MATRICE-7
MATRICE-7
1. Introduction
In this paper, in the study of k−potent elements we extend the results ob-
tained in [FB; 24] to generalised quaternion algebras and generalised octonion
algebras. As an application, by using the matrix representations for quaternions
and octonions, we give examples of classes of matrices which are k−potents and
a procedure to find such examples.
The paper is organised as follows: in the second paragraph we present matrix
representations and their properties which can be used to provide examples
of k−potent matrices and in the part three we study k−potents elements in
quaternion algebras and octonion algebras over the real field. The paper end
with conclusions and an idea for a further research.
2. Matrix representation
q = q0 − q1 f1 − q2 f2 − q3 f3
1
is called the conjugate of the element q. For q ∈ H(a, b), we consider the following
elements:
t (q) = q + q ∈ K
and
(q + q) q = q 2 + qq = q 2 + n (q) · 1
and
q 2 − t (q) q + n (q) = 0, ∀q ∈ H(a, b),
therefore the generalized quaternion algebra is a quadratic algebra.
If, for x ∈ H(a, b), the relation n (x) = 0 implies x = 0, then the algebra
H(a, b) is a division algebra. A quaternion non-division algebra is called a split
algebra.
Using the above notations, we remark that H (−1, −1) = −1,−1
R is a divi-
sion algebra.
A generalized octonion algebra over an arbitrary field K, with char K ̸= 2,
is an algebra of dimension 8, denoted O(a, b, c), with basis {1, f1 , ..., f7 } and
multiplication given in the following table:
· 1 f1 f2 f3 f4 f5 f6 f7
1 1 f1 f2 f3 f4 f5 f6 f7
f1 f1 α f3 af2 f5 af4 − f7 −af6
f2 f2 −f3 b −bf1 f6 f7 bf4 bf5
f3 f3 -af2 bf1 −ab f7 af6 −bf5 −abf4
f4 f4 −f5 − f6 − f7 c − cf1 −cf2 − cf3
f5 f5 -af4 − f7 - af6 cf1 -ac cf3 acf2
f6 f6 f7 −bf4 bf5 cf2 −cf3 -bc −bcf1
f7 f7 af6 −bf5 abf4 cf3 −acf2 bcf1 abc
t (x) = x + x ∈ K
and
2
These elements are called the trace, respectively, the norm of the element x ∈
O(a, b, c). It follows that
(x + x) x = x2 + xx = x2 + n (x) · 1
and
x2 − t (x) x + n (x) = 0, ∀x ∈ x ∈ O(a, b, c),
therefore the generalized octonion algebra is a quadratic algebra.
If, for x ∈ O(a, b, c), the relation n (x) = 0 implies x = 0, then the algebra
O(a, b, c) is a division algebra ( see [Sc; 54] and [Sc; 66]).
If we take a = b = c = −1, K = R, then we obtain H (−1, −1), the
quaternion division algebra, usually denoted by H and octonion division al-
gebra O (−1, −1, −1), usually denoted by O. For example, by taking a = −1
and b = 1, K = R, we obtain a split quaternion
algebra. In the following, we
will denote HK = −1,−1 and OK = −1,−1,−1
K K .
We know that a finite-dimensional associative algebra A over an arbitrary
field K is algebraically isomorphic to a subalgebra of a matrix algebra over the
same field K. Therefore, each element a ∈ A has a matrix representation. That
means, there is a map f : A → Mn (K) such that f (x) = Mx ∈ Mn (K), where
dimA = n. For an arbitrary quaternion algebra H (a, b), the map
φ : H (a, b) → M4 (K)
q0 aq1 bq2 −abq3
q1 q0 bq3 −bq2 , q ∈ H (a, b) ,
φ (q) = (1)
q2 −aq3 q0 aq1
q3 −q2 q1 q0
is called the left representation and the map
ρ : H (a, b) → M4 (K)
q0 aq1 bq2 −abq3
q1 q0 −bq3 bq2
, q ∈ H (a, b) ,
ρ (q) = (2)
q2 aq3 q0 −aq1
q3 q2 −q1 q0
where q = q0 + q1 f1 + q2 f2 + q3 f3 , is called the right representation ( see [Ti;
00]).
In the same paper [Ti; 00], were defined, for real octonions, two representa-
tions maps, left and right representations, by using the maps φ and ρ, defined on
real quaternions. These maps can be defined for all octonion algebras O (a, b, c)
over an arbitrary field K, namely
Φ : O (a, b, c) → M8 (K) ,
φ (x′ ) −ρ (x′′ ) E4
Φ (x) = ,
φ (x′′ ) E4 ρ(x′ )
3
the left representation, where the octonion x can be written under the form
x = x′ + x′′ f , with x′ , x′′ ∈ H (a, b), by using the Cayley-Dickson process. In
the same way, we define the right representation
Ψ : O (a, b, c) → M8 (K) ,
ρ (x′ ) −φ x′′
Ψ (x) = ,
φ (x′ ) ρ(x′ )
where x′ , x′′ are the conjugates of the quaternions x′ and x′′ and E4 = diag (1, −1, −1, −1).
For x ∈ O (a, b, c), x = x0 + x1 f1 + x2 f2 + x3 f3 + x4 f4 + x5 f5 + x6 f6 + x7 f7
with {1, f1 , ...., f7 } the base in O (a, b, c), we have
x0 ax1 bx2 −abx3 cx4 −acx5 −bcx6 abcx7
x1
x0 bx3 −bx2 cx5 −cx4 bcx7 −bcx6
x2 −ax3 x 0 ax 1 cx6 −acx 7 −cx 4 acx5
x3 −x2 x1 x0 cx7 −cx6 cx5 −cx4
Φ (x) = (3)
x4 −ax5 −bx6 abx7 x0 ax1 bx2 −abx3
x5 −x4 −bx7 bx6 x1 x0 −bx3 bx2
x6 ax7 −x4 −ax5 x2 ax3 x0 −ax1
x7 x6 −x5 −x4 x3 x2 −x1 x0
and
x0 ax1 bx2 −abx3 cx4 −acx5 −bcx6 abcx7
x1 x0 −bx3 bx2 −cx5 cx4 −bcx7 bcx6
x2 ax3 x0 −ax1 −cx6 acx7 cx4 −acx5
x3 x2 −x1 x0 −cx7 cx6 −cx5 cx4
Ψ (x) = (4)
x4 ax5 bx6 −abx7 x0 −ax1 −bx2 abx3
x5 x4 bx7 −bx6 −x1 x0 bx3 −bx2
x6 −ax7 x4 ax5 −x2 −ax3 x0 ax1
x7 −x6 x5 x4 −x3 −x2 x1 x0
Proposition 1. ([Ti; 00], Lemma 1.2) With the above notations, for ε ∈
{φ, ρ}, we have:
i) ε (x + y) = ε (x) + ε (y) ,
ii) ε (xy) = ε (x) ε (y) ,
iii) ε (λx) = λε (x) , λ ∈ K, ε (1) = I4
T
iv) ε (x) = ε (x)
−1
v) ε x = ε−1 (x)
vi) ε (x) = ε (y) if and only if x = y, with x, y ∈ H.
We remark that the above properties are proved in the real case for division
algebra H, but these are true in the general case, for an arbitrary field with
characteristic different from 2. Properties i), iii), iv) and v) are also satisfied for
the maps Φ and Ψ over reals. Moreover, the following properties were proved
over the real field.
4
Proposition 2. ([Ti;00],Theorem 2.5, Theorem 2.9, Theorem 2,10, Theo-
rem 2.11)
With the above notations, for ε ∈ {Φ, Ψ}, we have
i) ε (x +y) = ε (x) + ε (y) ,
2
ii) ε x2 = ε (x)
iii) ε (xyx) = ε (x) ε (y) ε (x) ,
iv) ε (λx) = λε (x) , λ ∈ K, ε (1) = I8
v) ε (x) = εT (x)
vi) ε (x) = ε (y) if and only if x = y, where x, y ∈ O.
Properties ii) and iii) from the above proposition were proved by using al-
ternativity and the the Moufang identities. But, since Moufang identities are
true in any octonion algebra over a field of characteristic not two (see [Sc; 66]),
these properties are also true for K = Zp , p a prime number, p ̸= 2.
Proposition 3. With the notations from the above proposition, we have
that
ε (xn ) = εn (x) ,
for x ∈ O (a, b, c) over an arbitrary field K and n a positive integer.
Proof. We use induction. From condition ii) from the above proposition,
taking x = y,we have ε x3 = ε3 (x). From the same condition, for y = x2 ,
we have ε x4 = ε4 (x) and so on.
In the paper [FB; 24], we studied some properties of k−potent elements over
algebras obtained by the Cayley-Dickson process.
Definition 4.
i) The element x in the ring R is called nilpotent if there is a positive integer
n such that xn = 0. The number n is the smallest with this property and is
called the nilpotency index. ii) The element x in the ring R is called a k-potent
element, for k > 1, a positive integer, if k is the smallest number such that
xk = x. The number k is called the k-potency index. For k = 2, we have
idempotent elements, for k = 3, we have tripotent elements, and so on.
Remark 5. From the above definition and Proposition 3, if x ∈ H (a, b)
or x ∈ O (a, b, c) is a k−potent element, it results that the matrices ε (x) , for
ε ∈ {φ, ρ} or ε ∈ {Φ, Ψ} are k−potent matrices over the field K.
Example 6. i) By using results obtained in [FB; 24], we consider quater-
nions over the field K = Z5 and the element x = 2 + 3i + j + 3k which is a
5−potent element over HZ5 . Indeed, x = 2 + 3γ, γ = i + 2j + k, with γ 2 = −1
and nx = 1. Therefore, x2 = 2γ and x4 = 1. The matrices
2 −3 −1 −3 2 −3 −1 −3
3 2 −3 1 3 −1
and ρ (x) = 3 2
φ (x) =
1 3 2 −3 1 −3 2 3
3 −1 3 2 3 1 −3 2
5
ii) With the same arguments as above, if we consider octonions over the field
K = Z13 and the element x ∈ OZ13 , x = 3 + 2f1 + f2 + f3 + f4 + f5 + f6 + f7 ,
we obtain that x is a 13−potent element . The matrices
3 −2 −1 −1 −1 −1 −1 −1
2 3 −1 1 −1 1 1 −1
1 1
3 −2 −1 −1 1 1
1 −1 2 3 −1 1 −1 1
Φ (x) =
1 1 1 1 3 −2 −1 −1
1 −1 1 −1 2 3 1 −1
1 −1 −1 1 1 −1 3 2
1 1 −1 −1 1 1 −2 3
and
3 −2 −1 −1 −1 −1 −1 −1
2 3 1 −1 1 −1 −1 1
1 −1 3 2 1 1 −1 −1
1 1 −2 3 1 −1 1 −1
Ψ (x) =
1 −1 −1 −1 3 2 1 1
1 1 −1 1 −2 3 −1 1
1 1 1 −1 −1 1 3 −2
1 −1 1 1 −1 −1 2 3
are also 13-potent matrices.
xk−1 = 1. (5)
In the following, we will provide solutions for this equation. From relation (5),
we obtain that nk−1 x = 1. Since nx is a positive number, we obtain that nx = 1.
Let x ∈ H, x = x0 + x1 f1 + x2 f2 + x3 f3 , with nx = x20 + x21 + x22 + x23 = 1, with
x0 , x1 , x2 , x3 ∈ (−1, 1). We denote with
q
cos α = x0 , sin α = x21 + x22 + x23 ,
x1 f1 + x2 f2 + x3 f3
θ= p ∈ H, with θ2 = −1.
x21 + x22 + x23
6
and the element x can be write under the form
x = cos α + θ sin α.
2
It is clear that x2 = (cos α + θ sin α) = cos2 α − sin2 θ + (2 cos α sin α)θ =
cos 2α + θ sin 2α. Using induction, we obtain that
By using the above relation, the element x ∈ H satisfing condition xk−1 = 1 has
the form
2π 2π
x = cos + θ sin .
k−1 k−1
Example 7. i) We consider x = 12 + 12 f1 + 12 f2 + 12 f3 = cos π3 + θ sin π3 ,
where θ = f1 +f
√2 +f3 . We obtain that x6 = 1, then x7 = x and x is a 7−potent
3
element. Therefore, the matrices
1
− 12 − 12 − 21
1
− 12 − 12 − 21
2 2
1 1
− 12 1 1 1 1
− 21
φ (x) = 2 2 2 and ρ (x) = 2 2 2
1 1 1 1 1 1 1 1
2 2 2 −2
2 −2 2 2
1
2 − 12 1
2
1
2
1
2
1
2 − 21 1
2
− 12 − 12 1
− 12
1
− 12 1
− 21
2 −2 2
1
2 − 12 − 12 − 12 1
and ρ (q) = 21 − 12 1
2
1
2
φ (q) =
− 12 1
2 − 12 − 12 −
2 − 12 − 21 1
2
1 1 1
2 2 2 − 12 1
2 − 12 − 21 − 12
are 4−potent matrices. √ √
iii) For x = 12 f1 − 21 f2 + 22 f3 = cos π2 + θ sin π2 , where θ = f1 −f22+ 2f3 . We
obtain that x4 = 1, then x5 = x and x is 5−potent element. The matrices
√ √
0 − 12 1
2√ − 22 0 − 21 1
√2
− 22
0 − 22 − 12 2
1 1 1
0√
φ (q) = 21 √2
2 2 2
and ρ (q) =
−2 1
1 2 1
√ 2 0 − 2
− 2 −
√ 2 0 2
2 1 1 2
2 2 2 0 2 − 21 − 12 0
are 5−potent matrices.
Remark 8. In the following, we consider the quaternion split algebra
H (a, b) or octonion split algebra O (a, b, c) and K = R. If an element x ∈ H (a, b)
or x ∈ O (a, b, c) is k−potent with nx = 0 and tx ̸= 0, it results that tkx = tx ,
therefore tx = 1 or tx = −1 and x0 = t2x . We have that t2x = tx or t3x = tx , there-
fore there are only idempotent and tripotent elements in H (a, b) or O (a, b, c)
over R.
7
Example 9.
i) We consider the split quaternion algebra H (1, 1) and the quaternion q =
1
2 (1 + f1 + f2 + f3 ), with n (q) = 0. We have that the matrices
1 1 1
− 21
1 1 1
− 12
2 2 2 2 2 2
1 1 1
− 21 1 1 −1 1
φ (q) = 2 2 2 2 21 2 2 ,
1 and ρ (q) = 1
1 −1 1 1
− 12
2 2 2 2 2 2 2
1
2 − 12 12 1
2
1
2
1
2 − 21 1
2
1
are idempotent matrices. For w = 2 (−1 + f1 + f2 + f3 ), the matrices
− 21 1 1
− 21
1 1 1
− 21
2 2 −2 2 2
1
2 − 12 1
2 − 21 1
2 − 12 − 21 1
2
φ (w) = 1 1 and ρ (w) = 1
2 − 12 − 12 2 2
1
2 − 21 − 21
1
2 − 12 1
2 − 21 1
2
1
2 − 21 − 21
1 2 2
− 21
1 1
0√ 0 −
φ (z) =
2 2 and ρ (z) = 12
√ 2 2
1
− 22 1 2
− 12
√2
0 2
2
√ 2 0
2 2
2 − 21 1
2 0 2
1
2 − 12 0
are idempotent. √
If we take q2 = 12 −1 + f1 + f2 + 36 f3 , with n (q2 ) = 0,we have that the
matrices
1 3
√ 1 3
√
−2 1 √2
− 6 −2 1 2√ − 6
6
1 −√12 − 23 1 − 1
− 26 3
φ (q1 ) = 21 2 2 √2 2
and ρ (q ) =
1
− 36 − 12 √12 6
− 1
−1
√2 1 3 2
6 1 1 1 6 1 1 1
6 −2 2 −2 6 2 −2 −2
are tripotent. √
1
If we take q3 = 2 1+ 2f1 + f2 + f3 , the matrices
8
1
√ 3
1
√ 3
√2
2 2 −3 √2
2 2 −3
2 1 3 3 2 1
− − 32 3
φ (q3 ) =
2
1
2 2
1
√2 and ρ (q3 ) =
2
1
2
1
√
2
−1 2 1 − 2
2 √2
2 2√
2
1
2 − 12 2
1
2
1
2
1
2 − 22 1
2
1
√
are idempotent. For q4 = 2 −1 + 2f1 + f2 + f3 , we obtain the following
tripotent matrices
1
√ 3
1
√ 3
−
√2
2 2 −3 −
√2
2 2 −3
2 2
− 12 3
− 3
− 12 − 32 3
φ (q4 ) =
2
1
2
1
√2 and ρ (q4 ) =
2
1
√
2 .
2 −1 −
√2
2 2 1 −√12 − 2
2
1
2 − 12 2 − 12 1
2
1
2 − 22 − 21
References
[FB; 24] Flaut, C., Baias, A., Some Remarks Regarding Special Elements in
Algebras Obtained by the Cayley–Dickson Process over Zp ,Axioms, 13(6)(2024),
351; https://doi.org/10.3390/axioms13060351
[Sc; 54] Schafer, R. D., On the algebras formed by the Cayley-Dickson process,
Amer. J. Math., 76(1954), 435-446.
[Sc; 66] Schafer, R. D., An Introduction to Nonassociative Algebras, Aca-
demic Press, New-York, 1966.
[Ti; 00] Tian, Y., Matrix Representatios of Octonions and their Applications,
https://arxiv.org/pdf/math/0003166, 2000.
Cristina FLAUT
Faculty of Mathematics and Computer Science, Ovidius University,
Bd. Mamaia 124, 900527, Constanţa, România,
http://www.univ-ovidius.ro/math/
e-mail: [email protected]; cristina [email protected]
9
Andreea BAIAS
PhD student at Doctoral School of Mathematics,
Ovidius University of Constanţa, România,
e-mail: [email protected]
10