Since the square of every element of D
is proportional to the
identity operator
,
we have fifteen two-element, one-generator subgroups, each of
them composed of the identity and one of the other D
operators. We denote these subgroups by
,
where
is the generic symbol corresponding to one of the
non-identity elements of D
.
Obviously, only one choice of
the generator is possible for every of the two-element
subgroups.
Similarly, group D
has 35 different four-element subgroups,
which can be called the two-generator subgroups, and are denoted by symbols
.
pertaining to their generators.
The two-generator subgroups contain, in addition to
and
,
also the identity
and the product
.
Since this product is also one of the D
operators, we have in each of the four-element, two-generator subgroups three
possibilities to select the generators.
Finally, there are 15 different eight-element, three-generator subgroups
of D
,
denoted by
.
Each of these subgroups
contains the identity
,
the three generators,
three products of pairs of generators, and the product of
all three generators. Hence, to choose the generators
we may first pick any pair out of seven non-identity
elements (21 possibilities), and next pick any other
subgroup element, except the product of the first two,
(4 possibilities). Since the order in which we pick the
generators is irrelevant, one has altogether 28
possibilities of choosing the three generators in each of
the eight-element, three-generator subgroups of D
.
In the same way one can calculate that there is 168 different
choices of the four generators of the whole D
group;
one of them is, e.g., the set
.
This illustrates the degree of arbitrariness in implementing
calculations for which the whole group D
is conserved. Similar
freedom, although to a lesser degree, is available when conserving
any of the subgroups of D
.
Of course, the freedom of choosing
generators cannot influence the final results, however, it
allows using different quantum numbers, phase conventions, and
structure of matrix elements, as discussed in Sec. 3.
A classification of all the 65 non-trivial subgroups of D
(we do not include trivial subgroups
and D
itself)
is presented in
Table 1. Every subgroup is assigned to a certain type,
and described by a symbol given in the first column of the Table.
The types are defined according to: (i) the number of generators
in the subgroup (1, 2, or 3), (ii) the number of Cartesian axes
involved in the subgroup (0, I, or III standing for 0, 1, or 3),
and (iii) the number of signature operators in the subgroup (A, B,
or D standing for 0, 1, or 3).
The classification is based on two important characteristics of
each subgroup. As shown in Ref.[6], every conserved
symmetry, labeled by one of the Cartesian directions x, y, or
z, induces a specific symmetry of local densities, related to
this particular direction. Therefore, the number of Cartesian
axes involved in the subgroup gives us the number of symmetries of
local densities induced by the given subgroup. In addition,
the number of signature operators illustrates the way in which the
given subgroup is located with respect to the standard D
subgroup, which is composed of the three signatures.
Type | Generators | Generic | Total |
1-0A: |
![]() ![]() ![]() | 3 | 3 |
1-IA: |
![]() ![]() ![]() | 3 | 9 |
1-IB: |
![]() | 1 | 3 |
Total number of one-generator subgroups: | 7 | 15 | |
2-0A: |
![]() | 1 | 1 |
2-IA: |
![]() ![]() ![]() | 3 | 9 |
2-IB: |
![]() ![]() ![]() | 3 | 9 |
2-IIIA: |
![]() | 1 | 6 |
2-IIIB: |
![]() ![]() ![]() | 3 | 9 |
2-IIID: |
![]() | 1 | 1 |
Total number of two-generator subgroups: | 12 | 35 | |
3-IB: |
![]() | 1 | 3 |
3-IIIB: |
![]() ![]() ![]() | 3 | 9 |
3-IIID: |
![]() ![]() ![]() | 3 | 3 |
Total number of three-generator subgroups: | 7 | 15 | |
Total number of subgroups: | 26 | 65 |