See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/303364491

Errors in The Feynman Lectures on Physics: Symmetry and Crystals

Article · May 2016

CITATIONS READS

0 129

1 author:

Rajesh Prasad
Indian Institute of Technology Delhi
30 PUBLICATIONS   225 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Helicoids associated with dislocations View project

Quasicrystals View project

All content following this page was uploaded by Rajesh Prasad on 26 May 2016.

The user has requested enhancement of the downloaded file.

 GENERAL  ARTICLE

Errors in The Feynman Lectures on Physics

Symmetry and Crystals

Rajesh Prasad

We discuss some errors in The Feynman Lectures on Physics

related to the concept of symmetry. We also suggest a possible
correction to Fig.1.4 of Vol. I. The discussion may be useful
for students of crystallography, solid state physics and solid
state chemistry.
Rajesh Prasad teaches
1. Introduction: materials science at the
Indian Institute of
The Feynman Lectures on Physics [1] are justifiably famous as an Technology Delhi. He is
also interested in science
interesting and inspiring introduction to physics at the under-
education at the school
graduate level. Here I point out some inadvertent mistakes related level.
to symmetry and crystallography that have crept into the text. A
discussion of these errors can be illuminating for beginning
students of physics, chemistry and materials science.

2. Basic Crystallography

To put the errors in a proper context, we discuss briefly elemen-

tary crystallography. A crystal is a 3D periodic arrangement of
atoms. A lattice is a 3D periodic arrangement of points. Thus, a
crystal can be considered as a lattice with an atom or group of
atoms associated with each lattice point. The atom or group of
atoms associated with each lattice point is called a motif or basis
of a crystal. This relationship between a crystal and the associated
lattice and motif is sometimes expressed as

Crystal = Lattice + Motif.

Keywords
Or, in other words, the lattice tells us how to repeat but the motif Symmetry, 3-fold axis, 6-fold
gives what to repeat. A lattice is a geometrical scaffolding to axis, hexagonal symmetry, 2-
which atoms are attached to get the physical crystal. fold axis, mirror plane, bilateral
symmetry, lattice, crystal, motif,
The fact that crystals are 3D periodic arrangements of atoms can body-centred cubic lattice, face-
centred cubic lattice.
also be expressed by stating that they have 3D translational

RESONANCE May 2016 429

 GENERAL ARTICLE

The crystallographic symmetry. Thus translational symmetry is the defining require-

restriction theorem ment of a crystal. We can define translational symmetry as
forbids the presence follows. If on shifting all points of a crystal by a vector r, we
of a five-fold axis in a arrive at points that are indistinguishable from the starting points,
translationally then we say that the crystal possesses translational symmetry.
periodic crystal. But
it can be present in As already stated, all crystals, by definition possess translational
quasicrystals, which symmetry. However, what makes the study of crystals interesting
lack translational is that they may (and often do) also possess another kind of
periodicity. The 2011 symmetry called point symmetry. In general, a symmetry opera-
Nobel Prize in tion is one that brings an object into self-coincidence. Transla-
Chemistry was tional symmetry discussed above is one example. By definition, a
awarded for the point symmetry is an operation that brings the object into self-
discovery of coincidence, and also leaves a set of points unmoved or invariant.
quasicrystals. As translation operation moves all points it is not a point symme-
try. Two examples of point symmetry are rotation and reflection.

If a crystal is brought into self-coincidence by a rotation about an

axis it is said to possess rotational symmetry. This is obviously a
point symmetry operation as points on the rotation axis remain
unmoved. If T is the minimum angle of rotation about an axis
which brings a crystal into self-coincidence, then in one complete
rotation it will come into self-coincidence n times where n =
360°/T. The crystal is then said to possess an n-fold axis. An
equilateral triangle has a 3-fold axis passing through its centre,
and a square has a 4-fold axis.

An interesting theorem, known as the crystallographic restriction

theorem, restricts the values of n to 1, 2, 3, 4 and 6. This is because
other rotational symmetries are incompatible with the transla-
tional symmetry of a crystal. In particular, the absence of 5-fold
symmetry for periodic crystals should be noted. In 1984, Schetman
et al reported a solid phase of metallic alloy having 5-fold
symmetry. However, the phase, although ordered, lacked transla-
tional symmetry in accordance with the above stated theorem and
was called quasicrystal. Schetman was awarded the 2011 Nobel
Prize in chemistry for this achievement [2].

430 RESONANCE May 2016

 GENERAL  ARTICLE

Another point symmetry is the reflection symmetry. If a crystal is Any 2D periodic

brought into self-coincidence by reflection in a plane (called pattern with 6-fold
mirror plane) then it is said to possess reflection symmetry. This symmetry axes will
is obviously a point symmetry as points on the mirror plane also possess 3-
remain unmoved. This symmetry is also known as bilateral fold axes.
symmetry.

We are now prepared to discuss the errors in The Feynman

Lectures on Physics (henceforth FLP).

3. ERROR 1: Hexagonal vs. Trigonal Symmetry

Following is a quote from Chapter 1, Volume 1 of FLP (Fig.1-4

referred to in the text is reproduced as Figure 1 here.)

“Fig.1-4 is an invented arrangement for ice, and although it

contains many of the correct features of the ice, it is not the
true arrangement. One of the correct features is that there is
a part of the symmetry that is hexagonal. You can see that
if we turn the picture around an axis by 120°, the picture
returns to itself.”

Error in text: Hexagonal symmetry should involve a 6-fold

rotation axis with minimum angle of rotation of 360/6 = 60° and
not 120°.

Error in figure: The large dark circles represent oxygen atoms

Figure 1. A reproduction of
Fig.1-4 of FLP, Vol.I.

RESONANCE May 2016 431

 GENERAL ARTICLE

Figure 2. Fig.1-4 of FLP with

the 3-fold and a 6-fold axes
marked respectively by tri-
angles and hexagons. The
figure is centred on a 3-fold
axis which is a wrong choice
if one wishes to emphasize
the 6-fold symmetry of the
structure.

and the smaller white circles hydrogen in Figure 1 (Fig.1-4 of

FLP). As claimed by Feynman, this is not the true structure of ice
but a 2D representation attempting to show the hexagonal symmetry.
But where is the hexagonal symmetry axis in the figure? The
centre of the ring of atoms in the middle of the figure is a 3-fold
Figure 3. The suggested cor- axis and not a 6-fold one. This is shown in Figure 2 with a small
rection for Fig.1-4 of FLP. The triangle (standard symbol for such an axis). This is because the
structure shown in Figures 1 three alternating sides of the hexagonal ring each accommodate a
and 2 has been extended
pair of H, but the other three sides do not.
periodically to include more
atoms. The box is centred on The question now is how to correct this figure without drastically
a 6-fold axis.
changing the original. If we look carefully, there are three six-fold
axes marked by little hexagons in Figure 2. But
all of these are on the sides of the figure and not
in the centre. Thus the mistake in Fig.1-4 of
FLP is that the external box of the figure is
incorrectly centred on a 3-fold axis rather than
a 6-fold axis. We note that any 2D periodic
pattern having 6-fold axes will necessarily have
also 3-fold axes. Thus the error in the figure is
that although the intention is to show the 6-fold
(hexagonal) symmetry, the box is centred on
one of the 3-fold axes. With this realization the
correction is simple. All that one has to do is to
shift the box and centre it on a 6-fold axis. This
is shown in Figure 3.

432 RESONANCE May 2016

 GENERAL  ARTICLE

4. ERROR 2: Bilateral Symmetry vs. Two-Fold Axis

The following quote is from FLP, Vol 1, Section 11-1:

“The word “symmetry” is used here with a special mean-

ing, and therefore needs to be defined. When a thing is
symmetrical-how can we define it? When we have a pic-
ture that is symmetrical, one side is somehow the same as
the other side. Professor Hermann Weyl has given this
definition of symmetry: a thing is symmetrical if one can
subject it to certain operation and it appears exactly the
same after the operation. For instance, if we look at a vase
that is left-and-right symmetrical, then turn it 180o around Figure 4. An object having a
mirror plane need not have a
the vertical axis, it looks the same.”
2-fold rotational axis. This is
In the last sentence of the above quote, reflection or bilateral illustrated with the example
of a chair. There is a vertical
symmetry (a mirror plane) is casually being mixed up with 180°
mirror plane (shaded) which
rotational symmetry (a 2-fold rotation axis). If Feynman’s vase is brings it into self-coincidence.
left-and-right symmetrical, it has a mirror plane; however, it is not But there is no axis about
guaranteed to look the same after 180° rotation. The only way to which it can be rotated by
establish that the vase has a mirror plane is to carry out a 180° to bring it into self-coin-
reflection about the imaginary mirror plane in one’s mind, and not cidence.

a 180°rotation about an axis. Consider a chair as an example

(Figure 4). It does have a vertical mirror plane dividing it into
symmetric left and right halves. Still it cannot be rotated about
any axis by 180° to bring it into self-coincidence.

By the way, the work of Hermann Weyl that Feynman refers to is

a collection of four lectures delivered on the concept of symmetry
on the eve of his retirement from Institute of Advanced Study,
Princeton. This series of lectures, now available as a book entitled
Symmetry is a highly readable introduction to the concept [3].
Weyl devotes the very first lecture entirely to bilateral symmetry.
The second lecture discusses translational and rotational symme-
try. He concludes the series of lectures by the final one on
symmetry of crystals.

RESONANCE May 2016 433

 GENERAL ARTICLE

A crystal having a 5. ERROR 3: Ignoring the Absence of a Translational

cubic unit cell with one Symmetry in a Crystal
kind of atoms at the
corners and another
The following is a quote from Chapter 30 entitled ‘The Internal
Geometry of crystals’ in Vol 2 of FLP (Section 30-4).
kind at body-centres
does not have a body- “We would like to describe first the simplest lattices,
centred cubic lattice. It which are characteristic of most of the metals and of the
is described as a solids formed of the inert gases. They are cubic lattices
crystal with simple which can occur in two forms: the body-centered cubic,
cubic lattice and a shown in Fig. 30-4 (a), and the face-centered cubic shown
motif consisting of two in Fig.30-4 (b). The drawings show, of course, only one
atoms. cube of the lattice; you are to imagine that the pattern is
Crystallographers call repeated indefinitely in three dimensions. Also, to make
such structures B2 or the drawing clearer, only the “centres” of the atoms are
CsCl. shown. In an actual crystal, the atoms are more like spheres
in contact with each other. The dark and light spheres in the
Figure 5. Redrawing of
drawings may, in general, stand for different kinds of
Fig.30-4 (a) and (b) of FLP,
atoms or may be the same kind.”
Vol.2, Chapter 30. (a) is sup-
posed to represent body-
Fig.30-4 of FLP referred to in the above quote is redrawn below
centred cubic (bcc) crystal
as Figure 5.
and (b) face-centred cubic
(fcc) crystal. It is claimed in
In Figure 5a the drawing shows a cube with dark spheres at
the text of FLP that the dark
corners and a bright one at the centre of a cube. For the lattice
and light atoms may repre-
sent different elements. This generated by translations of this unit cell to be body-centred cubic
is incorrect. For these figures (bcc) the atom at the centre of the cube must (and not may as
to represent bcc and fcc the claimed in FLP) be of the same kind as that at the corners. If they
dark and light atoms must be are different, the translational symmetry from corner to body
of the same element.
centre is lost. This translational symmetry is a required symmetry
for the lattice to be bcc. In case of different
atoms occupying the centre and corners of
the unit cell (say A at the corner and B at the
body centre) the crystal can be described as a
simple cubic lattice with a two-atom motif.
Crystallographers refer to such structures as
type B2 or CsCl type.
(a) (b)

434 RESONANCE May 2016

 GENERAL  ARTICLE

A more complete description of motif should include the posi- A crystal having a
tions of atoms in the motif as fractional coordinates. The position cubic unit cell with
of atoms in a motif is given as their displacement coordinates with one kind of atom at
respect to the lattice point along the three crystallographic axes as the corners and
fractions of the respective lattice parameters (or unit cell edge another kind at face-
lengths). Thus the motif of CsCl crystal can be more accurately centres does not
have a face-centred
described as Cl at 000 and Cs at .
cubic lattice. It is
It is interesting here to consider this error in a bit more detail. It described as a
can be thought that the error arises from a failure to distinguish crystal with simple
between a lattice and a crystal. In a B2 structure (Figure 5a) the cubic lattice and a
atoms at the corners are different from the one at the body centre motif consisting of
of the cubic cell. Hence they do not form a lattice. Thus only the four atoms.
corners (or equivalently, only the body centres) can be selected as Crystallographers
lattice points. This makes the lattice simple cubic. However, to call such structures
get the crystal structure we now have to associate two atoms with L12 or AuCu3.
each lattice point: one at the corner 000 and the other at the body
centre . This pair of atoms constitutes the motif of the crystal.

A similar argument holds for Fig. 30-4 (b) of FLP (Figure 5b of

this article) which shows a cubic unit cell with dark atoms at eight
corners and light ones at the six face centres. If the cell has to
generate a crystal with a face-centred cubic (fcc) lattice, the
corner and face centre atoms must be the same and not may be.
If they are not the same, we get a crystal structure with a simple
cubic lattice and 4 atoms motif: a dark atom (say A) at a corner
and three white atoms (say B) at the centres of three faces meeting
at the corner. In terms of the fractional coordinates, the motif can

be described as A at 000 and 3 B’s at , , . An

example of such a structure is AuCu3 with Au and Cu atoms
playing the role of the dark and white atoms respectively. Crystal-
lographers call such structures as AuCu3 or L12 structure. Inter-
ested readers may consult [4] for details.

RESONANCE May 2016 435

 GENERAL ARTICLE

Failure to distinguish Failure to distinguish between a lattice and a crystal is rather

between a lattice and widespread in textbooks and on the web. Let us consider another
a crystal is rather example. There are only three cubic lattices: simple, body-centred
widespread in and face-centred. Obviously, there is no such thing as a diamond
textbooks and Web. cubic lattice. Instead, one should use the phrase diamond cubic
For example, structure. This structure has a face-centred cubic lattice with two-
incorrect phrases like atom motif with fractional coordinates 000 and . However,
‘diamond cubic
lattice’ and the incorrect phrase diamond cubic lattice is rather common. A
‘hexagonal close- google search for the exact phrase diamond cubic lattice (by
packed lattice’ or ‘hcp putting the phrase in quotes) gave me 6520 results in 0.37 seconds
lattice’ are quite on my computer! Similar is the case with the mistaken phrase
common. hexagonal close-packed (hcp) lattice. There is no such lattice, the
only hexagonal lattice being ‘simple hexagonal’. Hexagonal close-
packed structure can be described as a simple hexagonal lattice
with two-atom motif 000 and .

The present article alerts students to avoid pitfalls described

above. Also, it warns them to critically examine statements in
their textbooks even if the author is as famous as Feynman.

Note: In the digital edition of FLP at the official website

(www.feynmanlectures.caltech.edu) Error 1 has now been cor-
rected by redrawing Fig.1-4 as suggested in this article. An
attempt has also been made in the web edition to fix Error 2. But
this has been achieved by reducing the 3D ‘vase’of the original
FLP to a 2D ‘silhouette of a vase’ in the revised digital edition. In
my opinion this is not satisfactory. It would have probably been
better to change the phrase ‘turn it 180o around the vertical axis’
to ‘reflect it in a vertical mirror plane’ to preserve the three-
dimensionality. Error 3 remains uncorrected even in the latest
digital version. It could be corrected simply by dropping the
unnecessary last sentence of the paragraph, “The dark and light
spheres in the drawings may, in general, stand for different kinds
of atoms or may be the same kind ...” and changing the figure to
one having all atoms of the same colour.

436 RESONANCE May 2016

 GENERAL  ARTICLE

Acknowledgement

The author would like to thank Dr. Michael A Gottlieb, an editor

of the current editions of FLP, for fruitful discussions.
Address for Correspondence
Suggested Reading Rajesh Prasad
Department of Applied
[1] Richard P Feynman, Robert B Leighton and Matthew Sands, The
Mechanics
Feynman Lectures on Physics, 3 Volumes, Narosa, 1986.
Indian Institute of Technology
[2] T N Guru Row, Quasicrystals: A new state of matter, Resonance, Vol.18,
Delhi Hauz Khas
No.3, 2013.
New Delhi 110 106, India.
[3] Hermann Weyl, Symmetry, Princeton University Press, 1952.
Email: [email protected]
[4] Neil W Ashcroft and N David Mermin, Solid State Physics, Cengage,
2003.

RESONANCE May 2016 437