1. Introduction
The definition of twisted Heisenberg Lie algebras was motivated by the connection with differential geometry. More specifically, the twisted Heisenberg Lie algebra is the tangent algebra of a twisted Heisenberg group (certain semidirect product of a Heisenberg group with the real numbers) [Reference Calderón, Draper, Martín and Sánchez3]. Its importance is due to the relationship between twisted Heisenberg algebras and Lorentzian manifolds [Reference Adams1, Reference Günther10]. For example,
Theorem 1.1 ([Reference Adams1], Theorem 11.7.3)
Let
$M$ be a compact connected Lorentzian manifold and
$G$ a connected Lie group acting isometrically and locally faithfully on
$M$. Then its Lie algebra
$\mathfrak{g} = \mathfrak{k}\oplus \mathfrak{a} \oplus \mathfrak{s}$ is a direct sum of a compact semisimple Lie algebra
$\mathfrak{k}$, an abelian algebra
$\mathfrak{a}$ and a Lie algebra
$\mathfrak{s}$, which is either trivial, or isomorphic to
$\mathfrak{aff}(\mathbb{R})$, to a Heisenberg algebra
$H_n$, to a twisted Heisenberg algebra
$H^{\lambda}
_n$ with
$\lambda\in \mathbb{Q}_{+}^{(n-1)/2}$, or to
$\mathfrak{sl}_2(\mathbb{R})$.
Algebraically, the twisted Heisenberg Lie algebra can be viewed as a certain one-dimensional extension of the corresponding Heisenberg Lie algebra, as shown in the following definitions:
Definition 1.2. For a non-negative integer
$m$, the Heisenberg Lie algebra
$\mathfrak{h}_{m}$ is the
$(2m+1)$-dimensional vector space over a field
$\mathbb F$ spanned by the elements
with the non-vanishing Lie brackets
for
$1\le i\le m$.
Definition 1.3. For a non-negative integer
$m$, the twisted Heisenberg Lie algebra
$\mathfrak{h}^{\lambda}_{m}$, parameterized by
$\lambda=(\lambda_1,\ldots,\lambda_m)\in(\mathbb F^{\times})^{m}$, is the
$(2m+2)$-dimensional vector space over a field
$\mathbb F$ spanned by the elements
with the non-vanishing Lie brackets
for
$1\le i\le m$.
The cohomology with trivial coefficients of Heisenberg Lie algebras was computed by Santharoubane [Reference Santharoubane14] over fields of characteristic zero in 1983. In 2005, Sköldberg [Reference Sköldberg16] used algebraic Morse theory to compute the Poincaré polynomial of Heisenberg Lie algebras over fields of characteristic
$p=2$. In 2008, Cairns and Jambor [Reference Cairns and Jambor2] gave the
$n$-th Betti number of
$\mathfrak{h}_m$ over fields of prime characteristic for
$n \le m$.
In 1937, Jacobson [Reference Jacobson13] introduced the notion of restricted Lie algebras over fields of positive characteristic and proved that a Lie algebra is restrictable precisely when all
$p$-powers of the inner derivations are still inner. This gives us many examples of restricted Lie algebras, like 2-step nilpotent Lie algebras, the derivation algebra
$\text{Der}(A)$ of any algebra
$A$, the Lie algebra of an algebraic group, the primitive elements of an irreducible co-commutative Hopf algebra, etc. However, certain technical tools are available for restricted Lie algebras which are not available for arbitrary modular Lie algebras, see [Reference Seligman15] for examples.
The restricted structures on
$\mathfrak{h}_{m}$ were classified and characterized by Jacobson’s theorem in [Reference Evans, Fialowski and Yang6]. However, much less is known about restricted twisted Heisenberg algebras. In this note, we study the family of restricted structures on the twisted Heisenberg Lie algebra
$\mathfrak{h}_m^{\lambda,\mu}$ parameterized by the elements
$\mu \in \mathbb F^{2m+2}$. By considering
$\mathfrak{h}_m^{\lambda,\mu}$ as a restricted one-dimensional central extension of the restricted Heisenberg Lie algebra, we compute the ordinary and restricted cohomology spaces
$H^q(\mathfrak{h}_m^{\lambda,\mu})$ and
$H_*^q(\mathfrak{h}_m^{\lambda,\mu})$ for
$q=1,2$, and give the bracket structures and
$[p]$-operators for the corresponding restricted one-dimensional central extensions of
$\mathfrak{h}_m^{\lambda,\mu}$.
2. Restricted Lie algebras and cohomology
In this section, we recall the definition of a restricted Lie algebra, the Chevalley–Eilenberg cochain complex, and the restricted cochain complex [Reference Evans and Fuchs7]. Everywhere in this section,
$\mathbb F$ denotes an algebraically closed field of characteristic
$p$ and
$\mathfrak g$ denotes a finite-dimensional Lie algebra over
$\mathbb F$ with an ordered basis
$\{e_1,\dots , e_n\}$. For
$j\geq 2$ and
$g_1,\ldots,g_ j\in \mathfrak{g}$, we denote the
$j$-fold bracket
2.1. Restricted Lie algebras
Suppose that
$\mathrm{char}\ \mathbb F=p \gt 0$. A restricted Lie algebra
$\mathfrak{g}$ over
$\mathbb F$ is a Lie algebra
$\mathfrak{g}$ over
$\mathbb F$ together with a mapping
$\cdot^{[p]}:\mathfrak g\to\mathfrak g$, written
$g\mapsto g^{[p]}$, such that for all
$a\in\mathbb F$, and all
$g,h\in \mathfrak{g}$
(1)
$(a g)^{[p]}=a^{p}g^{[p]}$;(2)
$(g+h)^{[p]}=g^{[p]}+h^{[p]}+ \sum\limits_{i=1}^{p-1} s_i(g,h)$, where
$is_i(g,h)$ is the coefficient of
$t^{i-1}$ in the formal expression
$(\mathrm{ad}(tg+h))^{p-1}(g)$; and(3)
$(\mathrm{ad}\ g)^{p}=\mathrm{ad}\ g^{[p]}$.
The mapping
$\cdot^{[p]}$ is called a
$[p]$-operator on
$\mathfrak{g}$ and a Lie algebra is restrictable if it is possible to define a
$[p]$-operator on it. By Jacobson’s theorem [Reference Jacobson13, Reference Strade and Farnsteiner17], a finite-dimensional Lie algebra
$\mathfrak g$ is restrictable if and only if it admits a basis
$\{e_1,\dots, e_n\}$ such that
$(\mathop{\rm ad} e_i)^p$ is an inner derivation for all
$1\le i\le n$. Choosing
$e_i^{[p]}\in\mathfrak g$ with
$\mathop{\rm ad} e_i^{[p]}=(\mathop{\rm ad} e_i)^p$ for all
$i$ completely determines a
$[p]$-operator on
$\mathfrak{g}$.
2.2. Chevalley–Eilenberg Lie algebra cohomology
We only describe the Chevalley–Eilenberg cochain spaces
$C^q(\mathfrak g)=C^q(\mathfrak g,\mathbb F)$ for
$q=0,1,2,3$ and differentials
$d^q:C^q(\mathfrak g)\to C^{q+1}(\mathfrak g)$ for
$q=0,1,2$. For more details on the Chevalley–Eilenberg cochain complex we refer the reader to [Reference Chevalley and Eilenberg4] and [Reference Fuchs9]. Set
$C^0 (\mathfrak g)=\mathbb F$ and
$C^q (\mathfrak g)= (\wedge^q\mathfrak g)^*$ for
$q=1,2,3$. We will use the following bases, ordered lexicographically, throughout the paper.
\begin{align*}
C^0 (\mathfrak g):& \{1\}&\\
C^1 (\mathfrak g):& \{e^k\ |\ 1\le k\le n\}&\\
C^2 (\mathfrak g):& \{e^{i,j}\ |\ 1\le i \lt j\le n\}&\\
C^3 (\mathfrak g):& \{e^{u,v,w}\ |\ 1\le u \lt v \lt w\le n\}&\\
\end{align*} Here
$e^k$,
$e^{i,j}$ and
$e^{u,v,w}$ denote the dual vectors of the basis vectors
$e_k\in \mathfrak g$,
$e_{i,j}=e_i\wedge e_j\in \wedge^2\mathfrak g $ and
$e_{u,v,w}=e_u\wedge e_v\wedge e_w\in \wedge^3\mathfrak g$, respectively. The differentials
$d^q:C^q (\mathfrak g)\to C^{q+1}(\mathfrak g)$ are defined for
$\psi\in C^1 (\mathfrak g)$,
$\varphi\in C^2 (\mathfrak g)$ and
$g,h,f\in\mathfrak g$ by
\begin{align*}
d^0: C^0 (\mathfrak g)\to C^1 (\mathfrak g), &\ d^0=0&\\
d^1:C^1 (\mathfrak g)\to C^2 (\mathfrak g), &\ d^1(\psi)(g\wedge h)=\psi([g,h])&\\
d^2:C^2 (\mathfrak g)\to C^3 (\mathfrak g), &\ d^2(\varphi)(g\wedge h\wedge f)=\varphi([g,h]\wedge f)-\varphi([g,f]\wedge h)+\varphi([h,f]\wedge g). &\\
\end{align*} The mappings
$d^q$ satisfy
$d^{q}d^{q-1}=0$ and
$H^q(\mathfrak g)=H^q(\mathfrak g,\mathbb F)=\ker(d^q)/\mathop{\rm im}(d^{q-1})$.
2.3. Restricted Lie algebra cohomology
In this subsection, we recall the definitions and results on the restricted cochain complex only for the case of trivial coefficients. For a more general treatment of the cohomology of restricted Lie algebras, we refer the reader to [Reference Evans and Fuchs7] and [Reference Feldvoss8].
Given
$\varphi\in C^2(\mathfrak g)$, a mapping
$\omega:\mathfrak g\to\mathbb F$ is
$\varphi$-compatible if for all
$g,h\in\mathfrak g$ and all
$a\in\mathbb F$
$\omega(a g)=a^p \omega (g)$ and
\begin{equation}
\omega(g+h)=\omega(g)+\omega(h) + \sum_{\substack{g_i={\scriptsize \textit{g}\
or\ \textit{h}}\\ g_1=g, g_2=h}}
\frac{1}{\#(g)}\varphi([g_1,g_2,g_3,\dots,g_{p-1}]\wedge g_p)
\end{equation} where
$\#(g)$ is the number of factors
$g_i$ equal to
$g$. We define
to be the space of Frobenius homomorphisms from
$\mathfrak g$ to
$\mathbb F$. A mapping
$\omega:\mathfrak g\to\mathbb F$ is
$0$-compatible if and only if
$\omega\in {\rm Hom}_{\rm Fr}(\mathfrak g,\mathbb F)$.
Given
$\varphi\in C^2(\mathfrak g)$, we can define a mapping
$\omega: \mathfrak{g} \to \mathbb F $ that is
$\varphi$-compatible by assigning the values of
$\omega$ arbitrarily on a basis for
$\mathfrak{g}$ and use (1) to extend it to
$\mathfrak{g}$ and this
$\omega$ is unique because its values are completely determined by
$\omega(e_i)$ and
$\varphi$ (c.f. [Reference Evans and Fialowski5]). In particular, given
$\varphi$, we can define
$\tilde\varphi(e_i)=0$ for all
$i$ and use (1) to determine a unique
$\varphi$-compatible mapping
$\widetilde\varphi:\mathfrak g\to\mathbb F$. Note that, in general,
$\widetilde\varphi\ne 0$ but
$\widetilde\varphi (0)=0$. Moreover, if
$\varphi_1,\varphi_2\in C^2(\mathfrak g)$ and
$a\in\mathbb F$, then
$\widetilde{(a\varphi_1+\varphi_2)} = a\widetilde\varphi_1 + \widetilde\varphi_2$.
Given
$\zeta\in C^3(\mathfrak g)$, a mapping
$\eta:\mathfrak g\times \mathfrak g\to\mathbb F$ is
$\zeta$-compatible if for all
$a\in\mathbb F$ and all
$g,h,h_1,h_2\in\mathfrak g$,
$\eta(\cdot,h)$ is linear in the first coordinate,
$\eta(g,a h)=a^p\eta(g,h)$ and
\begin{align*}
\eta(g,h_1+h_2) &=
\eta(g,h_1)+\eta(g,h_2)-\nonumber \\
& \sum_{\substack{l_1,\dots,l_p=1 {\rm or}2\\ l_1=1,
l_2=2}}\frac{1}{\#\{l_i=1\}}\zeta (g\wedge
[h_{l_1},\cdots,h_{l_{p-1}}]\wedge h_{l_{p}}).
\end{align*} The restricted cochain spaces are defined as
$C^0_*(\mathfrak g)=C^0 (\mathfrak g)$,
$C^1_*(\mathfrak g)=C^1 (\mathfrak g)$,
For
$1\le i\le n$, define
$\overline e^i:\mathfrak g\to\mathbb F$ by
$\overline e^i \left(\sum_{j=1}^n a_j e_j\right ) = a_i^p.$ The set
$\{\overline e^i\ |\ 1\le i\le n\}$ is a basis for
${\rm Hom}_{\rm Fr}(\mathfrak g,\mathbb F)$. Since
\begin{equation*}\dim C^2_*(\mathfrak g) = \binom{\dim\mathfrak g+1}{2}=\binom{\dim\mathfrak g}{2}+\dim\mathfrak g,\end{equation*}we have that
\begin{equation*}
\{(e^{i,j},\widetilde{e^{i,j}})\ |\ 1\le i \lt j\le n\} \cup \{(0,\overline e^i)\ |\ 1\le i\le n\}
\end{equation*} is a basis for
$C^2_*(\mathfrak g)$. We will use this basis in the following computations.
Define
$d_*^0=d^0$. For
$\psi\in C^1_*(\mathfrak g)$, define the mapping
${\rm ind}^1(\psi):\mathfrak g\to\mathbb F$ by
${\rm ind}^1(\psi)(g)=\psi(g^{[p]})$. The mapping
${\rm ind}^1(\psi)$ is
$d^1(\psi)$-compatible for all
$\psi\in C^1_*(\mathfrak g)$, and the differential
$d^1_*:C^1_*(\mathfrak g)\to C^2_*(\mathfrak g)$ is defined by
\begin{equation}
d^1_*(\psi) = (d^1(\psi),{\rm ind}^1(\psi)).
\end{equation} For
$(\varphi,\omega)\in C^2_*(\mathfrak g)$, define the mapping
${\rm ind}^2(\varphi,\omega):\mathfrak g \times\mathfrak g \to\mathbb F$ by the formula
\begin{equation*}{\rm ind}^2(\varphi,\omega)(g,h)=\varphi(g\wedge h^{[p]})-\varphi([g,\underbrace{h,\ldots,h}_{p-1}],h).\end{equation*} The mapping
${\rm ind}^2(\varphi,\omega)$ is
$d^2(\varphi)$-compatible for all
$\varphi\in C^2(\mathfrak g)$, and the differential
$d^2_*:C^2_*(\mathfrak g)\to C^3_*(\mathfrak g)$ is defined by
\begin{equation}
d^2_*(\varphi,\omega) =
(d^2(\varphi),{\rm ind}^2(\varphi,\omega)).
\end{equation} These mappings
$d_*^q$ satisfy
$d_*^{q}d_*^{q-1}=0$ and we define
Note that if
$\omega_1$ and
$\omega_2$ are both
$\varphi$-compatible, then
${\rm ind}^2(\varphi, \omega_1)={\rm ind}^2(\varphi, \omega_2)$.
Lemma 2.1 ([Reference Evans, Fialowski and Yang6])
If
$(\varphi,\omega)\in C^2_*(\mathfrak g)$ and
$\varphi=d^1(\psi)$ with
$\psi\in C^1 (\mathfrak g)$, then
$(\varphi,{\rm ind}^1(\psi))\in C^2_*(\mathfrak g)$ and
${\rm ind}^2(\varphi,\omega)={\rm ind}^2(\varphi,{\rm ind}^1(\psi))$.
If
$\mathfrak g$ is a restricted Lie algebra, there is a six-term exact sequence [3.2, Reference Viviani18] that relates the ordinary and restricted 1- and 2-cohomology spaces:
\begin{equation}
\begin{aligned}
0&\longrightarrow H^1_*(\mathfrak g)\longrightarrow H^1(\mathfrak g)\longrightarrow {\rm Hom}_{\rm Fr}(\mathfrak g,\mathbb F) \longrightarrow H^2_*(\mathfrak g)\longrightarrow H^2(\mathfrak g)\longrightarrow\\
&\stackrel{\Delta}{\longrightarrow} {\rm Hom}_{\rm
Fr}(\mathfrak g,H^1(\mathfrak g)).
\end{aligned}
\end{equation}The mapping
in (4) is given explicitly by
\begin{equation}\Delta_\varphi(g)\cdot h =\varphi(g\wedge[h,\underbrace{g,\ldots,g}_{p-1}])-\varphi(g^{[p]}\wedge
h)
\end{equation} where
$g,h\in\mathfrak g$.
3. Ordinary cohomology
$H^q({\mathfrak{h}}_m^{\lambda})$ for
$q=1,2$
Suppose
$I$ is an ideal of a Lie algebra
$\mathfrak{g}$. Then there is a convergent spectral sequence called the Hochschild–Serre spectral sequence [Reference Hochschild and Serre12] such that
Here we associate the cohomology of a twisted Heisenberg Lie algebra with the cohomology of its Heisenberg Lie algebra by means of Hochschild–Serre spectral sequences. Regarding
${\mathfrak{h}}_{m}$ as an ideal of
${\mathfrak{h}}_{m}^{\lambda}$, we have
So we have
$E_{\infty}^{r,s}=E_{2}^{r,s}=0$ for
$r\geq 2$ and
$E_{\infty}^{0,s}=E_{2}^{0,s}$,
$E_{\infty}^{1,s}=E_{2}^{1,s}$. Then we have the following lemma.
Lemma 3.1.
$H^{k}({\mathfrak{h}}_{m}^{\lambda})=
H^{k}({\mathfrak{h}}_{m})^{\mathbb F e_{2m+2}}\oplus H^{1}(\mathbb F e_{2m+2},H^{k-1}({\mathfrak{h}}_{m}))$ for
$k\geq 0$.
Proof. For
$k\geq 0$, we have
\begin{eqnarray*}
H^{k}({\mathfrak{h}}_{m}^{\lambda})&=&\bigoplus_{r+s=k}E^{r,s}_{\infty}\\
&=&E_{2}^{0,k}\oplus E_{2}^{1,k-1}\\
&=&H^{0}(\mathbb F e_{2m+2},H^{k}({\mathfrak{h}}_{m}))\oplus H^{1}(\mathbb F e_{2m+2},H^{k-1}({\mathfrak{h}}_{m}))\\
&=&H^{k}({\mathfrak{h}}_{m})^{\mathbb F e_{2m+2}}\oplus H^{1}(\mathbb F e_{2m+2},H^{k-1}({\mathfrak{h}}_{m})).
\end{eqnarray*} The cohomology of the Heisenberg Lie algebra
${\mathfrak{h}}_m$ with trivial coefficients was first studied by Santharoubane in [Reference Santharoubane14] (in characteristic
$0$) and by Cairns and Jambor in [Reference Cairns and Jambor2] (in characteristic
$p \gt 0$).
Lemma 3.2 ([Reference Cairns and Jambor2, Reference Santharoubane14])
For
$q\leq m$,
(1) if
$\mathrm{char}\ \mathbb F=0$,
\begin{equation*}H^{q}({\mathfrak{h}}_m)=\bigwedge^{q}({\mathfrak{h}}_m/\mathbb F e_{2m+1})^{\ast}/(d
e^{2m+1}\wedge\bigwedge^{q-2}({\mathfrak{h}}_m/\mathbb F e_{2m+1})^{\ast});\end{equation*} (2) if
$\mathrm{char}\ \mathbb F=p \gt 0$,
\begin{align*}
H^{q}({\mathfrak{h}}_m)&=\bigwedge^{q}({\mathfrak{h}}_m/\mathbb F
e_{2m+1})^{\ast}/(de^{2m+1}\wedge\bigwedge^{q-2}({\mathfrak{h}}_m/\mathbb F
e_{2m+1})^{\ast}) \\
&\oplus((d e^{2m+1})^{p-1}\wedge
e^{2m+1}\wedge\bigwedge^{q-2p+1}({\mathfrak{h}}_m/\mathbb F
e_{2m+1})^{\ast})
\end{align*} where
$d e^{2m+1}=\sum^{m}_{i=1}e^{i,m+i}$.
If
$q=1$ or
$q=2$, then the second term in part (2) of Reference HochschildLemma 11.11 vanishes, so the cohomology spaces do not depend on
$p$. In particular
and
\begin{equation}
H^2({\mathfrak{h}}_m)=\bigwedge^{2}({\mathfrak{h}}_m/\mathbb F e_{2m+1})^{\ast}/(d
e^{2m+1}).
\end{equation} For
$1\leq i \lt j\leq m$, put
$\delta_{\lambda_i=\pm\lambda_j}=1$ when
$\lambda_i=\lambda_j$ or
$\lambda_i=-\lambda_j$ (mod
$p$) and
$\delta_{\lambda_i=\pm\lambda_j}=0$, otherwise. By Lemma 3.1 and Eqs. (6), (7), we get the following theorem.
Theorem 3.3. For
$m\ge 1$ and
$\lambda\in(\mathbb F^{\times})^{m}$,
(1)
$H^{1}({\mathfrak{h}}_{m}^{\lambda})= \mathbb{F}e^{2m+2}$;
(2) a basis of
$H^{2}({\mathfrak{h}}_{m}^{\lambda})$ consists of the classes of the cocycles
\begin{eqnarray*}
&&\{\delta_{\lambda_i=\pm\lambda_j}\left(e^{i,j}-\lambda_i\lambda_j^{-1}e^{m+i,m+j}\right), \delta_{\lambda_i=\pm\lambda_j}\left(e^{i,m+j}-\lambda_i\lambda_j^{-1}e^{m+i,j}\right)\\
&&\mid 1\leq i \lt j\leq m\} \bigcup\{e^{i,m+i}\mid 1\leq i\leq m-1\}.
\end{eqnarray*}In particular,
\begin{equation*}\dim H^{2}({\mathfrak{h}}_{m}^{\lambda})=2Card\{(i,j)\mid \lambda_i=\pm \lambda_j,1\leq i \lt j\leq m\}+m-1.\end{equation*}4. Restricted structures on the modular Lie algebra
$\mathfrak{h}^{\lambda}_m$
In this section, we determine all restricted structures on the twisted Heisenberg Lie algebra over a field of prime characteristic.
For
$e_{2m+2}$, we have
\begin{equation*}(\mathop{\rm ad} e_{2m+2})^{p}: \left\{
\begin{array}{ll}
e_i\mapsto \lambda^{2}_i e_{i} , e_{m+i}\mapsto \lambda^{2}_i e_{m+i}, & p=2;\\
e_i\mapsto \lambda^{p}_i e_{m+i}, e_{m+i}\mapsto \lambda^{p}_i e_i , & p \gt 2;
\end{array}
\right.\end{equation*} where
$1\leq i\leq m$. So
$(\mathrm{ad}\ e_{2m+2})^{p}$ is an inner derivation if and only if
$p \gt 2$ and
$\lambda^{p-1}_1=\cdots=\lambda^{p-1}_m$. In this case, we denote by
$|\lambda|=\lambda^{p-1}_i$ for
$1\leq i\leq m$ and we have
$(\mathrm{ad}\ e_{2m+2})^{p}=\mathrm{ad}\ (|\lambda|e_{2m+2})$. When
$p \gt 2$, we have
$(\mathrm{ad}\ e_i)^{p}=0$ for
$1\leq i\leq 2m+1$. By Jacobson’s theorem [Reference Jacobson13], we have the following theorem:
Theorem 4.1. For
$m\ge 1$ and
$\lambda\in(\mathbb F^{\times})^{m}$,
$\mathfrak{h}^{\lambda}_m$ is restrictable if and only if
$p \gt 2$ and
$\lambda^{p-1}_1=\cdots=\lambda^{p-1}_m$.
From now on, we assume
$p \gt 2$ and
$\lambda^{p-1}_1=\cdots=\lambda^{p-1}_m=|\lambda|$. We can use Jacobson’s theorem [Reference Jacobson13] and define a restricted Lie algebra
$[p]$-operator on
$\mathfrak{h}^{\lambda}_m$ by choosing
$e_i^{[p]}\in \mathfrak{h}^{\lambda}_m$ so that
$\mathop{\rm ad} e_{2m+2}^{[p]}=(\mathrm{ad}\ e_{2m+2})^{p}=\mathrm{ad}\ (|\lambda|e_{2m+2})$ and
$\mathop{\rm ad} e_i^{[p]}=(\mathop{\rm ad} e_i)^p=0$ for all
$i$. For each
$i$, we choose a scalar
$\mu_i\in \mathbb F$ and set
\begin{equation*}e_i^{[p]}=\mu_i e_{2m+1},\quad
e_{2m+2}^{[p]}=|\lambda|e_{2m+2}+\mu_{2m+2} e_{2m+1}.\end{equation*} We let
$\mu=(\mu_1,\dots, \mu_{2m+2})$ and denote the corresponding restricted Lie algebra
$\mathfrak{h}_m^{\lambda,\mu}$.
Since
$[[[{\mathfrak{h}}_m^{\lambda,\mu},{\mathfrak{h}}_m^{\lambda,\mu}],{\mathfrak{h}}_m],{\mathfrak{h}}^{\lambda,\mu}_m]=0$, by the definition of restricted Lie algebras, we have
$s_i(g,h)=0$ for
$i\geq 3$, and
\begin{equation*}s_1(g,h)=[g,\underbrace{h,\ldots,h}_{p-1}],\quad 2s_2(g,h)=[g,\underbrace{h,\ldots,h}_{p-2},g]\end{equation*} for all
$g\in \mathfrak{h}_{m}$ and
$h\in \mathfrak{h}_{m}^{\lambda,\mu}$. From this we get that if
$g=\sum^{2m+2}_{i=1} a_i e_i\in \mathfrak{h}_m^{\lambda,\mu}$, then
\begin{equation}
\begin{aligned}
g^{[p]}&=a_{2m+2}^{p-1}|\lambda|\sum_{i=1}^{2m}a_ie_i+a_{2m+2}^{p}|\lambda|e_{2m+2}\\
&+\left(\sum^{2m+2}_{i=1}a^{p}_i\mu_i+2^{-1}a_{2m+2}^{p-2}\sum_{i=1}^{m}\lambda_{i}^{p-2}(a_i^{2}-a_{m+i}^{2})\right) e_{2m+1}.
\end{aligned}
\end{equation}Remark 4.2. For
$g=\sum^{2m+1}_{i=1} a_i e_i\in \mathfrak{h}_m$, (8) gives the restricted structures on the Heisenberg Lie algebra
$\mathfrak{h}_m$ [Reference Evans, Fialowski and Yang6]:
\begin{equation*}g^{[p]}=\left(\sum^{2m+1}_{i=1}a^{p}_i\mu_i\right) e_{2m+1}.\end{equation*} Therefore,
$\mathfrak{h}_m$ is a restricted ideal of
$ \mathfrak{h}^{\lambda,\mu}_m$.
5. Restricted cohomology
$H^q_*({\mathfrak{h}}_m^{\lambda,\mu})$ for
$q=1,2$
By definition,
$[{\mathfrak{h}}_m^{\lambda,\mu},{\mathfrak{h}}_m^{\lambda,\mu}]= {\mathfrak{h}}_{m}$, and the
$[p]$-operator formula (8) imply
\begin{equation*} [{\mathfrak{h}}_m^{\lambda,\mu},{\mathfrak{h}}_m^{\lambda,\mu}]+\langle ({\mathfrak{h}}_m^{\lambda,\mu})^{[p]}\rangle_{\mathbb{F}}=
{\mathfrak{h}}_m^{\lambda,\mu}.\end{equation*} For any restricted Lie algebra
$\mathfrak g$, [Reference Hochschild11.11, Theorem 2.1] states that
\begin{equation*}
H^1_*(\mathfrak g)=(\mathfrak g/([\mathfrak g,\mathfrak g]+\langle \mathfrak g^{[p]}\rangle_{\mathbb{F}}))^*.\end{equation*} It follows that
$H^1_*({\mathfrak{h}}_m^{\lambda,\mu})=0$, and the six-term exact sequence (4) decouples to the exact sequence
\begin{equation*}
\begin{aligned}
0&\longrightarrow H^1({\mathfrak{h}}_m^{\lambda})\longrightarrow{\rm Hom}_{\rm Fr}({\mathfrak{h}}_m^{\lambda,\mu},\mathbb F) \longrightarrow H^2_*({\mathfrak{h}}_m^{\lambda,\mu})\longrightarrow H^2({\mathfrak{h}}_m^{\lambda,\mu})\longrightarrow \\
&\stackrel{\Delta}{\longrightarrow} {\rm Hom}_{\rm Fr}({\mathfrak{h}}_m^{\lambda,\mu},\mathbb F e^{2m+2}).
\end{aligned}
\end{equation*} If we let
$g=\sum a_ie_i\in{\mathfrak{h}}_m^{\lambda,\mu}$ and
$\omega:{\mathfrak{h}}_m^{\lambda,\mu}\to \mathbb F$ be any
$\varphi$-compatible mapping to a 2-cocycle
$\varphi$, then Theorem 3.3 (2) and (5) give
\begin{eqnarray*}
\Delta_\varphi(g)\cdot e_{2m+2}
&=&\varphi(g\wedge [e_{2m+2},\underbrace{g,\ldots,g}_{p-1}])-\varphi(g^{[p]}\wedge
e_{2m+2})\\
&=&-a_{2m+2}^{p-2}|\lambda|\varphi(g,g)\\
&=&0.
\end{eqnarray*} Then
$\Delta=0$. By Theorem 3.3 (1), we have the splitting exact sequence
where the mapping
$ {\rm Hom}_{\rm Fr}({\mathfrak{h}}_m^{\lambda,\mu},\mathbb F)/\mathbb F \overline e^{2m+2} \to H^2_*({\mathfrak{h}}_m^{\lambda,\mu}) $ sends
$\omega$ to the class of
$ (0,\omega)$ and the image is a
$(2m+1)$-dimensional subspace of
$ H^2_*({\mathfrak{h}}_m^{\lambda,\mu})$ spanned by the classes
$(0,\overline
e^i)$ for
$1\leq i\leq 2m+1$. It follows that
Theorem 3.3 (2) implies the following theorem:
Theorem 5.1. For
$m\ge 1$,
$\lambda\in(\mathbb F^{\times})^{m}$ and any form
$\mu:{\mathfrak{h}}^{\lambda,\mu}_m\to\mathbb F$,
\begin{equation*}\dim H^2_*({\mathfrak{h}}_m^{\lambda,\mu})=2Card\{(i,j)\mid \lambda_i=\pm \lambda_j,1\leq i \lt j\leq m\}+3m\end{equation*}and a basis consists of the classes of the cocycles
\begin{equation*} \{\delta_{\lambda_i=\pm\lambda_j}(e^{i,j}-\lambda_i\lambda^{-1}_je^{m+i,m+j},
\widetilde{e^{i,j}}-\lambda_i\lambda^{-1}_j\widetilde{e^{m+i,m+j}}),
\end{equation*}
\begin{equation*} \delta_{\lambda_i=\pm\lambda_j} (e^{i,m+j}-\lambda_i \lambda^{-1}_je^{m+i,j}, \widetilde{e^{i,m+j}}-\lambda_i\lambda^{-1}_j\widetilde{e^{m+i,j}})
\mid1\leq i \lt j\leq m\} \end{equation*}
\begin{equation*}\bigcup \{(e^{i,m+i},\widetilde{e^{i,m+i}})\ |\ 1\le i\le m-1\}\bigcup \{(0,\overline e^i)\ |\ 1\le i\le 2m+1\}.\end{equation*}6. Restricted one-dimensional central extensions of
${\mathfrak{h}}_m^{\lambda,\mu}$
Similar to the case of ordinary Lie algebras [[Reference Fuchs9], Chapter 1, Section 4.6], restricted one-dimensional central extensions of a restricted Lie algebra
$\mathfrak g$ are parameterized by the restricted cohomology group
$H^2_*(\mathfrak g)$ [[Reference Hochschild11.11], Theorem 3.3]. If
$(\varphi,\omega) \in C^2_{*}(\mathfrak g)$, then the corresponding restricted one-dimensional central extension
$\mathfrak{G}=\mathfrak g\oplus\mathbb F c$ has Lie bracket and
$[p]$-operator defined by
\begin{equation}
\begin{aligned}
{[g,h]}_\mathfrak{G}&={[g,h]}_{\mathfrak g} + \varphi(g\wedge h)c\\
[g,c]_\mathfrak{G}&=0\\
g^{[p]_\mathfrak{G}}&=g^{[p]_{\mathfrak g}} + \omega(g)c\\
c^{[p]_\mathfrak{G}}&=0
\end{aligned}
\end{equation} where
$[\cdot,\cdot]_{\mathfrak g}$ and
$\cdot^{{[p]}_{\mathfrak g}}$ denote the Lie bracket and
$[p]$-operator in
$\mathfrak g$, respectively ([Reference Evans and Fuchs7], equations (26) and (27)). We can the equations (9) together with Theorem 5.1 to explicitly describe the restricted one-dimensional central extensions of
${\mathfrak{h}}_m^{\lambda,\mu}$. Let
$g=\sum a_ie_i$ and
$h=\sum b_ie_i$ denote two arbitrary elements of
${\mathfrak{h}}_m^{\lambda,\mu}$. For
$1\leq s \lt t\leq 2m$, the equation (1) gives
\begin{equation}
\begin{aligned}
\widetilde{e^{s,t}}
(g)&= \widetilde{e^{s,t}}
\left(\sum_{i=1}^{2m+1}a_ie_i+a_{2m+2}e_{2m+2}\right)\\
&=2^{-1}a_{2m+2}^{p-2}\sum_{i,j=1}^{2m}a_ia_je^{s,t}([e_i,\underbrace{e_{2m+2},\ldots,e_{2m+2}}_{p-2}],e_j)\\
&= -2^{-1}a_{2m+2}^{p-2}
\bigg(
\sum_{i=1}^{m}\sum_{j=1}^{2m}\lambda_{i}^{p-2}a_i a_j
e^{s,t}(e_{m+i},e_j) \\
&+
\sum_{i=1}^{m}\sum_{j=1}^{2m}\lambda_{i}^{p-2}a_{m+i}a_je^{s,t}(e_{i},e_j) \bigg).
\end{aligned}
\end{equation} If
$1\le i \lt j\le m$ such that
$\lambda_i=\lambda_j$ or
$\lambda_i=-\lambda_j$, and
$\mathfrak{H}_{ij}={\mathfrak{h}}_m^{\lambda.\mu}\oplus \mathbb F c$ denotes the one-dimensional restricted central extension of
${\mathfrak{h}}_m^{\lambda,\mu}$ determined by the cohomology class of the restricted cocycle
\begin{equation*} \left(e^{i,j}-\lambda_i\lambda^{-1}_je^{m+i,m+j},
\widetilde{e^{i,j}}-\lambda_i\lambda^{-1}_j\widetilde{e^{m+i,m+j}}\right),\end{equation*} then (9) and (10) give the bracket and
$[p]$-operator in
$\mathfrak{H}_{ij}$:
\begin{equation*}
\begin{aligned}
{[g,h]}_{\mathfrak{H}_{ij}}& ={[g,h]}_{{\mathfrak{h}}_m^{\lambda,\mu}}+(a_ib_j-a_jb_i-\lambda_i\lambda^{-1}_ja_{m+i}b_{m+j}+\lambda_i\lambda^{-1}_ja_{m+j}b_{m+i})c;\\
g^{[p]_{\mathfrak{H}_{ij}}} & = g^{[p]_{{\mathfrak{h}}_m^{\lambda,\mu}}}-2^{-1}a_{2m+2}^{p-2}(\lambda^{p-2}_{i} a_{m+i}a_{j}-\lambda_{j}^{p-2}a_{m+j}a_{i}-|\lambda|\lambda_j^{-1}a_i a_{m+j}\\
&+\lambda_i\lambda_j^{p-3}a_ja_{m+i}) c.
\end{aligned}
\end{equation*} If
$1\le i \lt j\le m$ such that
$\lambda_i=\lambda_j$ or
$\lambda_i=-\lambda_j$, and
$\mathfrak{H}_{i,m+j}={\mathfrak{h}}_m^{\lambda.\mu}\oplus \mathbb F c$ denotes the one-dimensional restricted central extension of
${\mathfrak{h}}_m^{\lambda,\mu}$ determined by the cohomology class of the restricted cocycle
\begin{equation*}(e^{i,m+j}-\lambda_i \lambda^{-1}_je^{m+i,j}, \widetilde{e^{i,m+j}}-\lambda_i\lambda^{-1}_j\widetilde{e^{m+i,j}}),\end{equation*} then (9) and (10) give the bracket and
$[p]$-operator in
$\mathfrak{H}_{i,m+j}$:
\begin{equation*}
\begin{aligned}
{[g,h]}_{\mathfrak{H}_{i,m+j}} & ={[g,h]}_{{\mathfrak{h}}_m^{\lambda,\mu}}+(a_ib_{m+j}-a_{m+j}b_i-\lambda_i\lambda^{-1}_ja_{m+i}b_{j}+\lambda_i\lambda^{-1}_ja_{j}b_{m+i})c;\\
g^{[p]_{\mathfrak{H}_{i,m+j}}} & = g^{[p]_{{\mathfrak{h}}_m^{\lambda,\mu}}}-2^{-1}a_{2m+2}^{p-2}(\lambda^{p-2}_{i} a_{m+i}a_{m+j}-\lambda_{j}^{p-2}a_{j}a_{i}-|\lambda|\lambda_j^{-1}a_i a_{j}\\
&+\lambda_i\lambda_j^{p-3}a_{m+j}a_{m+i}) c.
\end{aligned}
\end{equation*} If
$1\le i\le m-1$ and
$\mathfrak{H}_{i,m+i}={\mathfrak{h}}_m^{\lambda.\mu}\oplus \mathbb F c$ denotes the one-dimensional restricted central extension of
${\mathfrak{h}}_m^{\lambda,\mu}$ determined by the cohomology class of the restricted cocycle
$(e^{i,m+i},\widetilde{e^{i,m+i}})$, then (9) and (10) give the bracket and
$[p]$-operator in
$\mathfrak{H}_{i,m+i}$:
\begin{equation*}
\begin{aligned}
{[g,h]}_{\mathfrak{H}_{i,m+i}} & ={[g,h]}_{{\mathfrak{h}}_m^{\lambda,\mu}}+(a_ib_{m+i}-a_{m+i}b_i)c;\\
g^{[p]_{\mathfrak{H}_{i,m+i}}} & = g^{[p]_{{\mathfrak{h}}_m^{\lambda,\mu}}}-2^{-1}a_{2m+2}^{p-2}\lambda^{p-2}_{i} (a^{2}_{m+i}-a_{i}^{2}) c.
\end{aligned}
\end{equation*} If
$1\le i\le 2m+1$ and
$\mathfrak{H}_i={\mathfrak{h}}_m^{\lambda.\mu}\oplus \mathbb F c$ denotes the one-dimensional restricted central extension of
${\mathfrak{h}}_m^{\lambda,\mu}$ determined by the cohomology class of the restricted cocycle
$(0,\overline e^i)$, then (9) gives the bracket and
$[p]$-operator in
$\mathfrak{H}_i$:
\begin{equation*}
\begin{aligned}
{[g,h]}_{\mathfrak{H}_i} & ={[g,h]}_{{\mathfrak{h}}_m^{\lambda,\mu}};\\
g^{[p]_{\mathfrak{H}_i}} & = g^{[p]_{{\mathfrak{h}}_m^{\lambda,\mu}}}+ a_i^p c.
\end{aligned}
\end{equation*} The central extensions
$\mathfrak{H}_i$ form a basis for the
$(2m+1)$-dimensional space of restricted one-dimensional central extensions that split as ordinary Lie algebra extensions (c.f. [Reference Evans and Fialowski5]).
Acknowledgements
The author thanks the referee for the suggestions, which improved the presentation. The work is supported by the Talent Project of the Tianchi Doctoral Program in Xinjiang Uygur Autonomous Region (Grant No. 5105250184d) and Natural Science Foundation of Xinjiang Uygur Autonomous Region (Grant No. 2025D01C279).
Conflict of interest
The authors have no conflict of interest to declare that is relevant to this article.
Data availability
Data sharing is not applicable to this article as no datasets were generated or analyzed in this study.









