bands, etc.'. Throughout the project you are expected to check how good-bad your calculations are on the basis of real experimental and theoretical results. (n general, your calculations should yield reasonable &within ./%0/1' answers, so these experimental checks are a good way to ensure that you$re on the right track. This project does involve both library research and significant matlab coding. 2uch of the information you need can be found in the books that are on reserve. +ll literature sources you use must be documented appropriately. +ll matlab code must be sent by email, as well as included, with comments, as an appendix to the project.

Part I- Cr(&tal Str$ t$re

This part of the project is relatively simple, but it does lay important ground work for the second and third parts of the project. 4hen you do the research for this section, be sure to make copies of any information you find on your material%%it will come in handy for parts (( and (((. 5o appropriate research to answer the following 6uestions about your material"

3. 4hat is the crystal structure7 .. 4hat is the lattice constant7 0. 4hat is the basis7 8. 4hat are the primitive lattice vectors7 9. 4hat is the structure of the reciprocal lattice7 :. 4hat are the primitive reciprocal lattice vectors7 ;. 4hat are the atomic form factors for your material7
The following can be done by hand, or by using matlab where appropriate"

<. =rovide pictures of the crystal and of the reciprocal lattice in the >3//?, >33/?, and >333? planes.
(ndicate the vertical positions of atoms with respect to the plane &see, for example, @ittel p. 3A'.

A. *alculate the structure factor for all reciprocal lattice vectors . of magnitude B.B. C 3:&.-a'.. Be sure
to include the atomic form factors.

3/. 5escribe in detail an experimental setup you could use to determine the crystal structure of your
material by x%ray diffraction. =rovide appropriate figures. Dpecify what wavelength&s' of x%ray is&are' necessary in your setup. Duggest possible sources.

33. *alculate and draw the expected x%ray diffraction pattern&s' from your setup. 5iscuss &with
calculations' how the pattern&s' would be used to determine the crystal structure.

3.. Try to find experimental results for x%ray diffraction on your material. *ompare with your predicted
results and comment on any discrepancies you find.

0
6.730 PHYSICS FOR SOLID STATE APPLICATIONS

Department of Ele tr! al En"!neer!n" an# Comp$ter S !en e %a&&a '$&ett& In&t!t$te of Te 'nolo"(
)REAL* SE%ICOND+CTOR PRO,ECT

Part II- P'onon& (n this part of the project, you will be developing a simple model for the phonon spectra of your semiconductor. The game plan here is to develop a general model for the phonon dispersion relationship, in terms of some unknown force constants, and then to use experimental data to fit the force constants to the material. The model you will be using is a Born force constant model that incorporates bond bending and bond stretching, as we discussed in class. A. /a 0"ro$n# 1$e&t!on& #or the Born force constant model, we can consider just nearest neighbor interactions, or we can make the model more accurate &and more complicated' by including next nearest neighbors, next%next nearest neighbors, etc. Before beginning on your calculation, answer the following general 6uestions about the Born model"

30. ow many force constants are re6uired for each bond7 4hy7 38. 4hat is the energy of a single bond in the Born model7 39. The model assumes that the bond is only slightly displaced from e6uilibrium.
ow would you modify the model to make the bond energy more realistically dependent on displacement from e6uilibrium%% what order would the corrections be, and of what sign7 Eustify your answer physicallyF include sketches if appropriate. material7 ow large will the dynamical matrix be7 4hat if you use nearest neighbor and next nearest neighbor couplings7

3:. (f you use only nearest%neighbor couplings, how many force constants will your model re6uire for your

3;. ow many independent elastic constants does your material possess7 4hat are they &give numbers'7
4hy will a nearest%neighbor approach not provide the most general solution for a cubic material7 /. Con&tr$ t!on of D(nam! al %atr!2 (n fact, a nearest neighbor calculation that accounts for bond bending actually does a good job &within ./1', and is significantly more tractable than the next nearest neighbors approach. 4e will therefore start out with a nearest neighbors approachF next nearest neighbors will be left as an extra credit bonanGa for the diligent among you.

3<. 5raw all the atoms in the basis and all their nearest neighbors. )abel the atoms according to their
lattice vector and basis vector.

3A. 4rite a general expression for the potential energy of all the atoms in the crystal in terms of their
displacement from e6uilibrium. Hse a Born force model which considers only nearest neighbor interactions to simplify this expression &your expression should contain force constants and displacements at this point'.

./. Hse your simplified expression for the potential energy to determine the force on a given atom in the
crystal in terms of its displacement and its neighbors$ displacements. *heck your answer by directly calculating the force from the spring constants and displacements.

.3. Hse your simplified expression for the potential energy to calculate the dynamical matrix. (s it
ermitian7

... 4rite a matlab code to find the phonon spectra for general values of your force constants and atomic

masses. =lot the phonon dispersion in appropriate units along the %I, I%), and %) directions using force constants of B!J5K3, K/..9, and 23K2.K3. , and for the highest optic and lowest acoustic modes at I and ). =rovide drawings of the atomic motion of these modes. ow many modes are there at 7

.0. #or these values of force constants and masses, determine the atomic displacements for all the modes at

C. %o#el Opt!m!3at!on an# Compar!&on to %a ro& op! Propert!e& )ook up experimental results for phonon dispersion in your material. Hsing whatever criteria you deem appropriate &sound velocities, elastic constants, Gone edge fre6uencies, etc.', determine optimum values of your force constants to match measured results. *alculate sound velocities, Gone edge fre6uencies, and elastic constants &c33, c3., and c88' using your model and compare with values from the literature. Dhow all your calculations. Hsing your model, do the following"

.8. =lot a comparison of your calculated phonon dispersion with theoretical and-or experimental results
along the %I, I%), and %) directions.

.9. =lot the total density of states &histogram method, include all modes' versus fre6uency. .:. *alculate the specific heat of your material versus temperature using &a' your calculated density of
states, &b' a 5ebye model, and &c' a combined 5ebye%!instein model &5ebye for acoustic modes, !instein for optical modes'. =lot your results for temperatures between / @ and 0 5. *omment on the strengths and weaknesses of your model. (f you so desire, extend it to next nearest neighbors &this is not re6uired, but if you$re bored one Daturday night...'

9
6.730 PHYSICS FOR SOLID STATE APPLICATIONS

Department of Ele tr! al En"!neer!n" an# Comp$ter S !en e %a&&a '$&ett& In&t!t$te of Te 'nolo"(
)REAL* SE%ICOND+CTOR PRO,ECT

Part III- Ele tron! /an# Str$ t$re (n this part of the project, you will be developing an )*+, picture for the valence and conduction bands of your semiconductor. +s in the phonon part of the project, you will first develop a general form for the solution. Lou will then apply that form to your material, compare your calculated results with results reported in the literature, and use your results to evaluate physically interesting material properties. A. /a 0"ro$n# 1$e&t!on& % Before developing the amiltonian matrix, answer the following general 6uestions about the )*+, method"

3. #or our )*+, model we construct extended orbital basis functions, | 405> , which are in turn used to

construct our trial wave function, | 405> . Typically use only the outermost &highest principal 6uantum number' orbitals to construct our different | 405> , so that in a system with b atoms in the basis and n outermost orbitals per atom, the total number of extended orbital basis functions is nb. ow many extended orbital basis functions will you have for your material7 4hy7 so, how many extended orbital basis functions would you have to use for your material7 ow large would your amiltonian matrix be in this case7 ow do you expect your results would differ from those you would get with just valence orbitals7 4hat if we used higher &totally unoccupied' orbitals, too7 ow many orbitals per atom would we have to use to get an exact band structure7 4hy7

3. *onceivably one could also construct | 405> out of core orbitals as well as valence orbitals. (f you did

/. Con&tr$ t!on of Ham!lton!an %atr!2

3. 4hat are the atomic configurations of the two atoms in your material7 4hich orbitals do you expect to
play a significant role in bonding7

.. 5raw all the atoms in the basis and all their nearest neighbors with appropriate orbitals on each atom.
)abel the orbitals according to their lattice vector, basis vector, orbital type, and the type atom they are associated with &cation or anion'. #or example, the s orbital on the cation at lattice vector R, basis vector would be labeled | & 4R6 5>

0. Dhow that one can approximate all the nearest neighbor interactions with one of the following &to within
a sign'. =ictorial arguments &with accompanying explanation' are acceptable. E&&=< & 405| H| &a4a789a789a785> E&p=< & 405| H| p2a4a789a789a785> E22=< p2 405| H| p2a4a789a789a785> E2(=< p2 405| H| p(a4a789a789a785> (s this approximation reasonable7 *omment on the validity of < & 405| H| p2a4a789a789a785> = < &a405| H| p2 4a789a789a785> Miven this approximation, can we write < & 405| H| & 405> = < &a4a789a789a785| H| &a4a789a789a785> 7 *alculate values for !ss, !sp, !xx, and !xy in terms of Nss, Nsp, etc. &see =D :'.

8. *onstruct a set of extended atomic orbitals. Be explicit about your choice of phase factors. 4rite your
trial wave function as a linear combination of these extended orbitals.

: 8. #ind the amiltonian matrix for the nearest neighbor approximation. Lour answer should contain only
!ss, !sp, !xx, !xy, the s and p energies of the anion and cation &!sa, !sc, !pa, and !pc' and phase factors. J,T!" Lou will be able to find this matrix in the literature. owever, several of the sources &including arrison, 3A</ and Dlater and @oster, 3A98' have minor errors in their matrices or matrix elements, so check your answer carefully. (f you are really motivated, find for next%nearest neighbors, too... C. /an# Cal $lat!on&

3. )ook up a real energy band diagram for your material. (nclude a copy in your report. .. 4rite a matlab program to plot the free%electron band structure for your material along the same

directions as used in the energy band you found in the literature >Jote" neglecting factors of -a, the symmetry points are K&/,/,/'F IK&3,/,/'F ) K &3-., 3-., 3-.'F @K&0-8, 0-8, /'F and 4K&3, 3-., /'?. (ndicate the degeneracies of the different bands. 4here is the #ermi level located7 ow does the free electron band structure compare to the real band structure7 eigenvectors, and what do they correspond to physically &hint" look back at =D 3'7 Hsing arrison$s Dolid Dtate Table &attached' find numerical values for ! sa, !sc, !pa, !pc, !ss, !sp, !xx, and !xy. *ompare your calculated energies at the point with values from the literature.

0. +lgebraically diagonaliGe the amiltonian matrix at the point. 4hat are the different energies and

8. 4rite a matlab program to plot the )*+, energy bands along the same directions as above, along with
the approximate location of the #ermi level. ow do your results compare &6ualitatively' with the band structure you found in the literature7 (f you wish, optimiGe the matrix elements for your material.

9. 4here are the valence band maximum and the conduction band minimum located7 4hat is the energy
gap7 (s your material direct or indirect7 D. Effe t!:e %a&&e&9 Con&tant Ener"( S$rfa e&9 Den&!t( of State& % Dince you are working in 05, it will probably prove rather challenging to plot the 05 constant energy surfaces near the valence band and conduction band edges. (f you can figure out a good way to do this &that doesn$t involve making an effective mass approximation' you will be rewarded with many bonus points and a Toscanini$s gift certificate. #ailing that,

3. =lot constant energy contours near the valence and conduction band edges for appropriate planes.

#or example, if your minimum is at 0m!n along %I, you should probably plot an energy contour for the kx%ky plane, and for the plane parallel to ky%kG that contains 0m!n. #or a minimum along %), the plane containing the %) direction and the plane perpendicular to that direction would make sense. points using a 6uadratic polynomial &be sure to think about your results from 53 when you do this'. Hse your results to find the effective masses for both the valence bands and the conduction band. ow do your results compare with results from the literature7 ow would you improve your results7

.. Dolve for the energy at a number of k points near the valence and conduction band edges. #it these

0. =lot the total density of states &histogram method, include all bands' versus energy. Hse your calculated
effective masses to determine an approximate expression for the density of states near the valence and conduction band edges. ow does this calculation compare with the total 5.,.D.7

8. Hsing your total 5.,.D., calculate the electronic specific heat of your material as a function of
temperature. *ompare this with your calculations for the phonons, and comment.

9. Hse the band structure you found in the literature to discuss the characteristics of your material. 4hat
electronic-optical applications would your material be good-bad for7 4hy7

6.730 PHYSICS FOR SOLID STATE APPLICATIONS Department of Ele tr! al En"!neer!n" an# Comp$ter S !en e %a&&a '$&ett& In&t!t$te of Te 'nolo"(

)REAL* SE%ICOND+CTOR PRO,ECT

Lour assigned material is" 0*%Di* +l+s Ma=%( Ma+s (n= (n+s MaDb%(