





Study with the several resources on Docsity
Earn points by helping other students or get them with a premium plan
Prepare for your exams
Study with the several resources on Docsity
Earn points to download
Earn points by helping other students or get them with a premium plan
The identification of translations and orthogonal transformations in the group of motions of the plane. It covers the properties of plane crystallographic groups, chasles's theorem, and the classification of motions as rotations, reflections, translations, or glide reflections. The document also explains how to determine the type of a lattice based on its symmetry group.
Typology: Study notes
1 / 9
This page cannot be seen from the preview
Don't miss anything!






Let T be the subgroup of the group of motions M (2) of the plane P which consists of translations. The subgroup T is a normal subgroup of M (2). For each g ∈ M (2) and ta ∈ T we have
g ◦ ta = tg(a) ◦ g. (1)
Fix a point O on the plane so that any vector is equal to a vector a = OP~. thus we can identify vectors with their end points P. Then vectors form a vector space of dimension 2. If we fix a rectangular coordinate system, then any vector a has coordinates (a, b). Let O(2) be the group of motions which fix the point P. Then we have proved in class that any g ∈ O(2) is a linear transformation and its matrix in the basis i, j is an orthogonal matrix A. It is of determinant 1 if g is a rotation, and of determinant −1 if g is reflection with respect to a line ` containing O. Consider a homomorphism
φ : M → O(2)
defined as follows. For any motion g we have
t−g(O) ◦ g(O) = O.
This shows that any element m ∈ M (2) is of the form m = ta ◦ g, where g ∈ O(2). Since a = m(O), the end point of the vector a, the vector a is defined uniquely by g, hence the orthogonal part g is defined uniquely. We set φ(ta ◦ g) = g.
By (1), we have
(ta ◦ g) ◦ (tb ◦ g′) = ta ◦ (g ◦ tb) ◦ g′^ = (ta ◦ tg(b)) ◦ (g ◦ g′) = ta+g(b) ◦ g ◦ g′.
Thus φ((ta ◦ g) ◦ (tb ◦ g′)) = g ◦ g′. This checks that φ is a homomorphism. By Chasles’s Theorem any motion is either rotation ρθ(c) about an angle θ with fixed point c, or a reflection rwith the mirror line, or a translation ta, or a glide reflection ta ◦ r`. We can write
ρθ(c) = tc ◦ ρθ ◦ t−c,
r= ta ◦ r′ ◦ t−a,
where ′^ is the line parallell to and containing O, and ta() =′. It follows from the definition of φ that φ(ta) = 1, the identity map. Since φ is a homomorphism, we get
φ(ρθ(c)) = ρθ, φ(r) = r′.
It follows from the definition of φ that its kernel is T and the image is the whole O(2). Thus M (2)/T ∼= O(2).
Lemma 1. Let g = ta ◦ r. Then g = tb ◦ r′ , where the vector b is parallel to the line ′. Here b is equal to the orthogonal projection of a to, and ′^ is the line parallel to which passes through the end point of the vector 1 2 (a^ −^ b).
Proof. Choose coordinate system so that is the x-axis. Let a = (a, b). Consider the line′^ given by the equation y = b/2. For any x = (x, y) we have
t(a,0) ◦ r′^ (x) = (x, y − 2(y − b/2)) + (a, 0) = (x + a, −y + b) = ta ◦ r(x).
b
a
A glide reflection is the motion g of the form ta ◦ r, where a is not orthogonal to the line (equivalently, its orthogonal projection to is not equal to zero). Note that if a is orthogonal to then g is the reflection with respect to a line parallel to `. The main difference between a glide reflection
Proof. If g ∈ Γ, then¯ h = ta ◦ g ∈ Γ for some translation ta ∈ T. For any x ∈ L we get
tg(x) = g ◦ tx ◦ g−^1 = (t−a ◦ h) ◦ tx ◦ (h−^1 ◦ ta)
= t−a ◦ (h ◦ tx ◦ h−^1 ) ◦ ta) = t−a ◦ th(x) ◦ ta = th(x ∈ Γ.
This shows that g(x) ∈ L.
Note that elements g in the point group Γ are not necessary in Γ. How-¯ ever, ta ◦ g ∈ Γ for some translation ta (not necessary in LΓ).
Lemma 3. The group ¯Γ is a finite group. Its intersection with O(2)+^ (the subgroup of rotations) is a cyclic group Cn of order n = 1, 2 , 3 , 4 or 6. If ¯Γ contains a reflection, then it is a group D 2 , D 3 , D 4 ,or D 6.
Proof. One proves that Γ is finite because Γ is discrete (we omit the proof).¯ Thus any g ∈ ¯Γ ∩ O(2)+^ is represented by an orthogonal matrix of finite order. It looks as
A =
cos θ − sin θ sin θ cos θ
, θ = 2π/n.
Let L = Za + Za. Since g(L) ⊂ L, we get
g(a) = m 1 a + n 1 b, g(b) = m 2 a + n 2 b,
for some integers m, n. The matrix
m 1 n 1 m 2 n 2
represents g in the basis (a, b). Hence B = CAC−^1 for some matrix C (the basis change matrix). Also, this implies that the traces of A and B coincide (it is the sum of the eigenvalues of g). This gives
T r(A) = 2 cos 2πi/n = T r(B) = m 1 + n 2 ∈ Z.
Thus 2 cos 2πi/n = 0, ± 1 , ±2. This gives 2π/n = π/ 2 , 3 π/ 2 , π/ 3 , 2 π/ 3 , 0 , π. This proves the first assertion. If Γ contains a reflection, then it is a dihedral group with the cyclic subgroup of index 2 of one of Cn from the assertion.
Definition 1. A lattice L is called rectangular (resp. hexagonal) if one can choose a rectangular fundamental parallelogram (resp. parallelogram with two equal side and the angle π/ 3 between them). A lattice is called half-rectangular if it is spanned by a side of a rectangular and its middle point.
In the following, if L is rectangular (resp. hexagonal) we will choose a basis (a, b) of a lattice L which defines a fundamental parallelogram which is a rectangle (resp. has equal length sides with the angle π/3 between the sides). However, note that there are bases of the same lattice of different shape.
Assume that a reflection r∈ Γ.¯ Then, for any x ∈ L, x + r(x) ∈ L. But it is easy to see that this vector is equal to 2pr(x), the twice of the orthogonal projection of x to. Thus is spanned by a vector from the lattice L. By taking a lattice vector on of smallest length, we may assume that ` is spanned by a. Now
2 pr(b) = b + r(b) ∈ L,
hence, after replacing b with b + ma for some m ∈ Z, we may assume that pr`(b) = 0, or 12 a. In the first case we get a rectangular lattice, in the second case we get a half-rectangular lattice. Now we are ready to classify all crystallographic groups.
In this case Γ contains only translations and glide reflections. Also the composition of two glide reflections whose glide vectors are not parallel has a fixed poins since
(tc ◦ r) ◦ (tc′^ ◦ r′^ ) = tc ◦ tr(c′) ◦ r′^ ◦ r′^ = tc+r(c′) ◦ ρ 2 θ,
where θ is the angle between and′. Thus all glide reflections s,c have lines parallel to the same vector. Let s = s,c ∈ Γ. By changing O we may assume that O ∈. Applying (2) we see that 2c ∈ L. Let a′^ be the smallest length vector in L which lies on . By composing s = tc ◦ r with tma with m ∈ Z we may replace c with a vector equal to a′/2. We know that r`(L) ⊂ L. As we observed before this implies that L is a a rectangular lattice or a half-rectangular lattice. Let ΩΓ be spanned by a′^ and b′.
Lemma 2 to write it as tc ◦ r′^ , where′^ is parallel to and the distance between the lines and ′^ is equal to the length of the vector 12 (x − pr(x)). Write x = ma + nb for some integers m, n. Then x − pr(x) = nb, and hence the length of 12 (x − pr(x)) is equal to n||b||/2.
III 2 : L = Za + Zb is a hexagonal lattice, Γ¯ ∼= C 2 is generated by r. The mirror lines of reflections from Γ are lines parallel to with distance between each other equal to an integer multiple of
3 ||b||/2. b
O^ a
Similar to the previous case. We note that the vector c = −a+b belongs to L. It is perpendicular to a and its length is
3 ||a|| =
3 ||b||.
III 3 : L = Za + Zb is a square lattice, Γ¯ ∼= D 2 is generated by ρπ, r`. The mirror lines of reflections from Γ are lines parallel to a or b with distance between parallel lines equal to an integer multiple of ||a||/2. b
O (^) a
We use (3) to see that lines spanned by a and b are mirror lines, and then get more applying the argument from case III 1.
III 4 : L = Za + Zb is a square lattice, Γ¯ ∼= D 4 is generated by ρπ/ 2 , r`. The mirror lines of reflections from Γ are lines parallel to a, b, c = a + b with distance between parallel lines equal to an integer multiple of
3 ||a||/2. b
O (^) a
III 5 : L = Za + Zb is a hexagonal lattice, ¯Γ ∼= D 3 is generated by ρ 2 π/ 3 , r`. The mirror lines of reflections from Γ are lines parallel to a or bwith distance
between parallel lines equal to an integer multiple of
3 ||a||/2. b
O^ a We use (3) to see that lines spanned by a and b are mirror lines, and then get more applying the argument from case III 2.
III 6 : L = Za + Zb is a hexagonal lattice, Γ¯ ∼= D 6 is generated by ρπ/ 3 , r`. The mirror lines of reflections from Γ are lines parallel to a, b, a − b with distance between parallel lines equal to an integer multiple of
3 ||a||/2. b
O^ a
Since Γ is not of type I it must contain either rotations or reflections. In fact, if it contains a reflection then the composition with a glide reflection will be a rotation. So, we may assume that Γ contains a non-trivial rotation ρθ. We choose a fundamental parallelogram such that c = a/2 and use that
s,c ◦ ρθ = (tc ◦ r) ◦ ρθ = tc ◦ rρθ/ 2 (`), (4)
This shows that the projection of 2c to the line ρθ/ 2 (`) belongs to the lattice LΓ.
IV 1 , IV 2 : L = Za + Zb is a rectangular lattice, Γ¯ ∼= D 2 is generated by ρπ, r. Γ contains glide reflections s,c parallel to a reflections with respect to lines parallel to b. The distances between the lines parallel to a (resp.b) are integer multiple of the half-lengths of b (resp. a). In this case θ = π, hence in (4) we have c is perpendicular to ρπ/ 2 (). As we had remarked before this means that s,c ◦ ρθ is a reflection with respect to a line parallel to a line perpendicular to `. This implies that L is rectangular (case IV 1 ) or half-rectangular (case IV 2 ). The other properties had been explained before.