Dismiss Notice
Join Physics Forums Today!
The friendliest, high quality science and math community on the planet! Everyone who loves science is here!

A question about band structure silicene

  1. Jun 28, 2015 #1
    Dear forum people
    I consider four atomic orbitals of the silicon atom: s, px , py and pz , thus Hi,j is a 4*4 matrix.
    Hi,i is the expression of the diagonal matrix and Hi,j is the expression off-diagonal matrix where the expression for Hi,i and Hi,j is attached.
    But I can not calculate the cosine of direction.
    Who can help me find the cosine of direction?
     

    Attached Files:

  2. jcsd
  3. Jun 28, 2015 #2

    DrDu

    User Avatar
    Science Advisor

    Could you attach a pdf?
     
  4. Jun 28, 2015 #3
     

    Attached Files:

  5. Jun 29, 2015 #4

    DrDu

    User Avatar
    Science Advisor

    So what is your problem? l, m, n are just the cartesian x, y and z coordinates of the distance vector of the two atoms.
     
  6. Jun 29, 2015 #5
    Position vectors of the 1 NN for silicene as follows:
    R1=(a/√3,0,0) , R2=(-a/2√3,a/2,0) , R3=(-a/2√3,-a/2,0)
    but n, l,m for the nearest neighbours is three values as we must have a value for the nearest neighbours.
     
  7. Jun 29, 2015 #6

    DrDu

    User Avatar
    Science Advisor

    Well, for example you take ##r_i= R0=(0,0,0)^T## which has the three nearest neighbours you specified. So e.g. for ##r_j=R2##:
    ##l=R1_x-R0_x=a/\sqrt{3}##.
     
  8. Jun 30, 2015 #7
    Dear DrDu
    Thank you very much.
     
  9. Jul 2, 2015 #8
    If you're trying to plot this in something like MATLAB you can form the two block matrices and run a loop that adds the phase factors for the hopping parameters and then plot over those points if you know the region it spans (which in your case is just the Brillouin zone)
     
  10. Jul 3, 2015 #9
    I can't calculate phase factors for hopping parameter. can help me?
     
  11. Jul 3, 2015 #10

    DrDu

    User Avatar
    Science Advisor

    It is just the product of the phase factors of the two orbitals involved.
     
  12. Jul 3, 2015 #11
    In graphene
    <si|H|sj>=t1*g0 and <si|H|p(x)j>=t2*g1
    where
    g0= 1+exp(ik.R(B))+exp(ik.R(C)) and g1= 1-(1/2)(exp(ik.R(B))+exp(ik.R(C)))
    How to calculate g0 and g1?
     
  13. Jul 3, 2015 #12

    DrDu

    User Avatar
    Science Advisor

    g0 and g1 are the sum of three hopping parameters. In g0, the first term is the hopping between the atoms in the same cell. So the phase factor for both orbitals is 1 and the contribution is t1*1*1. The other two hopping terms go to orbitals with phase factors exp(ik.R(B)) and exp(ik.R(C), respectively. So the second factor is t1*1*exp(ik.R(B)) and analogously for C.
    For the p_x orbitals, g also contains the direction cosine, which is 1 for A and cos(+-120 deg) =-1/2 for B and C.
     
  14. Jul 3, 2015 #13
    I'll post some snippets of what code in MATLAB would look like for a calculation like this.

    phase(i) = exp(1i*dot(k(:,index),R(:,i)))

    This gives you the phase you're going to multiply with the hopping parameter. You'd have to write a loop that runs from i = 1:3 in order to multiply each contribution to the hopping parameter such that the product of all 3 would get you the total hopping parameter. k and R here are matrices that contain generated wavevectors and the coordinates for the atoms, respectively.


    Code (Text):
    Es = -4.2;
    Ep = 1.715;
    orbitals = [Es Ep Ep Ep];
    Vpp_sigma = 2.72;
    Vpp_pii = -0.72;
    Vss_sigma = -2.08;
    Vsp_sigma = 2.48;

    a = 3.86;
    a1 = (a/2) * [sqrt(3), -1, 0];
    a2 = (a/2) * [sqrt(3), 1, 0];

    coordinates = [4.45714 0 0; %B
                   2.22857 0 0.46152]; %A

    R1 = coordinates(2, :)-coordinates(1,:);
    R2 = R1 + a1;
    R3 = R1 + a2;
    R = [R1' R2' R3'];

    n = dot(R,[ 0 0 1])/norm(R); %direction cosine
    l = dot(R,[1 0 0])/norm(R);
    m = dot(R, [0 1 0])/norm(R);
    tss = Vss_sigma;
    txs = l*Vss_sigma;
    tys = m*Vss_sigma;
    tzs = n*Vss_sigma;
    txy = l*m*Vpp_sigma - l*m*Vpp_pii;
    txz = l*n*Vpp_sigma - l*n*Vpp_pii;
    tyz = m*n*Vpp_sigma - m*n*Vpp_pii;
    tzz = n^2 *Vpp_sigma+(1-n^2)*Vpp_pii;
    txx = l^2 *Vpp_sigma+(1-l^2)*Vpp_pii;
    tyy = m^2 *Vpp_sigma+(1-m^2)*Vpp_pii;
     
    These will be your matrix elements that you'll want. Since the Hamiltonian matrix takes a block-diagonal form, you can create an A block and B block and set the Hamiltonian equal to [A B; B A]. Once you run the loop to get all the hopping parameters correct you can simply plot the eigenvalues of the full Hamiltonian matrix across the desired path to obtain the full band structure.
     
  15. Jul 5, 2015 #14
    In the l,m,n inner product matrix and vector is not possible so n,l,m how obtained?
     
  16. Jul 6, 2015 #15
    Hi
    In matlab program, cosine directions (n,l,m) not true because inner product R and vectors is impossible.
    I can't construction matrix Hamiltonian because txy txz , etc are 1*3 vectors while tss is 1*1 vector.
    can help me?
     
Know someone interested in this topic? Share this thread via Reddit, Google+, Twitter, or Facebook




Similar Discussions: A question about band structure silicene
  1. Band structure (Replies: 5)

Loading...