Created by Margret J. Geselbracht, Reed College (mgeselbr@reed.edu) and posted on VIPEr
(www.ionicviper.org) on March 19, 2011. Copyright Margret J. Geselbracht 2011. This work is licensed
under the Creative Commons Attribution Non-commercial Share Alike License. To view a copy of this
license visit http://creativecommons.org/about/license/.
Raman Spectroscopy of P-doped Sodium Silicide, Na4Si4
1. Susan Kauzlarich’s group at UC Davis has pioneered the use of sodium silicide, Na4Si4, as a
precursor to silicon nanoparticles. This so-called Zintl phase is composed of isolated
tetrahedral Si44– clusters. To explore potential precursors to P-doped silicon nanoparticles,
Kauzlarich and co-workers recently synthesized Na4Si4 doped with 1-3 atomic % phosphorus
in which the phosphorus atoms replaced silicon atoms in the tetrahedral clusters.
(a) Assume that the Si44– clusters have perfect Td symmetry. Conduct a complete normal
mode analysis and predict the number and symmetries of the Raman-active modes.
z1
y1
x1
The simplest orientation is to place all the atom
displacement vectors in the alignment shown on the
left (the unlabeled vectors would be x4, y4, and z4.
z2
z3
y2
x2
x3
y3
all motion =
E
8C3
3C2
6S4
6d
12
0
0
0
2
all motion = A1 + E + T1 + 2 T2
–  xyz = T2
–  Rx, Ry, Rz = T1
 vibrational = A1 + E + T2
Note that this matches 3(4) – 6 = 6 normal modes expected.
All of these modes should be Raman active, so we would expect to see 3 peaks or 3
unique frequencies in the Raman spectrum. One of these modes should have A1
symmetry, 1 should be a doubly degenerate set of modes with E symmetry, and 1 should
be a triply degenerate set of modes with T2 symmetry.
(b) The crystal structure of Na4Si4 shows that the tetrahedral clusters are actually distorted
with four edge lengths that are all unique. Use this information to help interpret the
observed Raman spectrum of the undoped Na4Si4 shown below.
Figure removed for copyright reasons.
See Figure 4 in:
Created by Margret J. Geselbracht, Reed College (mgeselbr@reed.edu) and posted on VIPEr
(www.ionicviper.org) on March 19, 2011. Copyright Margret J. Geselbracht 2011. This work is licensed
under the Creative Commons Attribution Non-commercial Share Alike License. To view a copy of this
license visit http://creativecommons.org/about/license/.
Jialing Wang, Sabyasachi Sen, Ping Yu, Nigel D. Browning, and Susan M. Kauzlarich,
“Synthesis and spectroscopic characterization of P-doped Na4Si4”
J. Solid State Chem. 2010, 183, 2522-2527.
This distortion basically removes all symmetry from the molecule leaving us with the C1
point group. There is only one representation in C1, which means that all 6 normal modes
should all be Raman active, and we would expect to see 6 peaks. However, if the
distortion is not too great, the frequencies of these normal modes might group together in
patterns reminiscent of the parent Td symmetry. In looking at Figure 4, we see a single
peak near 485 cm-1 that we can assign to the mode with A1 symmetry (in the ideal Td
molecule). There are 3 peaks between 340 and 378 cm-1, and we might expect that these
three modes are derived from the T2 set in the ideal Td symmetry. Finally, there are two
peaks at lower frequencies, ~280 and 305 cm-1, and these are the two modes derived from
the E set in the ideal Td symmetry.
(c) Suggest an explanation for the changes in the spectra with P-doping.
Your first thought should be that doping with phosphorus would lower the symmetry of
the molecule. For example, with a P atom substituting for one Si atom, this would lower
the idealized symmetry from Td to C3v. With 2 P atoms, the in the tetrahedron, the
molecule would drop to C2v, although substituting two atoms in the same tetrahedron
would be fairly unlikely at these low doping levels.
However, the distortions in the Si4 tetrahedra have already lowered the symmetry as low
as we can go, so doping with phosphorus will not change the number of peaks in the
Raman spectrum. Doping with P will change the reduced mass (viewing the vibrations as
harmonic oscillators, or balls on a spring) however as P is a slightly heavier atom than Si.
Since the vibrational frequency is inversely proportional to the reduced mass, this should
shift the normal modes to slightly lower frequencies with P-doping. The only mode for
which we have the data for all 3 levels of P-doping is the “A1” mode, which shifts from
486.4 cm-1 to 484 cm-1 with P-doping levels up to 3 atomic %. Granted this is not much,
but most of the molecules remember will be undoped.