This paper was converted on www.awesomepapers.org from LaTeX by an anonymous user.
Want to know more? Visit the Converter page.

Gerstenhaber bracket on Hopf algebra and Hochschild cohomologies

Tekin Karadağ
Abstract

We calculate the Gerstenhaber bracket on Hopf algebra and Hochschild cohomologies of the Taft algebra TpT_{p} for any integer p>2p>2 which is a nonquasi-triangular Hopf algebra. We show that the bracket is indeed zero on Hopf algebra cohomology of TpT_{p}, as in all known quasi-triangular Hopf algebras. This example is the first known bracket computation for a nonquasi-triangular algebra. Also, we find a general formula for the bracket on Hopf algebra cohomology of any Hopf algebra with bijective antipode on the bar resolution that is reminiscent of Gerstenhaber’s original formula for Hochschild cohomology.

Key words and phrases: Hochschild cohomology, Hopf algebra cohomology, Gerstenhaber bracket, Taft algebraPartially supported by NSF grant 1665286.

1 Introduction

Gerstenhaber brackets were originally defined on Hochschild cohomology by M. Gerstenhaber himself [3, Section 1.1]. In 2002, A. Farinati and A. Solotar showed that for any Hopf algebra AA, Hopf algebra cohomology H(A):=ExtA(k,k){}^{*}(A):=\operatorname{Ext}^{*}_{A}(k,k) is a Gerstenhaber algebra [2]. Hence, we can define a Gerstenhaber bracket on Hopf algebra cohomology. In the same year, R. Taillefer used a different approach and found a bracket on Hopf algebra cohomology [11] which is equivalent to the bracket constructed by A. Farinati and A. Solotar. The category of AA-modules and the category of AeA^{e}-modules are examples of strong exact monoidal categories. In 2016, Reiner Hermann [5, Theorem 6.3.12, Corollary 6.3.15] proved that if the strong exact monoidal category is lax braided, then the bracket is constantly zero. Therefore, the Gerstenhaber bracket on the Hopf algebra cohomology of a quasi-triangular Hopf algebra is trivial. However, we do not know the bracket structure for a nonquasi-triangular Hopf algebra. Taft algebras are nice examples of nonquasi-triangular Hopf algebras. In this paper, we show that the Gerstenhaber bracket on the Hochschild cohomology of a Taft Algebra is nontrivial. However, the bracket structure on Hopf algebra cohomology of a Taft algebra is constantly zero. Also, we take the Gerstenhaber bracket formula on Hochschild comology and find a general formula for Gerstenhaber bracket on Hopf algebra cohomology.

We start by giving some basic definitions and some tools to calculate the bracket on Hochschild cohomology in Section 2. Then, we compute the Gerstenhaber bracket on the Hochschild cohomology of A=k[x]/(xp)A=k[x]/(x^{p}) where the field kk has characteristic 0 and the integer p>2p>2 in Section 3. We use the technique introduced by C. Negron and S. Witherspoon [7] and note that they computed the bracket on Hochschild cohomology of AA for the case that kk has positive characteristic pp [7, Section 5].

In Section 4, we compute the Gerstenhaber bracket for the Taft algebra TpT_{p} which is a nonquasi-triangular Hopf algebra. We use a similar technique as in [7] to calculate the bracket on Hochschild cohomology of TpT_{p}. It is also known that the Hopf algebra cohomology of any Hopf algebra with a bijective antipode can be embedded in the Hochschild cohomology of the algebra [14, Theorem 9.4.5 and Corollary 9.4.7]. Since all finite dimensional Hopf algebras (also most of known infinite dimensional Hopf algebras) have bijective antipode, we can embed the Hopf algebra cohomology of TpT_{p} into the Hochschild cohomology of TpT_{p}. Then, we use this explicit embedding and find the bracket on the Hopf algebra cohomology of TpT_{p}. As a result of our calculation, we obtain that the bracket on Hopf algebra cohomology of TpT_{p} is also trivial.

In the last section, we derive a general expression for the bracket on Hopf algebra cohomology of any Hopf algebra AA with bijective antipode. We first consider a specific resolution that agrees with the bar resolution of AA and find a bracket formula for it. Then, we use the composition of various isomorphisms and an embedding from Hopf algebra cohomology into Hochschild cohomology in order to discover the bracket formula on Hopf algebra cohomology.

2 Gerstenhaber Bracket on Hochschild Cohomology

Let kk be a field, AA be a kk-algebra, and Ae=AkAopA^{e}=A\otimes_{k}A^{op} where AopA^{op} is the opposite algebra with reverse multiplication. For simplicity, we write \otimes instead of k\otimes_{k}. The following resolution B(A)B(A) is a free resolution of the AeA^{e}-module AA, called the bar resolution,

B(A):d3A4d2A3d1A2πA0,B(A):\cdots\stackrel{{\scriptstyle d_{3}}}{{\longrightarrow}}A^{\otimes 4}\stackrel{{\scriptstyle d_{2}}}{{\longrightarrow}}A^{\otimes 3}\stackrel{{\scriptstyle d_{1}}}{{\longrightarrow}}A^{\otimes 2}\stackrel{{\scriptstyle\pi}}{{\longrightarrow}}A\longrightarrow 0, (2.1)

where

dn(a0a1an+1)=i=0n(1)ia0a1aiai+1an+1d_{n}(a_{0}\otimes a_{1}\otimes\cdots\otimes a_{n+1})=\sum_{i=0}^{n}(-1)^{i}a_{0}\otimes a_{1}\otimes\cdots\otimes a_{i}a_{i+1}\otimes\cdots\otimes a_{n+1}

and π\pi is multiplication.

Consider the following complex that is derived by applying Hom(,A)Ae{}_{A^{e}}(-,A) to the bar resolution B(A)B(A)

0HomAe(A2,A)d1HomAe(A3,A)d2HomAe(A4,A)d30{\longrightarrow}\text{Hom}_{A^{e}}(A^{\otimes 2},A)\stackrel{{\scriptstyle d^{*}_{1}}}{{\longrightarrow}}\text{Hom}_{A^{e}}(A^{\otimes 3},A)\stackrel{{\scriptstyle d^{*}_{2}}}{{\longrightarrow}}\text{Hom}_{A^{e}}(A^{\otimes 4},A)\stackrel{{\scriptstyle d^{*}_{3}}}{{\longrightarrow}}\cdots (2.2)

where dn(f)=fdnd_{n}^{*}(f)=fd_{n}. The Hochshild cohomology of the algebra AA is the cohomology of the cochain complex (2.1), i.e.

HH(A,A)=n0ExtAen(A,A).\text{HH}^{*}(A,A)=\bigoplus_{n\geq 0}\operatorname{Ext}_{A^{e}}^{n}(A,A).

We also define the Hopf algebra cohomology of the Hopf algebra AA over the field kk as

H(A,k)=n0ExtAn(k,k)\textup{H}^{*}(A,k)=\bigoplus_{n\geq 0}\textup{Ext}^{n}_{A}(k,k)

under the cup product.

Let fHomk(Am,A)f\in\text{Hom}_{k}(A^{\otimes m},A) and gHomk(An,A)g\in\text{Hom}_{k}(A^{\otimes n},A). The Hochschild cohomology of AA is an algebra with the following cup product and the Gerstenhaber bracket structures. The cup product fgf\smile g\inHom(A(m+n),A)k{}_{k}(A^{\otimes(m+n)},A) is defined by

(fg)(a1am+n):=(1)mnf(a1am)g(am+1am+n)(f\smile g)(a_{1}\otimes\cdots\otimes a_{m+n}):=(-1)^{mn}f(a_{1}\otimes\cdots\otimes a_{m})g(a_{m+1}\otimes\cdots a_{m+n})

for all a1,,am+nAa_{1},\cdots,a_{m+n}\in A, and the Gerstenhaber bracket [f,g][f,g] is an element of
Hom(A(m+n1),A)k{}_{k}(A^{\otimes(m+n-1)},A) given by

[f,g]:=fg(1)(m1)(n1)gf[f,g]:=f\circ g-(-1)^{(m-1)(n-1)}g\circ f

where the circle product fgf\circ g is

(fg)(a1am+n1):=\displaystyle(f\circ g)(a_{1}\otimes\cdots\otimes a_{m+n-1}):=
i=1m(1)(n1)(i1)f(a1ai1g(aiai+n1)ai+nam+n1)\displaystyle\sum_{i=1}^{m}(-1)^{(n-1)(i-1)}f(a_{1}\otimes\cdots a_{i-1}\otimes g(a_{i}\otimes\cdots a_{i+n-1})\otimes a_{i+n}\otimes\cdots\otimes a_{m+n-1})

for all a1,,am+n1Aa_{1},\cdots,a_{m+n-1}\in A. We note that these definitions directly come from the bar resolution.

There is an identity between cup product and bracket [3, Section 1]:

[fg,h]=[f,h]g+(1)|f|(|h|1)f[g,h],[f^{*}\smile g^{*},h^{*}]=[f^{*},h^{*}]\smile g^{*}+(-1)^{|f^{*}|(|h^{*}|-1)}f^{*}\smile[g^{*},h^{*}], (2.3)

where f,g,f^{*},g^{*}, and hh^{*} are the images (in Hochschild cohomology) of the cocyles f,gf,g, and hh, respectively.

Computing the bracket on the bar resolution is not an ideal method. Instead, we can use another resolution, 𝔸μA\mathbb{A}\stackrel{{\scriptstyle\mu}}{{\rightarrow}}A, satisfying the following hypotheses [7, (3.1) and Lemma 3.4.1]:

(a) 𝔸\mathbb{A} admits an embedding ι:𝔸B(A)\iota:\mathbb{A}\to B(A) of complexes of AA-bimodules for which the following diagram commutes

𝔸{\mathbb{A}}B(A){B(A)}A{A}ι\scriptstyle{\iota}

(b) The embedding ι\iota admits a section π:B𝔸\pi:B\to\mathbb{A}, i.e. an AeA^{e}-chain map π\pi with πι=id𝔸\pi\iota=id_{\mathbb{A}}.

(c) There is a diagonal map that satisfies Δ𝔸(2)=(πAπAπ)ΔB(A)(2)ι\Delta_{\mathbb{A}}^{(2)}=(\pi\otimes_{A}\pi\otimes_{A}\pi)\Delta^{(2)}_{B(A)}\iota where Δ(2)=(idΔ)Δ\Delta^{(2)}=(id\otimes\Delta)\Delta.

We give the following theorem which is the combination of [7, Theorem 3.2.5] and [7, Lemma 3.4.1] that allows us to use a different resolution for the bracket calculation.

Theorem 2.4.

Suppose 𝔸μA\mathbb{A}\stackrel{{\scriptstyle\mu}}{{\rightarrow}}A is a projective AA-bimodule resolution of AA that satisfies the hypotheses (a)-(c). Let ϕ:𝔸A𝔸𝔸\phi:\mathbb{A}\otimes_{A}\mathbb{A}\to\mathbb{A} be any contracting homotopy for the chain map F𝔸:𝔸A𝔸𝔸F_{\mathbb{A}}:\mathbb{A}\otimes_{A}\mathbb{A}\to\mathbb{A} defined by F𝔸:=(μAid𝔸id𝔸Aμ)F_{\mathbb{A}}:=(\mu\otimes_{A}id_{\mathbb{A}}-id_{\mathbb{A}}\otimes_{A}\mu), i.e.

d(ϕ):=d𝔸ϕ+ϕd𝔸A𝔸=F𝔸.d(\phi):=d_{\mathbb{A}}\phi+\phi d_{\mathbb{A}\otimes_{A}\mathbb{A}}=F_{\mathbb{A}}. (2.5)

Then for cocycles ff and gg in Hom(𝔸,A)Ae{}_{A^{e}}(\mathbb{A},A), the bracket given by

[f,g]ϕ=fϕg(1)(|f|1)(|g|1)gϕf[f,g]_{\phi}=f\circ_{\phi}g-(-1)^{(|f|-1)(|g|-1)}g\circ_{\phi}f (2.6)

where the circle product is

fϕg=fϕ(id𝔸AgAid𝔸)Δ(2)f\circ_{\phi}g=f\phi(id_{\mathbb{A}}\otimes_{A}g\otimes_{A}id_{\mathbb{A}})\Delta^{(2)} (2.7)

agrees with the Gerstenhaber bracket on cohomology.

In general, it is not easy to calculate the map ϕ\phi by the formula (2.5). We use alternative way to find ϕ\phi.

Let hh be any kk-linear contracting homotopy for the identity map on the extended complex 𝔸A0\mathbb{A}\to A\to 0 where 𝔸\mathbb{A} is free. A contracting homotopy ϕi:(𝔸A𝔸)i𝔸i+1\phi_{i}:(\mathbb{A}\otimes_{A}\mathbb{A})_{i}\longrightarrow\mathbb{A}_{i+1} in Theorem 2.4 is constructed by the following formula [7, Lemma 3.3.1]:

ϕi=hi((F𝔸)iϕi1d(𝔸A𝔸)i).\phi_{i}=h_{i}((F_{\mathbb{A}})_{i}-\phi_{i-1}d_{(\mathbb{A}\otimes_{A}\mathbb{A})_{i}}). (2.8)

3 Bracket on Hochschild cohomology of A=k[x]/(xp)A=k[x]/(x^{p})

Let A=k[x]/(xp)A=k[x]/(x^{p}) where kk is a field of characteristic 0 and p>2p>2 is an integer. We compute the Lie bracket on Hochschild cohomology of AA by Theorem 2.4. We work on a smaller resolution of AA than the bar resolution of AA. Consider the following AeA^{e}-module resolution of AA:

𝔸:v.Aeu.Aev.Aeu.AeπA0,\mathbb{A}:\cdots\stackrel{{\scriptstyle v.}}{{\longrightarrow}}A^{e}\stackrel{{\scriptstyle u.}}{{\longrightarrow}}A^{e}\stackrel{{\scriptstyle v.}}{{\longrightarrow}}A^{e}\stackrel{{\scriptstyle u.}}{{\longrightarrow}}A^{e}\stackrel{{\scriptstyle\pi}}{{\longrightarrow}}A\longrightarrow 0, (3.1)

where u=x11xu=x\otimes 1-1\otimes x, v=xp11+xp2x++xxp2+1xp1v=x^{p-1}\otimes 1+x^{p-2}\otimes x+\cdots+x\otimes x^{p-2}+1\otimes x^{p-1}, and π\pi is the multiplication.

The bracket on AA where kk is a field with positive characteristic, is calculated by C. Negron and S. Witherspoon [7, Section 5]. We adopt the contracting homotopy hh for the identity map from that calculation and obtain a new map hh for our setup. Let ξi\xi_{i} be the element 111\otimes 1 of 𝔸i\mathbb{A}_{i}. The following maps hn:𝔸n𝔸n+1h_{n}:\mathbb{A}_{n}\longrightarrow\mathbb{A}_{n+1} form a contracting homotopy for identity map, as we can see by direct calculation:

h1(xi)=ξ0xi,h0(xiξ0xj)=l=0i1xlξ1xi+j1l,h1(xiξ1xj)=δi,p1xjξ2,h2n(xiξ2nxj)=l=0j1xi+j1lξ2n+1xl (n2) ,h2n+1(xiξ2n+1xj)=δj,p1xiξ2n+2 (n2) .\displaystyle\begin{split}h_{-1}(x^{i})&=\xi_{0}x^{i},\\ h_{0}(x^{i}\xi_{0}x^{j})&=\sum_{l=0}^{i-1}x^{l}\xi_{1}x^{i+j-1-l},\\ h_{1}(x^{i}\xi_{1}x^{j})&=\delta_{i,p-1}x^{j}\xi_{2},\\ h_{2n}(x^{i}\xi_{2n}x^{j})&=-\sum_{l=0}^{j-1}x^{i+j-1-l}\xi_{2n+1}x^{l}\text{ $(n\geq 2)$ ,}\\ h_{2n+1}(x^{i}\xi_{2n+1}x^{j})&=\delta_{j,p-1}x^{i}\xi_{2n+2}\text{ $(n\geq 2)$ .}\end{split} (3.2)

Then, we take ϕ1=0\phi_{-1}=0 and construct the following AeA^{e}-linear maps ϕi:(𝔸A𝔸)i𝔸i+1\phi_{i}:(\mathbb{A}\otimes_{A}\mathbb{A})_{i}\longrightarrow\mathbb{A}_{i+1} for degree 1 and 2 by (2.8):

ϕ0(ξ0Axiξ0)=l=0i1xlξ1xi1l,ϕ1(ξ1Axiξ0)=δi,p1ξ2,ϕ1(ξ0Axiξ1)=δi,p1ξ2.\displaystyle\begin{split}&\phi_{0}(\xi_{0}\otimes_{A}x^{i}\xi_{0})=\sum_{l=0}^{i-1}x^{l}\xi_{1}x^{i-1-l},\\ &\phi_{1}(\xi_{1}\otimes_{A}x^{i}\xi_{0})=-\delta_{i,p-1}\xi_{2},\\ &\phi_{1}(\xi_{0}\otimes_{A}x^{i}\xi_{1})=\delta_{i,p-1}\xi_{2}.\end{split} (3.3)

Lastly, we form the following diagonal map Δ:𝔸𝔸A𝔸\Delta:\mathbb{A}\longrightarrow\mathbb{A}\otimes_{A}\mathbb{A}:

Δ0(ξ0)=ξ0Aξ0,Δ1(ξ1)=ξ1Aξ0+ξ0Aξ1,Δ2n(ξ2n)=i=0nξ2iAξ2n2i+i=0n1a+b+c=p2xaξ2i+1Axbξ2n2i1xc, for n1Δ2n+1(ξ2n+1)=i=02n+1ξiAξ2n+1i, for n1.\displaystyle\begin{split}&\Delta_{0}(\xi_{0})=\xi_{0}\otimes_{A}\xi_{0},\\ &\Delta_{1}(\xi_{1})=\xi_{1}\otimes_{A}\xi_{0}+\xi_{0}\otimes_{A}\xi_{1},\\ &\Delta_{2n}(\xi_{2n})=\sum_{i=0}^{n}\xi_{2i}\otimes_{A}\xi_{2n-2i}+\sum_{i=0}^{n-1}\sum_{a+b+c=p-2}x^{a}\xi_{2i+1}\otimes_{A}x^{b}\xi_{2n-2i-1}x^{c},\text{ for }n\geq 1\\ &\Delta_{2n+1}(\xi_{2n+1})=\sum_{i=0}^{2n+1}\xi_{i}\otimes_{A}\xi_{2n+1-i},\text{ for }n\geq 1.\end{split} (3.4)

It can be seen that the map Δ\Delta is a chain map lifting the canonical isomorphism AAAAA\stackrel{{\scriptstyle\sim}}{{\rightarrow}}A\otimes_{A}A by direct calculation.

Now, we are ready to calculate the brackets on cohomology in low degrees. By applying Hom(,A)Ae{}_{A^{e}}(-,A) to 𝔸\mathbb{A}, we see that the differentials are all 0 in odd degrees and (pxp1)(px^{p-1})\cdot in even degrees. In each degree, the term in the Hom complex is the free AA-module Hom(Ae,A)AeA{}_{A^{e}}(A^{e},A)\cong A. Moreover, since pp is not divisible by the characteristic of kk, we deduce HH(A)0A,HH2i+1(A)(x), and HH2i(A)A/(xp1){}^{0}(A)\cong A,\text{HH}^{2i+1}(A)\cong(x),\text{ and }\text{HH}^{2i}(A)\cong A/(x^{p-1}) [14, Section 1.1].

Let xjξix^{j}\xi^{*}_{i}\in Hom(Ae,A)Ae{}_{A^{e}}(A^{e},A) denote the function that takes ξi\xi_{i} to xjx^{j}. Since the characteristic of kk does not divide pp, the Hochschild cohomology as an AA-algebra is generated by ξ1\xi_{1}^{*} and ξ2\xi_{2}^{*} [14, Example 2.2.2]. We only calculate the brackets of the elements of degrees 1 and 2 which can be extended to higher degrees by the formula (2.3). Hence, we have the following calculations:

The bracket of the elements of degrees 1 and 1:

(xiξ1ϕxjξ1)(ξ1)\displaystyle(x^{i}\xi_{1}^{*}\circ_{\phi}x^{j}\xi_{1}^{*})(\xi_{1})
=xiξ1ϕ(1Axjξ1A1)Δ(2)(ξ1)\displaystyle=x^{i}\xi_{1}^{*}\phi(1\otimes_{A}x^{j}\xi_{1}^{*}\otimes_{A}1)\Delta^{(2)}(\xi_{1})
=xiξ1ϕ(1Axjξ1A1)(ξ1Aξ0Aξ0+ξ0Aξ1Aξ0+ξ0Aξ0ξ1)\displaystyle=x^{i}\xi_{1}^{*}\phi(1\otimes_{A}x^{j}\xi_{1}^{*}\otimes_{A}1)(\xi_{1}\otimes_{A}\xi_{0}\otimes_{A}\xi_{0}+\xi_{0}\otimes_{A}\xi_{1}\otimes_{A}\xi_{0}+\xi_{0}\otimes_{A}\xi_{0}\otimes\xi_{1})
=xiξ1ϕ(ξ0Axjξ0)\displaystyle=x^{i}\xi_{1}^{*}\phi(\xi_{0}\otimes_{A}x^{j}\xi_{0})
=xiξ1(ξ1xj1+xξ1xj2++xj1ξ1)\displaystyle=x^{i}\xi_{1}^{*}(\xi_{1}x^{j-1}+x\xi_{1}x^{j-2}+\cdots+x^{j-1}\xi_{1})
=jxi+j1\displaystyle=jx^{i+j-1}

and by symmetry (xjξ1ϕxiξ1)(ξ1)=ixi+j1(x^{j}\xi_{1}^{*}\circ_{\phi}x^{i}\xi_{1}^{*})(\xi_{1})=ix^{i+j-1}. Therefore, we have

[xiξ1,xjξ1]=(ji)xi+j1ξ1.[x^{i}\xi_{1}^{*},x^{j}\xi_{1}^{*}]=(j-i)x^{i+j-1}\xi_{1}^{*}.

The bracket of the elements of degrees 1 and 2:

(xiξ1ϕxjξ2)(ξ2)\displaystyle(x^{i}\xi_{1}^{*}\circ_{\phi}x^{j}\xi_{2}^{*})(\xi_{2})
=xiξ1ϕ(1Axjξ2A1)Δ(2)(ξ2)\displaystyle=x^{i}\xi_{1}^{*}\phi(1\otimes_{A}x^{j}\xi_{2}^{*}\otimes_{A}1)\Delta^{(2)}(\xi_{2})
=xiξ1ϕ(1Axjξ2A1)(ξ0Aξ0Aξ2+ξ0Aξ2Aξ0+ξ2Aξ0Aξ0\displaystyle=x^{i}\xi_{1}^{*}\phi(1\otimes_{A}x^{j}\xi_{2}^{*}\otimes_{A}1)(\xi_{0}\otimes_{A}\xi_{0}\otimes_{A}\xi_{2}+\xi_{0}\otimes_{A}\xi_{2}\otimes_{A}\xi_{0}+\xi_{2}\otimes_{A}\xi_{0}\otimes_{A}\xi_{0}
+ξ0Aa+b+c=p2(xaξ1Axbξ1xc)+a+b+c=p2xaξ1Axb(ξ0Aξ1+ξ1Ax0)xc)\displaystyle+\xi_{0}\otimes_{A}\sum_{\begin{subarray}{c}a+b+c\\ =p-2\end{subarray}}(x^{a}\xi_{1}\otimes_{A}x^{b}\xi_{1}x^{c})+\sum_{\begin{subarray}{c}a+b+c\\ =p-2\end{subarray}}x^{a}\xi_{1}\otimes_{A}x^{b}(\xi_{0}\otimes_{A}\xi_{1}+\xi_{1}\otimes_{A}x_{0})x^{c})
=xiξ1ϕ(ξ0Axjξ0)=xiξ1(ξ1xj1+xξ1xj2++xj1ξ1)=jxi+j1.\displaystyle=x^{i}\xi_{1}^{*}\phi(\xi_{0}\otimes_{A}x^{j}\xi_{0})=x^{i}\xi_{1}^{*}(\xi_{1}x^{j-1}+x\xi_{1}x^{j-2}+\cdots+x^{j-1}\xi_{1})=jx^{i+j-1}.

The circle product in the reverse order is

(xjξ2ϕxp1ξ1)(ξ2)\displaystyle(x^{j}\xi_{2}^{*}\circ_{\phi}x^{p-1}\xi_{1}^{*})(\xi_{2})
=xjξ2ϕ(1Axp1ξ1A1)Δ(2)(ξ2)\displaystyle=x^{j}\xi_{2}^{*}\phi(1\otimes_{A}x^{p-1}\xi_{1}^{*}\otimes_{A}1)\Delta^{(2)}(\xi_{2})
=xjξ2ϕ(1Axp1ξ1A1)(ξ0Aξ0Aξ2+ξ0Aξ2Aξ0+ξ2Aξ0Aξ0\displaystyle=x^{j}\xi_{2}^{*}\phi(1\otimes_{A}x^{p-1}\xi_{1}^{*}\otimes_{A}1)(\xi_{0}\otimes_{A}\xi_{0}\otimes_{A}\xi_{2}+\xi_{0}\otimes_{A}\xi_{2}\otimes_{A}\xi_{0}+\xi_{2}\otimes_{A}\xi_{0}\otimes_{A}\xi_{0}
+ξ0Aa+b+c=p2(xaξ1Axbξ1xc)+a+b+c=p2xaξ1Axb(ξ0Aξ1+ξ1Ax0)xc)\displaystyle+\xi_{0}\otimes_{A}\sum_{\begin{subarray}{c}a+b+c\\ =p-2\end{subarray}}(x^{a}\xi_{1}\otimes_{A}x^{b}\xi_{1}x^{c})+\sum_{\begin{subarray}{c}a+b+c\\ =p-2\end{subarray}}x^{a}\xi_{1}\otimes_{A}x^{b}(\xi_{0}\otimes_{A}\xi_{1}+\xi_{1}\otimes_{A}x_{0})x^{c})
=xjξ2ϕ(a+b+c=p2(ξ0Axa+b+iξ1xcxaξ1Axb+iξ0xc))\displaystyle=x^{j}\xi_{2}^{*}\phi(\sum_{\begin{subarray}{c}a+b+c\\ =p-2\end{subarray}}(\xi_{0}\otimes_{A}x^{a+b+i}\xi_{1}x^{c}-x^{a}\xi_{1}\otimes_{A}x^{b+i}\xi_{0}x^{c}))
=xjξ2(a+b+c=p2(δa+b+i,p1ξ2xc+xaδb+i,p1ξ2xc))\displaystyle=x^{j}\xi_{2}^{*}(\sum_{\begin{subarray}{c}a+b+c\\ =p-2\end{subarray}}(\delta_{a+b+i,p-1}\xi_{2}x^{c}+x^{a}\delta_{b+i,p-1}\xi_{2}x^{c}))
=xjξ2((pi)ξ2xi1+a+c=i1xaξ2xc)\displaystyle=x^{j}\xi_{2}^{*}((p-i)\xi_{2}x^{i-1}+\sum_{\begin{subarray}{c}a+c\\ =i-1\end{subarray}}x^{a}\xi_{2}x^{c})
=(pi)xi+j1+a+c=i1xa+c+j=(pi)xi+j1+ixi+j1=pxi+j1.\displaystyle=(p-i)x^{i+j-1}+\sum_{\begin{subarray}{c}a+c\\ =i-1\end{subarray}}x^{a+c+j}=(p-i)x^{i+j-1}+ix^{i+j-1}=px^{i+j-1}.

Therefore, we obtain

[xiξ1,xjξ2]=(jp)xi+j1ξ2.[x^{i}\xi_{1}^{*},x^{j}\xi_{2}^{*}]=(j-p)x^{i+j-1}\xi^{*}_{2}.

Lastly, the bracket of the elements of degrees 2 and 2:

(xiξ2ϕxjξ2)(ξ3)\displaystyle(x^{i}\xi_{2}^{*}\circ_{\phi}x^{j}\xi_{2}^{*})(\xi_{3}) =xiξ2ϕ(1Axjξ2A1)Δ(2)(ξ3)\displaystyle=x^{i}\xi_{2}^{*}\phi(1\otimes_{A}x^{j}\xi_{2}^{*}\otimes_{A}1)\Delta^{(2)}(\xi_{3})
=xiξ2ϕ(ξ1Axjξ0+ξ0Axjξ1)=xiξ2(0)=0\displaystyle=x^{i}\xi_{2}^{*}\phi(\xi_{1}\otimes_{A}x^{j}\xi_{0}+\xi_{0}\otimes_{A}x^{j}\xi_{1})=x^{i}\xi_{2}^{*}(0)=0

and by symmetry (xjξ2ϕxiξ2)(ξ3)=0(x^{j}\xi_{2}^{*}\circ_{\phi}x^{i}\xi_{2}^{*})(\xi_{3})=0. Therefore, we have

[(xiξ2,xjξ2)]=0.[(x^{i}\xi_{2}^{*},x^{j}\xi_{2}^{*})]=0.

As a consequence, the brackets for the elements of degrees 1 and 2 are

[(xiξ1,xjξ1)]\displaystyle[(x^{i}\xi_{1}^{*},x^{j}\xi_{1}^{*})] =(ji)xi+j1ξ1,\displaystyle=(j-i)x^{i+j-1}\xi_{1}^{*},
[(xiξ1,xjξ2)]\displaystyle[(x^{i}\xi_{1}^{*},x^{j}\xi_{2}^{*})] =(jp)xi+j1ξ2,\displaystyle=(j-p)x^{i+j-1}\xi^{*}_{2},
[(xiξ2,xjξ2)]\displaystyle[(x^{i}\xi_{2}^{*},x^{j}\xi_{2}^{*})] =0.\displaystyle=0.

Brackets in higher degrees can be determined from these and the identity (2.3) since the Hochschild cohomology is generated as an AA-algebra under the cup product in degrees 1 and 2.

L. Grimley, V. C. Nguyen, and S. Witherspoon [4] calculated Gerstenhaber brackets on Hochschild cohomology of a twisted tensor product of algebras. S. Sanchez-Flores [9] also calculated the bracket on group algebras of a cyclic group over a field of positive characteristic which is isomorphic to A=k[x]/(xp)A=k[x]/(x^{p}). C. Negron and S. Witherspoon [7] calculated the bracket on group algebras of a cyclic group over a field of positive characteristic as well with the same h,ϕh,\phi, and Δ\Delta maps. Our calculation agrees with those except slightly different [(xiξ1,xjξ2)][(x^{i}\xi_{1}^{*},x^{j}\xi_{2}^{*})].

4 Bracket on Hopf algebra cohomology of a Taft algebra

The Taft algebra TpT_{p} with p>2p>2 is a kk-algebra generated by gg and xx satisfying the relations : gp=1,xp=0, and xg=ωgxg^{p}=1,x^{p}=0,\text{ and }xg=\omega gx where ω\omega is a primitive pp-th root of unity. It is a Hopf algebra with the structure:

  • Δ(g)=gg\Delta(g)=g\otimes g, Δ(x)=1x+xg\Delta(x)=1\otimes x+x\otimes g

  • ε(g)=1,ε(x)=0\varepsilon(g)=1,\varepsilon(x)=0

  • S(g)=g1,S(x)=xg1.S(g)=g^{-1},S(x)=-xg^{-1}.

Note that as an algebra, TpT_{p} is a skew group algebra AkGA\rtimes kG where A=k[x]/(xp)A=k[x]/(x^{p}) and G=<ggp=1>G=<g\mid g^{p}=1>. The action of GG on AA is given by xg=ωx{}^{g}x=\omega x.

In this section, our main goal is to calculate the bracket on Hochschild cohomology of TpT_{p} with the same technique in Section 3 and find the bracket on Hopf algebra cohomology of TpT_{p} by using the embedding of H(Tp,k)\text{H}^{*}(T_{p},k) into HH(Tp,Tp)\text{HH}^{*}(T_{p},T_{p}).

We first find the bracket on Hochschild cohomology of TpT_{p}. Let 𝒟\mathcal{D} be the skew group algebra AeGA^{e}\rtimes G where the action of GG on AeA^{e} is diagonal, i.e. (ab)g=(ga)(gb){}^{g}(a\otimes b)=(^{g}a)\otimes(^{g}b). Then, there is the following isomorphism [1, Section 2]

𝒟=AeGgGAgAg1Tpe.\mathcal{D}=A^{e}\rtimes G\cong\underset{g\in G}{\bigoplus}Ag\otimes Ag^{-1}\subset T_{p}^{e}.

Hence 𝒟\mathcal{D} is isomorphic to a subalgebra of TpeT_{p}^{e} via a1a2ga1g(g1a2g1)a_{1}\otimes a_{2}\otimes g\mapsto a_{1}g\otimes(^{g^{-1}}a_{2}g^{-1}). Moreover, AA is a 𝒟\mathcal{D}-module under the following left and right action [1, Section 4]:

(a1ga2g1)a3=a1ga3a2g1=a1(g(a3a2))(a_{1}g\otimes a_{2}g^{-1})a_{3}=a_{1}ga_{3}a_{2}g^{-1}=a_{1}(^{g}(a_{3}a_{2}))
a3(a1ga2g1)=a2g1a3a1g=a2(g1(a3a1)).a_{3}(a_{1}g\otimes a_{2}g^{-1})=a_{2}g^{-1}a_{3}a_{1}g=a_{2}(^{g^{-1}}(a_{3}a_{1})).

Remember the resolution (3.1)

𝔸:v.Aeu.Aev.Aeu.AeπA0.\mathbb{A}:\cdots\stackrel{{\scriptstyle v.}}{{\longrightarrow}}A^{e}\stackrel{{\scriptstyle u.}}{{\longrightarrow}}A^{e}\stackrel{{\scriptstyle v.}}{{\longrightarrow}}A^{e}\stackrel{{\scriptstyle u.}}{{\longrightarrow}}A^{e}\stackrel{{\scriptstyle\pi}}{{\longrightarrow}}A\longrightarrow 0.

This is also a 𝒟\mathcal{D}-projective resolution of A and the action of GG on AeA^{e} is given by

  • g(a1a2)=(ga1)(ga2)g\cdot(a_{1}\otimes a_{2})=(^{g}a_{1})\otimes(^{g}a_{2}) in even degrees,

  • g(a1a2)=ω(ga1)(ga2)g\cdot(a_{1}\otimes a_{2})=\omega(^{g}a_{1})\otimes(^{g}a_{2}) in odd degrees.

From the resolution 𝔸\mathbb{A}, we construct the following TpeT_{p}^{e} resolution of TpT_{p}:

Tpe𝒟𝔸:Tpe𝒟AeTpe𝒟AeTpe𝒟AeTpe𝒟A0.T_{p}^{e}\otimes_{\mathcal{D}}\mathbb{A}:\cdots{\longrightarrow}T_{p}^{e}\otimes_{\mathcal{D}}{A^{e}}{\longrightarrow}T_{p}^{e}\otimes_{\mathcal{D}}{A^{e}}{\longrightarrow}T_{p}^{e}\otimes_{\mathcal{D}}{A^{e}}{\longrightarrow}T_{p}^{e}\otimes_{\mathcal{D}}{A}\longrightarrow 0. (4.1)

It is known that, TpTpe𝒟AT_{p}\cong T_{p}^{e}\otimes_{\mathcal{D}}{A} as TpT_{p}-bimodules via the map sending xigkx^{i}\otimes g^{k} to (1gk)𝒟xi(1\otimes g^{k})\otimes_{\mathcal{D}}x^{i} [14, Section 3.5]. Then we have ATpTpe𝒟AeA\otimes T_{p}\cong T_{p}^{e}\otimes_{\mathcal{D}}{A^{e}} with the TpT_{p}-bimodule isomorphism given by

κ(xi(xjgk))=(1gk)𝒟(xixj).\kappa(x^{i}\otimes(x^{j}\otimes g^{k}))=(1\otimes g^{k})\otimes_{\mathcal{D}}(x^{i}\otimes x^{j}). (4.2)

Then, we obtain the following resolution 𝔸~\tilde{\mathbb{A}} which is isomorphic to the resolution (4.1), i.e.

𝔸~:u~.ATpv~.ATpu~.ATpπ~.Tp0\tilde{\mathbb{A}}:\cdots\stackrel{{\scriptstyle\tilde{u}.}}{{\longrightarrow}}A\otimes T_{p}\stackrel{{\scriptstyle\tilde{v}.}}{{\longrightarrow}}A\otimes T_{p}\stackrel{{\scriptstyle\tilde{u}.}}{{\longrightarrow}}A\otimes T_{p}\stackrel{{\scriptstyle\tilde{\pi}.}}{{\longrightarrow}}T_{p}\longrightarrow 0 (4.3)

where v~=vidkG,u~=uidkG\tilde{v}=v\otimes id_{kG},\tilde{u}=u\otimes id_{kG}, and π~=πidkG\tilde{\pi}=\pi\otimes id_{kG}.

The following lemma gives us a contracting homotopy for the identity map on the resolution 𝔸~\tilde{\mathbb{A}}.

Lemma 4.4.

Let hnh_{n} be a contracting homotopy in (3.2). Then h~n=hn1kG\tilde{h}_{n}=h_{n}\otimes 1_{kG} forms a contracting homotopy for the identity map on 𝔸~\tilde{\mathbb{A}}.

Proof.

For n0n\geq 0, the domain of hn1kGh_{n}\otimes 1_{kG} is AAkGA\otimes A\otimes kG which is ATpA\otimes T_{p} as a vector space. Moreover, by definition of contracting homotopy, hnh_{n} satisfy

hi1di+di+1hi=id𝔸i.h_{i-1}d_{i}+d_{i+1}h_{i}=id_{\mathbb{A}_{i}}.

Then,

h~i1d~i+d~i+1h~i\displaystyle\tilde{h}_{i-1}\tilde{d}_{i}+\tilde{d}_{i+1}\tilde{h}_{i} =(hi1idkG)(diidkG)+(di+1idkG)(hiidkG)\displaystyle=(h_{i-1}\otimes id_{kG})(d_{i}\otimes id_{kG})+(d_{i+1}\otimes id_{kG})(h_{i}\otimes id_{kG})
=(hi1diidkG)+(di+1hiidkG)=(hi1di+di+1hi)idkG\displaystyle=(h_{i-1}d_{i}\otimes id_{kG})+(d_{i+1}h_{i}\otimes id_{kG})=(h_{i-1}d_{i}+d_{i+1}h_{i})\otimes id_{kG}
=id𝔸iidkG=id𝔸~i\displaystyle=id_{\mathbb{A}_{i}}\otimes id_{kG}=id_{\tilde{\mathbb{A}}_{i}}

and that implies h~n\tilde{h}_{n} is a contracting homotopy for 𝔸~\tilde{\mathbb{A}}. The proof is similar for n=1n=-1. ∎

We abbreviate a1a2gATpa_{1}\otimes a_{2}\otimes g\in A\otimes T_{p} by a1a2ga_{1}\otimes a_{2}g. By the Lemma 4.4, we obtain

h~1(xig)\displaystyle\tilde{h}_{-1}(x^{i}g) =ξ0xig,\displaystyle=\xi_{0}x^{i}g,
h~0(xiξ0xjg)\displaystyle\tilde{h}_{0}(x^{i}\xi_{0}x^{j}g) =l=0i1xlξ1xi+j1lg,\displaystyle=\sum_{l=0}^{i-1}x^{l}\xi_{1}x^{i+j-1-l}g,
h~1(xiξ1xjg)\displaystyle\tilde{h}_{1}(x^{i}\xi_{1}x^{j}g) =δi,p1xjξ2g,\displaystyle=\delta_{i,p-1}x^{j}\xi_{2}g,
h~2n(xiξ2nxjg)\displaystyle\tilde{h}_{2n}(x^{i}\xi_{2n}x^{j}g) =l=0j1xi+j1lξ2n+1xlg,\displaystyle=-\sum_{l=0}^{j-1}x^{i+j-1-l}\xi_{2n+1}x^{l}g,
h~2n+1(xiξ2n+1xjg)\displaystyle\tilde{h}_{2n+1}(x^{i}\xi_{2n+1}x^{j}g) =δj,p1xiξ2n+2g.\displaystyle=\delta_{j,p-1}x^{i}\xi_{2n+2}g.

We need a lemma to have the linear maps ϕ~i:(𝔸~Tp𝔸~)i𝔸~i+1\tilde{\phi}_{i}:(\tilde{\mathbb{A}}\otimes_{T_{p}}\tilde{\mathbb{A}})_{i}\longrightarrow\tilde{\mathbb{A}}_{i+1}. However, we first mention that there is an isomorphism from (ATp)Tp(ATp)(A\otimes T_{p})\otimes_{T_{p}}(A\otimes T_{p}) to (AA)A(AA)kG(A\otimes A)\otimes_{A}(A\otimes A)\otimes kG as TpeT_{p}^{e}-modules given by

ψ((xi1xj1gk1)Tp(xi2xj2gk2))=ωk1(i2+j2)(xi1xj1)A(xi2xj2)g(k1+k2).\psi((x^{i_{1}}\otimes x^{j_{1}}g^{k_{1}})\otimes_{T_{p}}(x^{i_{2}}\otimes x^{j_{2}}g^{k_{2}}))=\omega^{k_{1}(i_{2}+j_{2})}(x^{i_{1}}\otimes x^{j_{1}})\otimes_{A}(x^{i_{2}}\otimes x^{j_{2}})g^{(k_{1}+k_{2})}. (4.5)
Lemma 4.6.

Let F𝔸=(πAid𝔸id𝔸Aπ)F_{\mathbb{A}}=(\pi\otimes_{A}id_{\mathbb{A}}-id_{\mathbb{A}}\otimes_{A}\pi) be the chain map for the resolution 𝔸\mathbb{A} in (3.1) which is used for calculation of ϕ\phi in (3.3). Then F𝔸~:𝔸~Tp𝔸~𝔸~F_{\tilde{\mathbb{A}}}:\tilde{\mathbb{A}}\otimes_{T_{p}}\tilde{\mathbb{A}}\to\tilde{\mathbb{A}} defined by (π~Tpid𝔸~id𝔸~Tpπ~)(\tilde{\pi}\otimes_{T_{p}}id_{\tilde{\mathbb{A}}}-id_{\tilde{\mathbb{A}}}\otimes_{T_{p}}\tilde{\pi}) is exactly (F𝔸idkG)ψ(F_{\mathbb{A}}\otimes id_{kG})\psi. Moreover ϕ~:=(ϕidkG)ψ\tilde{\phi}:=(\phi\otimes id_{kG})\psi is a contracting homotopy for F𝔸~F_{\tilde{\mathbb{A}}}.

Proof.

Let (xi1xj1gk1)Tp(xi2xj2gk2)(ATp)Tp(ATp)(x^{i_{1}}\otimes x^{j_{1}}g^{k_{1}})\otimes_{T_{p}}(x^{i_{2}}\otimes x^{j_{2}}g^{k_{2}})\in(A\otimes T_{p})\otimes_{T_{p}}(A\otimes T_{p}). Note that F𝔸~F_{\tilde{\mathbb{A}}} is zero if degrees of (xi1xj1gk1)(x^{i_{1}}\otimes x^{j_{1}}g^{k_{1}}) and (xi2xj2gk2)(x^{i_{2}}\otimes x^{j_{2}}g^{k_{2}}) are both nonzero since π~\tilde{\pi} is only defined on degree zero. Also remember that π~=πidkG\tilde{\pi}=\pi\otimes id_{kG} for the resolution 𝔸~\tilde{\mathbb{A}}.

We check the case that the degree of (xi1xj1gk1)(x^{i_{1}}\otimes x^{j_{1}}g^{k_{1}}) is zero and the degree of (xi2xj2gk2)(x^{i_{2}}\otimes x^{j_{2}}g^{k_{2}}) is nonzero. By using definition of F𝔸~F_{\tilde{\mathbb{A}}}, we obtain

F𝔸~((xi1xj1gk1)Tp(xi2xj2gk2))\displaystyle F_{\tilde{\mathbb{A}}}((x^{i_{1}}\otimes x^{j_{1}}g^{k_{1}})\otimes_{T_{p}}(x^{i_{2}}\otimes x^{j_{2}}g^{k_{2}})) =(xi1+j1gk1)Tp(xi2xj2gk2)\displaystyle=(x^{i_{1}+j_{1}}g^{k_{1}})\otimes_{T_{p}}(x^{i_{2}}\otimes x^{j_{2}}g^{k_{2}})
=ωk1(i2+j2)xi1+i2+j1xi2gk1+k2.\displaystyle=\omega^{k_{1}(i_{2}+j_{2})}x^{i_{1}+i_{2}+j_{1}}\otimes x^{i_{2}}g^{k_{1}+k_{2}}.

On the other hand, we also have

(F𝔸idkG)ψ((xi1xj1gk1)Tp(xi2xj2gk2))\displaystyle(F_{\mathbb{A}}\otimes id_{kG})\psi((x^{i_{1}}\otimes x^{j_{1}}g^{k_{1}})\otimes_{T_{p}}(x^{i_{2}}\otimes x^{j_{2}}g^{k_{2}}))
=(F𝔸idkG)(ωk1(i2+j2)(xi1xj1)A(xi2xj2)gk1+k2)\displaystyle=(F_{\mathbb{A}}\otimes id_{kG})(\omega^{k_{1}(i_{2}+j_{2})}(x^{i_{1}}\otimes x^{j_{1}})\otimes_{A}(x^{i_{2}}\otimes x^{j_{2}})g^{k_{1}+k_{2}})
=ωk1(i2+j2)xi1+i2+j1xi2gk1+k2.\displaystyle=\omega^{k_{1}(i_{2}+j_{2})}x^{i_{1}+i_{2}+j_{1}}\otimes x^{i_{2}}g^{k_{1}+k_{2}}.

The proof for other cases are similar. Hence F𝔸~F_{\tilde{\mathbb{A}}} and (F𝔸idkG)ψ(F_{\mathbb{A}}\otimes id_{kG})\psi are identical.

In order to prove ϕ~:=(ϕidkG)ψ\tilde{\phi}:=(\phi\otimes id_{kG})\psi is a contracting homotopy for F𝔸~F_{\tilde{\mathbb{A}}}, we need to show that

d~𝔸~ϕ~+ϕ~d~𝔸~Tp𝔸~=F𝔸~.\tilde{d}_{\tilde{\mathbb{A}}}\tilde{\phi}+\tilde{\phi}\tilde{d}_{\tilde{\mathbb{A}}\otimes_{T_{p}}\tilde{\mathbb{A}}}=F_{\tilde{\mathbb{A}}}.

It is clear that

d~𝔸~ϕ~=(d𝔸idkG)(ϕidkG)ψ=(d𝔸ϕidkG)ψ.\tilde{d}_{\tilde{\mathbb{A}}}\tilde{\phi}=(d_{\mathbb{A}}\otimes id_{kG})(\phi\otimes id_{kG})\psi=(d_{\mathbb{A}}\phi\otimes id_{kG})\psi. (4.7)

We now claim that

ψd~𝔸~Tp𝔸~=(d𝔸A𝔸idkG)ψ.\psi\tilde{d}_{\tilde{\mathbb{A}}\otimes_{T_{p}}\tilde{\mathbb{A}}}=(d_{\mathbb{A}\otimes_{A}\mathbb{A}}\otimes id_{kG})\psi. (4.8)

By definition

d~𝔸~Tp𝔸~=d~𝔸~TpidTp+(1)idTpTpd~𝔸~\tilde{d}_{\tilde{\mathbb{A}}\otimes_{T_{p}}\tilde{\mathbb{A}}}=\tilde{d}_{\tilde{\mathbb{A}}}\otimes_{T_{p}}id_{T_{p}}+(-1)^{*}id_{T_{p}}\otimes_{T_{p}}\tilde{d}_{\tilde{\mathbb{A}}}

where * is the degree of the element in left ATpA\otimes T_{p}. Moreover, (ATp)Tp(ATp)(A\otimes T_{p})\otimes_{T_{p}}(A\otimes T_{p}) is generated by ξm1GTpxiξn1G\xi_{m}1_{G}\otimes_{T_{p}}x^{i}\xi_{n}1_{G} as TpT_{p}-bimodule. Without loss of generality, assume mm and nn are odd. Then we have the following calculation:

ψd~𝔸~Tp𝔸~(ξm1GTpxiξn1G)=ψ((xξm1Gξmx1G)Tpxiξn1Gξm1GTp(xi+1ξn1Gxiξnx1G))=(xξmξmx)Axiξn1GξmA(xi+1ξnxiξnx)1G\displaystyle\begin{aligned} &\psi\tilde{d}_{\tilde{\mathbb{A}}\otimes_{T_{p}}\tilde{\mathbb{A}}}(\xi_{m}1_{G}\otimes_{T_{p}}x^{i}\xi_{n}1_{G})\\ &=\psi((x\xi_{m}1_{G}-\xi_{m}x1_{G})\otimes_{T_{p}}x^{i}\xi_{n}1_{G}-\xi_{m}1_{G}\otimes_{T_{p}}(x^{i+1}\xi_{n}1_{G}-x^{i}\xi_{n}x1_{G}))\\ &=(x\xi_{m}-\xi_{m}x)\otimes_{A}x^{i}\xi_{n}1_{G}-\xi_{m}\otimes_{A}(x^{i+1}\xi_{n}-x^{i}\xi_{n}x)1_{G}\end{aligned}

and

(d𝔸A𝔸idkG)ψ(ξm1GTpxiξn1G)=(d𝔸A𝔸idkG)(ξmAxiξn1G)=(xξmξmx)Axiξn1GξmA(xi+1ξnxiξnx)1G.\displaystyle\begin{aligned} &(d_{\mathbb{A}\otimes_{A}\mathbb{A}}\otimes id_{kG})\psi(\xi_{m}1_{G}\otimes_{T_{p}}x^{i}\xi_{n}1_{G})\\ &=(d_{\mathbb{A}\otimes_{A}\mathbb{A}}\otimes id_{kG})(\xi_{m}\otimes_{A}x^{i}\xi_{n}1_{G})\\ &=(x\xi_{m}-\xi_{m}x)\otimes_{A}x^{i}\xi_{n}1_{G}-\xi_{m}\otimes_{A}(x^{i+1}\xi_{n}-x^{i}\xi_{n}x)1_{G}.\end{aligned}

The calculation is similar for the other cases of mm and nn. Therefore,

ϕ~d~𝔸~Tp𝔸~=(ϕidkG)ψd~𝔸~Tp𝔸~=(ϕidkG)(d𝔸A𝔸idkG)ψ=(ϕd𝔸A𝔸idkG)ψ.\tilde{\phi}\tilde{d}_{\tilde{\mathbb{A}}\otimes_{T_{p}}\tilde{\mathbb{A}}}=(\phi\otimes id_{kG})\psi\tilde{d}_{\tilde{\mathbb{A}}\otimes_{T_{p}}\tilde{\mathbb{A}}}=(\phi\otimes id_{kG})(d_{\mathbb{A}\otimes_{A}\mathbb{A}}\otimes id_{kG})\psi=(\phi d_{\mathbb{A}\otimes_{A}\mathbb{A}}\otimes id_{kG})\psi. (4.9)

By combining (4.7) and (4.9), we obtain

d~𝔸~ϕ~+ϕ~d~𝔸~Tp𝔸~=((d𝔸ϕ+ϕd𝔸A𝔸)idkG)ψ=(F𝔸idkG)ψ=F𝔸~\tilde{d}_{\tilde{\mathbb{A}}}\tilde{\phi}+\tilde{\phi}\tilde{d}_{\tilde{\mathbb{A}}\otimes_{T_{p}}\tilde{\mathbb{A}}}=((d_{\mathbb{A}}\phi+\phi d_{\mathbb{A}\otimes_{A}\mathbb{A}})\otimes id_{kG})\psi=(F_{\mathbb{A}}\otimes id_{kG})\psi=F_{\tilde{\mathbb{A}}}

whence ϕ~=(ϕidkG)ψ\tilde{\phi}=(\phi\otimes id_{kG})\psi is a contracting homotopy for F𝔸~F_{\tilde{\mathbb{A}}}. ∎

We use the Lemma 4.6 and find the following TpeT_{p}^{e}-linear maps ϕ~i:(𝔸~Tp𝔸~)i𝔸~i+1\tilde{\phi}_{i}:(\tilde{\mathbb{A}}\otimes_{T_{p}}\tilde{\mathbb{A}})_{i}\longrightarrow\tilde{\mathbb{A}}_{i+1}:

ϕ~0(ξ01GTpxiξ01G)=l=0i1xlξ1xi1l1G,\displaystyle\tilde{\phi}_{0}(\xi_{0}1_{G}\otimes_{T_{p}}x^{i}\xi_{0}1_{G})=\sum_{l=0}^{i-1}x^{l}\xi_{1}x^{i-1-l}1_{G},
ϕ~1(ξ11GTpxiξ01G)=δi,p1ξ21G,\displaystyle\tilde{\phi}_{1}(\xi_{1}1_{G}\otimes_{T_{p}}x^{i}\xi_{0}1_{G})=-\delta_{i,p-1}\xi_{2}1_{G},
ϕ~1(ξ01GTpxiξ11G)=δi,p1ξ21G.\displaystyle\tilde{\phi}_{1}(\xi_{0}1_{G}\otimes_{T_{p}}x^{i}\xi_{1}1_{G})=\delta_{i,p-1}\xi_{2}1_{G}.

Next, we give a lemma to find the the diagonal map.

Lemma 4.10.

The map Δ~:=ψ1(ΔidkG)\tilde{\Delta}:=\psi^{-1}(\Delta\otimes id_{kG}) is a diagonal map on 𝔸~\tilde{\mathbb{A}} where Δ\Delta is in (3.4).

Proof.

We need to check that Δ~\tilde{\Delta} is a chain map. The following equations are straightforward by considering the fact that Δ\Delta is a chain map and (4.8):

d~𝔸~Tp𝔸~Δ~\displaystyle\tilde{d}_{\tilde{\mathbb{A}}\otimes_{T_{p}}\tilde{\mathbb{A}}}\tilde{\Delta} =d~𝔸~Tp𝔸~ψ1(ΔidkG)=ψ1(d𝔸A𝔸idkG)(ΔidkG)\displaystyle=\tilde{d}_{\tilde{\mathbb{A}}\otimes_{T_{p}}\tilde{\mathbb{A}}}\psi^{-1}(\Delta\otimes id_{kG})=\psi^{-1}(d_{\mathbb{A}\otimes_{A}\mathbb{A}}\otimes id_{kG})(\Delta\otimes id_{kG})
=ψ1(d𝔸A𝔸ΔidkG)=ψ1(Δd𝔸idkG)=ψ1(ΔidkG)(d𝔸idkG)\displaystyle=\psi^{-1}(d_{\mathbb{A}\otimes_{A}\mathbb{A}}\Delta\otimes id_{kG})=\psi^{-1}(\Delta d_{\mathbb{A}}\otimes id_{kG})=\psi^{-1}(\Delta\otimes id_{kG})(d_{\mathbb{A}}\otimes id_{kG})
=Δ~d~𝔸~.\displaystyle=\tilde{\Delta}\tilde{d}_{\tilde{\mathbb{A}}}.

Lemma 4.10 allows us to compute the TpT_{p}-linear map Δ~:𝔸~i+1(𝔸~Tp𝔸~)i\tilde{\Delta}:\tilde{\mathbb{A}}_{i+1}\longrightarrow(\tilde{\mathbb{A}}\otimes_{T_{p}}\tilde{\mathbb{A}})_{i} as follows:

Δ~0(ξ01G)\displaystyle\tilde{\Delta}_{0}(\xi_{0}1_{G}) =ξ01GTpξ01G,\displaystyle=\xi_{0}1_{G}\otimes_{T_{p}}\xi_{0}1_{G},
Δ~1(ξ11G)\displaystyle\tilde{\Delta}_{1}(\xi_{1}1_{G}) =ξ11GTpξ01G+ξ01GTpξ11G,\displaystyle=\xi_{1}1_{G}\otimes_{T_{p}}\xi_{0}1_{G}+\xi_{0}1_{G}\otimes_{T_{p}}\xi_{1}1_{G},
Δ~2n(ξ2n1G)\displaystyle\tilde{\Delta}_{2n}(\xi_{2n}1_{G}) =i=0nξ2i1GTpξ2n2i1G\displaystyle=\sum_{i=0}^{n}\xi_{2i}1_{G}\otimes_{T_{p}}\xi_{2n-2i}1_{G}
+i=0n1a+b+c=p2xaξ2i+11GTpxbξ2n2i1xc1G, for n1\displaystyle+\sum_{i=0}^{n-1}\sum_{\begin{subarray}{c}a+b+c\\ =p-2\end{subarray}}x^{a}\xi_{2i+1}1_{G}\otimes_{T_{p}}x^{b}\xi_{2n-2i-1}x^{c}1_{G},\text{ for }n\geq 1
Δ~2n+1(ξ2n+11G)\displaystyle\tilde{\Delta}_{2n+1}(\xi_{2n+1}1_{G}) =i=02n+1ξi1GTpξ2n+1i1G, for n1.\displaystyle=\sum_{i=0}^{2n+1}\xi_{i}1_{G}\otimes_{T_{p}}\xi_{2n+1-i}1_{G},\text{ for }n\geq 1.

Before computing the bracket on Hochschild cohomology of TpT_{p}, we need to find a basis of HomTpe(𝔸~,Tp)\text{Hom}_{T_{p}^{e}}(\tilde{\mathbb{A}},T_{p}). In particular, we must find a basis of HomTpe(ATp,Tp)\text{Hom}_{T_{p}^{e}}(A\otimes T_{p},T_{p}) as it is an invariant in each degree.

It is known that

HH(Tp):=ExtTpe(Tp,Tp)Ext𝒟(A,Tp)ExtAe(A,Tp)G.\operatorname{HH}^{*}(T_{p}):=\text{Ext}^{*}_{T_{p}^{e}}(T_{p},T_{p})\cong\text{Ext}^{*}_{\mathcal{D}}(A,T_{p})\cong\text{Ext}^{*}_{A^{e}}(A,T_{p})^{G}.

The Eckmann-Shapiro Lemma (Lemma 5.3) and (4.2) imply the first isomorphism and see [14, Theorem 3.6.2] for the second isomorphism.

Consider the following resolution

HomAe(𝔸,Tp)G:0HomAe(Ae,Tp)GHomAe(Ae,Tp)G\text{Hom}_{A^{e}}(\mathbb{A},T_{p})^{G}:0{\longrightarrow}\text{Hom}_{A^{e}}(A^{e},T_{p})^{G}{\longrightarrow}\text{Hom}_{A^{e}}(A^{e},T_{p})^{G}{\longrightarrow}\cdots (4.11)

where the action of GG on HomAe(Ae,Tp)G\text{Hom}_{A^{e}}(A^{e},T_{p})^{G} is defined by

gf(a1a2)=fg(g1(a1a2)).g\cdot f(a_{1}\otimes a_{2})={{}^{g}}f(^{g^{-1}}(a_{1}\otimes a_{2})). (4.12)

This resolution is clearly isomorphic to

0TpGTpGTpG0{\longrightarrow}T_{p}^{G}{\longrightarrow}T_{p}^{G}{\longrightarrow}T_{p}^{G}\longrightarrow\cdots (4.13)

with the correspondence

ftt where ft(ξ)=t for all tTp.f_{t}\mapsto t\text{ where }f_{t}(\xi_{*})=t\text{ for all }t\in T_{p}. (4.14)

We claim that HomTpe(ATp,Tp)TpG\text{Hom}_{T_{p}^{e}}(A\otimes T_{p},T_{p})\cong T_{p}^{G}. Suppose xigjTpGx^{i}g^{j}\in T_{p}^{G}. Then, we have fxigjHomAe(Ae,Tp)Gf_{x^{i}g^{j}}\in\text{Hom}_{A^{e}}(A^{e},T_{p})^{G} defined by fxigj(xkxl):=xk+l+igjf_{x^{i}g^{j}}(x^{k}\otimes x^{l}):=x^{k+l+i}g^{j} where xAx^{*}\in A. Now observe that, fxigjHomAe(Ae,Tp)Gf_{x^{i}g^{j}}\in\text{Hom}_{A^{e}}(A^{e},T_{p})^{G} is a 𝒟\mathcal{D}-module homomorphism since

fxigj((xkξxlg)(a1a2))\displaystyle f_{x^{i}g^{j}}((x^{k}\xi_{*}x^{l}g)(a_{1}\otimes a_{2})) =fxigj((xkξxl1G)g(a1a2))=(xkξxl1G)fxigj(g(a1a2))\displaystyle=f_{x^{i}g^{j}}((x^{k}\xi_{*}x^{l}1_{G})g(a_{1}\otimes a_{2}))=(x^{k}\xi_{*}x^{l}1_{G})f_{x^{i}g^{j}}(g(a_{1}\otimes a_{2}))
=(xkξxl1G)gfxigj(a1a2)=(xkξxlg)fxigj(a1a2)\displaystyle=(x^{k}\xi_{*}x^{l}1_{G})gf_{x^{i}g^{j}}(a_{1}\otimes a_{2})=(x^{k}\xi_{*}x^{l}g)f_{x^{i}g^{j}}(a_{1}\otimes a_{2})

where xkξxlg𝒟,a1a2Aex^{k}\xi_{*}x^{l}g\in\mathcal{D},a_{1}\otimes a_{2}\in A^{e}. Moreover, if fHom𝒟(Ae,Tp)f\in\text{Hom}_{\mathcal{D}}(A^{e},T_{p}), then ff is GG-invariant as

gf(a1a2)=fg(g1(a1a2))=f(gg1)(a1a2)=f(a1a2)g\cdot f(a_{1}\otimes a_{2})={{}^{g}}f(^{g^{-1}}(a_{1}\otimes a_{2}))={{}^{(gg^{-1})}}f(a_{1}\otimes a_{2})=f(a_{1}\otimes a_{2})

where gG,a1a2Aeg\in G,a_{1}\otimes a_{2}\in A^{e}. Hence, the isomorphism from HomAe(Ae,Tp)G\text{Hom}_{A^{e}}(A^{e},T_{p})^{G} to Hom𝒟(Ae,Tp)\text{Hom}_{\mathcal{D}}(A^{e},T_{p}) is the identity, so that fxigjf_{x^{i}g^{j}} is also in Hom𝒟(Ae,Tp)\text{Hom}_{\mathcal{D}}(A^{e},T_{p}). We next use the Eckmann-Shapiro lemma (Lemma 5.3) which implies that Ext𝒟(A,Tp)ExtTpe(Tpe𝒟A,Tp)\text{Ext}_{\mathcal{D}}^{*}(A,T_{p})\cong\text{Ext}_{T_{p}^{e}}^{*}(T_{p}^{e}\otimes_{\mathcal{D}}A,T_{p}) and the isomorphism is given by

σ(fxigj)(xmgsxngr𝒟xkxl)\displaystyle\sigma(f_{x^{i}g^{j}})(x^{m}g^{s}\otimes x^{n}g^{r}\otimes_{\mathcal{D}}x^{k}\otimes x^{l}) =xmgsxngrfxigj(xkxl)=xmgsxngr(xk+l+igj)\displaystyle=x^{m}g^{s}\otimes x^{n}g^{r}f_{x^{i}g^{j}}(x^{k}\otimes x^{l})=x^{m}g^{s}\otimes x^{n}g^{r}(x^{k+l+i}g^{j})
=(xmgs)(xk+l+igj)(xngr)\displaystyle=(x^{m}g^{s})(x^{k+l+i}g^{j})(x^{n}g^{r})
=((xm(gsxk+l+i))gs+j)(xngr)\displaystyle=((x^{m}(^{g^{s}}x^{k+l+i}))g^{s+j})(x^{n}g^{r})
=ωs(k+l+i)(xm+k+l+igs+j)(xngr)\displaystyle=\omega^{s(k+l+i)}(x^{m+k+l+i}g^{s+j})(x^{n}g^{r})
=ωs(k+l+i)(xm+k+l+i(gs+jxn))gj+s+r\displaystyle=\omega^{s(k+l+i)}(x^{m+k+l+i}(^{g^{s+j}}x^{n}))g^{j+s+r}
=ωs(k+l+i+n)+jnxi+k+l+m+ngj+s+r.\displaystyle=\omega^{s(k+l+i+n)+jn}x^{i+k+l+m+n}g^{j+s+r}.

Hence, σ(fxigj)\sigma(f_{x^{i}g^{j}}) is in HomTpe(Tpe𝒟Ae,Tp)\text{Hom}_{T_{p}^{e}}(T_{p}^{e}\otimes_{\mathcal{D}}A^{e},T_{p}). Lastly, recall that Tpe𝒟AeATpT_{p}^{e}\otimes_{\mathcal{D}}A^{e}\cong A\otimes T_{p} via κ\kappa (4.2); so that,

κ(σ(fxigj))(xkxlgr)=σ(fxigj)((1Tpξgr)𝒟xkxl)=xi+k+lgj+r\displaystyle\kappa^{*}(\sigma(f_{x^{i}g^{j}}))(x^{k}\otimes x^{l}g^{r})=\sigma(f_{x^{i}g^{j}})((1_{T_{p}}\otimes\xi_{*}g^{r})\otimes_{\mathcal{D}}x^{k}\otimes x^{l})=x^{i+k+l}g^{j+r}

which implies κ(σ(fxigj))HomTpe(ATp,Tp)\kappa^{*}(\sigma(f_{x^{i}g^{j}}))\in\text{Hom}_{T_{p}^{e}}(A\otimes T_{p},T_{p}). For simplicity, we define f~xigj:=κ(σ(fxigj))\tilde{f}_{x^{i}g^{j}}:=\kappa^{*}(\sigma(f_{x^{i}g^{j}})).

The action of GG on TpT_{p} given by (4.12) and (4.14) depends on degree. Since TpGT_{p}^{G} is spanned by {1,g,,gp1}\{1,g,\cdots,g^{p-1}\} in even degrees and {x,xg,,xgp1}\{x,xg,\cdots,xg^{p-1}\} in odd degrees [8, Section 8.2], we have {f~1,f~g,,f~gp1}\{\tilde{f}_{1},\tilde{f}_{g},\cdots,\tilde{f}_{g^{p-1}}\} in even degrees and {f~x,f~xg,,f~xgp1}\{\tilde{f}_{x},\tilde{f}_{xg},\cdots,\tilde{f}_{xg^{p-1}}\} in the odd degrees as a basis of HomTpe(ATp,Tp)\text{Hom}_{T_{p}^{e}}(A\otimes T_{p},T_{p}).

We only calculate the bracket in degree 1 and 2 as before so we can extend it to higher degrees by the relation between cup product and the bracket. Since ATpAekGA\otimes T_{p}\cong A^{e}\otimes kG as vector spaces, ξi1G\xi_{i}1_{G} generates ATpA\otimes T_{p} as a TpT_{p}-bimodule. Through the calculation, idid represents idATpid_{A\otimes T_{p}} and \otimes represents Tp\otimes_{T_{p}}.

The circle product of two elements in degree one is

(f~xgiϕ~f~xgj)(ξ11G)\displaystyle(\tilde{f}_{xg^{i}}\circ_{\tilde{\phi}}\tilde{f}_{xg^{j}})(\xi_{1}1_{G}) =f~xgiϕ~(idf~xgjid)Δ~(2)(ξ11G)\displaystyle=\tilde{f}_{xg^{i}}\tilde{\phi}(id\otimes\tilde{f}_{xg^{j}}\otimes id)\tilde{\Delta}^{(2)}(\xi_{1}1_{G})
=f~xgiϕ~(idf~xgjid)(ξ01Gξ01Gξ11G+ξ01Gξ11Gξ01G\displaystyle=\tilde{f}_{xg^{i}}\tilde{\phi}(id\otimes\tilde{f}_{xg^{j}}\otimes id)(\xi_{0}1_{G}\otimes\xi_{0}1_{G}\otimes\xi_{1}1_{G}+\xi_{0}1_{G}\otimes\xi_{1}1_{G}\otimes\xi_{0}1_{G}
+ξ11Gξ01Gξ01G)\displaystyle+\xi_{1}1_{G}\otimes\xi_{0}1_{G}\otimes\xi_{0}1_{G})
=f~xgiϕ~(ξ01Gxξ0gj)=f~xgi(ξ1gj)=xgi+j.\displaystyle=\tilde{f}_{xg^{i}}\tilde{\phi}(\xi_{0}1_{G}\otimes x\xi_{0}g^{j})=\tilde{f}_{xg^{i}}(\xi_{1}g^{j})=xg^{i+j}.

Because of the symmetry, (f~xgjϕ~f~xgi)(ξ11G)=xgi+j(\tilde{f}_{xg^{j}}\circ_{\tilde{\phi}}\tilde{f}_{xg^{i}})(\xi_{1}1_{G})=xg^{i+j}. Therefore

[f~xgi,f~xgj](ξ11G)=xgi+j(1)0xgi+j=0.[\tilde{f}_{xg^{i}},\tilde{f}_{xg^{j}}](\xi_{1}1_{G})=xg^{i+j}-(-1)^{0}xg^{i+j}=0.

The circle product of the elements of degrees 1 and 2:

(f~xgiϕ~f~gj)(ξ21G)\displaystyle(\tilde{f}_{xg^{i}}\circ_{\tilde{\phi}}\tilde{f}_{g^{j}})(\xi_{2}1_{G}) =f~xgiϕ~(idf~gjid)Δ~(2)(ξ21G)=f~xgiϕ~(idf~gjid)\displaystyle=\tilde{f}_{xg^{i}}\tilde{\phi}(id\otimes\tilde{f}_{g^{j}}\otimes id)\tilde{\Delta}^{(2)}(\xi_{2}1_{G})=\tilde{f}_{xg^{i}}\tilde{\phi}(id\otimes\tilde{f}_{g^{j}}\otimes id)
(ξ01Gξ01Gξ21G+ξ01Gξ21Gξ01G\displaystyle(\xi_{0}1_{G}\otimes\xi_{0}1_{G}\otimes\xi_{2}1_{G}+\xi_{0}1_{G}\otimes\xi_{2}1_{G}\otimes\xi_{0}1_{G}
+ξ01Ga+b+c=p2(xaξ11Gxbξ1xc1G)+ξ21Gξ01Gξ01G\displaystyle+\xi_{0}1_{G}\otimes\sum_{\begin{subarray}{c}a+b+c\\ =p-2\end{subarray}}(x^{a}\xi_{1}1_{G}\otimes x^{b}\xi_{1}x^{c}1_{G})+\xi_{2}1_{G}\otimes\xi_{0}1_{G}\otimes\xi_{0}1_{G}
+a+b+c=p2(xaξ11G(xbξ01Gξ1xc1G+xbξ11Gξ0xc1G)))\displaystyle+\sum_{\begin{subarray}{c}a+b+c\\ =p-2\end{subarray}}(x^{a}\xi_{1}1_{G}\otimes(x^{b}\xi_{0}1_{G}\otimes\xi_{1}x^{c}1_{G}+x^{b}\xi_{1}1_{G}\otimes\xi_{0}x^{c}1_{G})))
=f~xgiϕ~(ξ01Gξ0gj)=0.\displaystyle=\tilde{f}_{xg^{i}}\tilde{\phi}(\xi_{0}1_{G}\otimes\xi_{0}g^{j})=0.

And the circle product on the reverse order:

(f~gjϕ~f~xgi)(ξ21G)\displaystyle(\tilde{f}_{g^{j}}\circ_{\tilde{\phi}}\tilde{f}_{xg^{i}})(\xi_{2}1_{G}) =f~gjϕ~(idf~xgiid)Δ~(2)(ξ21G)=f~gjϕ~(idf~xgiid)\displaystyle=\tilde{f}_{g^{j}}\tilde{\phi}(id\otimes\tilde{f}_{xg^{i}}\otimes id)\tilde{\Delta}^{(2)}(\xi_{2}1_{G})=\tilde{f}_{g^{j}}\tilde{\phi}(id\otimes\tilde{f}_{xg^{i}}\otimes id)
(ξ01Gξ01Gξ21G+ξ01Gξ21Gξ01G\displaystyle(\xi_{0}1_{G}\otimes\xi_{0}1_{G}\otimes\xi_{2}1_{G}+\xi_{0}1_{G}\otimes\xi_{2}1_{G}\otimes\xi_{0}1_{G}
+ξ01Ga+b+c=p2(xaξ11Gxbξ1xc1G)+ξ21Gξ01Gξ01G\displaystyle+\xi_{0}1_{G}\otimes\sum_{\begin{subarray}{c}a+b+c\\ =p-2\end{subarray}}(x^{a}\xi_{1}1_{G}\otimes x^{b}\xi_{1}x^{c}1_{G})+\xi_{2}1_{G}\otimes\xi_{0}1_{G}\otimes\xi_{0}1_{G}
+a+b+c=p2(xaξ11G(xbξ01Gξ1xc1G+xbξ11Gξ0xc1G)))\displaystyle+\sum_{\begin{subarray}{c}a+b+c\\ =p-2\end{subarray}}(x^{a}\xi_{1}1_{G}\otimes(x^{b}\xi_{0}1_{G}\otimes\xi_{1}x^{c}1_{G}+x^{b}\xi_{1}1_{G}\otimes\xi_{0}x^{c}1_{G})))
=f~gjϕ~(a+b+c=p2ωi(b+c)ξ01Gxa+b+1ξ1xcgi+ωicxaξ11Gxb+1ξ0xcgi)\displaystyle=\tilde{f}_{g^{j}}\tilde{\phi}(\sum_{\begin{subarray}{c}a+b+c\\ =p-2\end{subarray}}\omega^{i(b+c)}\xi_{0}1_{G}\otimes x^{a+b+1}\xi_{1}x^{c}g^{i}+\omega^{ic}x^{a}\xi_{1}1_{G}\otimes x^{b+1}\xi_{0}x^{c}g^{i})
=f~gj(a+b+c=p2ωi(b+c)δa+b+1,p1xcξ2giωicδb+1,p1xa+cξ2gi)\displaystyle=\tilde{f}_{g^{j}}(\sum_{\begin{subarray}{c}a+b+c\\ =p-2\end{subarray}}\omega^{i(b+c)}\delta_{a+b+1,p-1}x^{c}\xi_{2}g^{i}-\omega^{ic}\delta_{b+1,p-1}x^{a+c}\xi_{2}g^{i})
=f~gj(b=0p2ωibξ2gi)f~gj(ξ2gi)\displaystyle=\tilde{f}_{g^{j}}(\sum_{b=0}^{p-2}\omega^{ib}\xi_{2}g^{i})-\tilde{f}_{g^{j}}(\xi_{2}g^{i})
={(p2)gj,for i=0(ωi+1)gi+j,for i0.\displaystyle=\left\{\begin{array}[]{lr}(p-2)g^{j},&\text{for }i=0\\ -(\omega^{-i}+1)g^{i+j},&\text{for }i\neq 0\end{array}\right..

Therefore, we obtain

[f~xgi,f~gj]={(p2)gj,for i=0(ωi+1)gi+j,for i0.[\tilde{f}_{xg^{i}},\tilde{f}_{g^{j}}]=\left\{\begin{array}[]{lr}-(p-2)g^{j},&\text{for }i=0\\ (\omega^{-i}+1)g^{i+j},&\text{for }i\neq 0\end{array}\right..

Lastly, the bracket of the elements of degrees 2 and 2:

(f~giϕ~f~gj)(ξ31G)\displaystyle(\tilde{f}_{g^{i}}\circ_{\tilde{\phi}}\tilde{f}_{g^{j}})(\xi_{3}1_{G}) =f~giϕ~(idf~gjid)Δ~(2)(ξ31G)=f~giϕ~(idf~gjid)\displaystyle=\tilde{f}_{g^{i}}\tilde{\phi}(id\otimes\tilde{f}_{g^{j}}\otimes id)\tilde{\Delta}^{(2)}(\xi_{3}1_{G})=\tilde{f}_{g^{i}}\tilde{\phi}(id\otimes\tilde{f}_{g^{j}}\otimes id)
(ξ01Gξ01Gξ31G+ξ01Gξ11Gξ21G+ξ01Gξ21Gξ11G\displaystyle(\xi_{0}1_{G}\otimes\xi_{0}1_{G}\otimes\xi_{3}1_{G}+\xi_{0}1_{G}\otimes\xi_{1}1_{G}\otimes\xi_{2}1_{G}+\xi_{0}1_{G}\otimes\xi_{2}1_{G}\otimes\xi_{1}1_{G}
+ξ01Gξ31Gξ01G+ξ11Gξ21Gξ01G+ξ11Gξ01Gξ21G\displaystyle+\xi_{0}1_{G}\otimes\xi_{3}1_{G}\otimes\xi_{0}1_{G}+\xi_{1}1_{G}\otimes\xi_{2}1_{G}\otimes\xi_{0}1_{G}+\xi_{1}1_{G}\otimes\xi_{0}1_{G}\otimes\xi_{2}1_{G}
+ξ21Gξ11Gξ01G+ξ21Gξ01Gξ11G+ξ31Gξ01Gξ01G)\displaystyle+\xi_{2}1_{G}\otimes\xi_{1}1_{G}\otimes\xi_{0}1_{G}+\xi_{2}1_{G}\otimes\xi_{0}1_{G}\otimes\xi_{1}1_{G}+\xi_{3}1_{G}\otimes\xi_{0}1_{G}\otimes\xi_{0}1_{G})
=f~giϕ~(ξ01Gξ1gj+ξ11Gξ0gj)=0\displaystyle=\tilde{f}_{g^{i}}\tilde{\phi}(\xi_{0}1_{G}\otimes\xi_{1}g^{j}+\xi_{1}1_{G}\otimes\xi_{0}g^{j})=0

and by symmetry (f~gjϕ~f~gi)(ξ31G)=0(\tilde{f}_{g^{j}}\circ_{\tilde{\phi}}\tilde{f}_{g^{i}})(\xi_{3}1_{G})=0. Therefore, we have [f~gi,f~gj]=0[\tilde{f}_{g^{i}},\tilde{f}_{g^{j}}]=0. As a consequence, the bracket for the elements of degree 1 and 2 are

[f~xgi,f~xgj]=0,[f~xgi,f~gj]={(p2)gj,for i=0(ωi+1)gi+j,for i0,[f~gi,f~gj]=0.[\tilde{f}_{xg^{i}},\tilde{f}_{xg^{j}}]=0,[\tilde{f}_{xg^{i}},\tilde{f}_{g^{j}}]=\left\{\begin{array}[]{lr}-(p-2)g^{j},&\text{for }i=0\\ (\omega^{-i}+1)g^{i+j},&\text{for }i\neq 0\end{array}\right.,[\tilde{f}_{g^{i}},\tilde{f}_{g^{j}}]=0.

By the identity (2.3), brackets in higher degrees can be determined, since the Hochschild cohomology is generated as an algebra under cup product in degrees 1 and 2.

Hopf algebra cohomology of TpT_{p} and Hochschild cohomology of TpT_{p} were calculated before by V. C. Nguyen [8, Section 8] as the Hopf algebra cohomology

Hn(Tp,k)={kif n is even,0if n is odd,\text{H}^{n}(T_{p},k)=\begin{cases}k&\text{if $n$ is even,}\\ 0&\text{if $n$ is odd,}\end{cases}

and the Hochschild cohomology

HHn(Tp,k)={kif n is even,Spank{x}if n is odd.\text{HH}^{n}(T_{p},k)=\begin{cases}k&\text{if $n$ is even,}\\ Span_{k}\{x\}&\text{if $n$ is odd.}\end{cases}

It is known that for any Hopf algebra with bijective antipode, the Hopf algebra cohomology can be embedded into the Hochschild cohomology [14, Theorem 9.4.5 and Corollary 9.4.7]. Since any finite dimensional Hopf algebra has a bijective antipode, the Taft algebra TpT_{p} is also a Hopf algebra with a bijective antipode. The embedding of Hn(Tp,k)\text{H}^{n}(T_{p},k) into HHn(Tp,Tp)\text{HH}^{n}(T_{p},T_{p}) turns out to be the map that is identity in even degrees and zero on odd degrees. Then, the corresponding bracket in Hopf algebra cohomology is

[f~gi,f~gj]=0,[\tilde{f}_{g^{i}},\tilde{f}_{g^{j}}]=0,

so that, the bracket on Hopf algebra cohomology for the elements of all degrees is 0 by the identity (2.3).

This is the first example of the Gerstenhaber bracket on the Hopf algebra cohomology of a nonquasi-triangular Hopf algebra and our calculation shows that the bracket on Hopf algebra cohomology of a Taft algebra is zero as it is on the Hopf algebra cohomology of any quasi-triangular algebra. A natural question that arises whether the bracket structure on the Hopf algebra cohomology is always trivial. In the next section, we explore a general expression for the bracket on the Hopf algebra cohomology that may help us to approach this question with a more theoretical perspective in the future researches.

5 Gerstenhaber bracket for Hopf algebras

In this section, we want to explore an expression for Gerstenhaber bracket on a Hopf algebra AA with a bijective antipode SS.

We give the following lemma which helps us to define the Gerstenhaber bracket on an equivalent resolution to the bar resolution of AA as an AA-bimodule.

Lemma 5.1.

Let AA be a Hopf algebra with bijective antipode. Let PP_{\bullet} be the bar resolution of kk as a left AA-module:

P:d3A3d2A2d1Aεk0,P_{\bullet}:\cdots\stackrel{{\scriptstyle d_{3}}}{{\longrightarrow}}A^{\otimes 3}\stackrel{{\scriptstyle d_{2}}}{{\longrightarrow}}A^{\otimes 2}\stackrel{{\scriptstyle d_{1}}}{{\longrightarrow}}A\stackrel{{\scriptstyle\varepsilon}}{{\longrightarrow}}k\longrightarrow 0,

with differentials

dn(a0a1an)=i=0n1(1)ia0a1aiai+1an+(1)nε(an)a0an1d_{n}(a_{0}\otimes a_{1}\otimes\cdots\otimes a_{n})=\sum_{i=0}^{n-1}(-1)^{i}a_{0}\otimes a_{1}\otimes\cdots\otimes a_{i}a_{i+1}\otimes\cdots\otimes a_{n}+(-1)^{n}\varepsilon(a_{n})a_{0}\otimes\cdots\otimes a_{n-1}

Then X=AeAPX_{\bullet}=A^{e}\otimes_{A}P_{\bullet} is equivalent to the bar resolution of AA as an AA-bimodule.

Proof.

Since SS is bijective [14, Lemma 9.2.9], AeA^{e} is projective as a right AA-module. Also there is an AeA^{e}-module isomorphism ρ:AAeAk\rho:A\to A^{e}\otimes_{A}k defined by ρ(a)=a11\rho(a)=a\otimes 1\otimes 1 for all aAa\in A [14, Lemma 9.4.2].

For each nn, define θn:XnA(n+2)\theta_{n}:X_{n}\to A^{\otimes(n+2)} by

θn((ab)A(1c1c2cn))=ac11c12c1nS(c21c22c2n)b\theta_{n}((a\otimes b)\otimes_{A}(1\otimes c^{1}\otimes c^{2}\otimes\cdots\otimes c^{n}))=\sum a\otimes c_{1}^{1}\otimes c_{1}^{2}\otimes\cdots\otimes c_{1}^{n}\otimes S(c_{2}^{1}c_{2}^{2}\cdots c_{2}^{n})b

for all a,b,c1,cnA.a,b,c^{1},\cdots c^{n}\in A.

Now, we show that θ\theta is a chain map:

θ\displaystyle\theta dnn1((ab)A(1c1c2cn)){}_{n-1}d_{n}((a\otimes b)\otimes_{A}(1\otimes c^{1}\otimes c^{2}\otimes\cdots\otimes c^{n}))
=\displaystyle= θn1((ab)A(c1c2cn)\displaystyle\theta_{n-1}((a\otimes b)\otimes_{A}(c^{1}\otimes c^{2}\otimes\cdots\otimes c^{n})
+i=1n1(1)i(ab)A(1c1c2cici+1cn)\displaystyle+\sum_{i=1}^{n-1}(-1)^{i}(a\otimes b)\otimes_{A}(1\otimes c^{1}\otimes c^{2}\otimes\cdots\otimes c^{i}c^{i+1}\otimes\cdots\otimes c^{n})
+(1)n(ab)A(ε(cn)c1c2cn1))\displaystyle+(-1)^{n}(a\otimes b)\otimes_{A}(\varepsilon(c^{n})\otimes c^{1}\otimes c^{2}\otimes\cdots\otimes c^{n-1}))
=\displaystyle= θn1((ac11S(c21)b)A(1c2cn)\displaystyle\theta_{n-1}(\sum(ac_{1}^{1}\otimes S(c_{2}^{1})b)\otimes_{A}(1\otimes c^{2}\otimes\cdots\otimes c^{n})
+i=1n1(1)i(ab)A(1c1c2cici+1cn)\displaystyle+\sum_{i=1}^{n-1}(-1)^{i}(a\otimes b)\otimes_{A}(1\otimes c^{1}\otimes c^{2}\otimes\cdots\otimes c^{i}c^{i+1}\otimes\cdots\otimes c^{n})
+(1)n(ε(cn)ab)A(1c1c2cn1))\displaystyle+(-1)^{n}(\varepsilon(c^{n})a\otimes b)\otimes_{A}(1\otimes c^{1}\otimes c^{2}\otimes\cdots\otimes c^{n-1}))
=\displaystyle= ac11c12c1nS(c22c2n)S(c21)b\displaystyle\sum ac_{1}^{1}\otimes c_{1}^{2}\otimes\cdots\otimes c_{1}^{n}\otimes S(c_{2}^{2}\cdots c_{2}^{n})S(c_{2}^{1})b
+i=1n1(1)iac11c1ic1i+1c1nS(c21c2n)b\displaystyle+\sum_{i=1}^{n-1}(-1)^{i}\sum a\otimes c_{1}^{1}\otimes\cdots\otimes c_{1}^{i}c_{1}^{i+1}\otimes\cdots\otimes c_{1}^{n}\otimes S(c_{2}^{1}\cdots c_{2}^{n})b
+(1)nac11c1n1ε(cn)S(c21c2n1)b\displaystyle+\sum(-1)^{n}a\otimes c_{1}^{1}\otimes\cdots\otimes c_{1}^{n-1}\otimes\varepsilon(c^{n})S(c_{2}^{1}\cdots c_{2}^{n-1})b

and

dn\displaystyle d_{n} θn((ab)A(1c1c2cn))\displaystyle\theta_{n}((a\otimes b)\otimes_{A}(1\otimes c^{1}\otimes c^{2}\otimes\cdots\otimes c^{n}))
=\displaystyle= dn(ac11c12c1nS(c21c22c2n)b)\displaystyle d_{n}(\sum a\otimes c_{1}^{1}\otimes c_{1}^{2}\otimes\cdots\otimes c_{1}^{n}\otimes S(c_{2}^{1}c_{2}^{2}\cdots c_{2}^{n})b)
=\displaystyle= ac11c12c1nS(c21c22c2n)b\displaystyle\sum ac_{1}^{1}\otimes c_{1}^{2}\otimes\cdots\otimes c_{1}^{n}\otimes S(c_{2}^{1}c_{2}^{2}\cdots c_{2}^{n})b
+i=1n1(1)iac11c1ic1i+1c1nS(c21c2n)b\displaystyle+\sum\sum_{i=1}^{n-1}(-1)^{i}a\otimes c_{1}^{1}\otimes\cdots\otimes c_{1}^{i}c_{1}^{i+1}\otimes\cdots\otimes c_{1}^{n}\otimes S(c_{2}^{1}\cdots c_{2}^{n})b
+(1)nac11c1n1c1nS(c21c2n)b.\displaystyle+\sum(-1)^{n}a\otimes c_{1}^{1}\otimes\cdots\otimes c_{1}^{n-1}\otimes c_{1}^{n}S(c_{2}^{1}\cdots c_{2}^{n})b.

Since SS is an algebra anti-homomorphism that is convolution inverse to the identity map,

c1nS(c21c2n)=c1nS(c2n)S(c2n1)S(c21)=ε(cn)S(c21c2n1)\sum c_{1}^{n}S(c_{2}^{1}\cdots c_{2}^{n})=\sum c_{1}^{n}S(c_{2}^{n})S(c_{2}^{n-1})\cdots S(c_{2}^{1})=\sum\varepsilon(c^{n})S(c_{2}^{1}\cdots c_{2}^{n-1})

and

S(c22c2n)S(c21)=S(c21c22c2n)S(c_{2}^{2}\cdots c_{2}^{n})S(c_{2}^{1})=S(c_{2}^{1}c_{2}^{2}\cdots c_{2}^{n})

so that the two expressions are equal which follows θ\theta is a chain map.

Lastly, one can see that the AeA^{e}-module homomorphism

ψn(ac1c2cnb)=(ac21c22c2nb)A(1c11c12c1n)\psi_{n}(a\otimes c^{1}\otimes c^{2}\otimes\cdots c^{n}\otimes b)=\sum(a\otimes c_{2}^{1}c_{2}^{2}\cdots c_{2}^{n}b)\otimes_{A}(1\otimes c_{1}^{1}\otimes c_{1}^{2}\otimes\cdots\otimes c_{1}^{n})

is the inverse of θn\theta_{n} by using the property that SS is an algebra anti-homomorphism that is convolution inverse to the identity map. ∎

Let fxf_{x}\inHom(Xm,A)Ae{}_{A^{e}}(X_{m},A) and gxg_{x}\inHom(Xn,A)Ae{}_{A^{e}}(X_{n},A). Then we define the XX-bracket [fx,gx]X[f_{x},g_{x}]_{X}\inHom(Xm+n1,A)Ae{}_{A^{e}}(X_{m+n-1},A) to be a composition XθB(A)[ψfx.ψgx]AX\stackrel{{\scriptstyle\theta}}{{\longrightarrow}}B(A)\xrightarrow{[\psi^{*}f_{x}.\psi^{*}g_{x}]}A; so that, we have

[fx,gx]X=[ψfx,ψgx]θ=(ψfxψgx)θ(1)(m1)(n1)(ψgxψfx)θ[f_{x},g_{x}]_{X}=[\psi^{*}f_{x},\psi^{*}g_{x}]\theta=(\psi^{*}f_{x}\circ\psi^{*}g_{x})\theta-(-1)^{(m-1)(n-1)}(\psi^{*}g_{x}\circ\psi^{*}f_{x})\theta

where

(ψ\displaystyle(\psi fxψgx)θm+n1((ab)A1c1cm+n1){}^{*}f_{x}\circ\psi^{*}g_{x})\theta_{m+n-1}((a\otimes b)\otimes_{A}1\otimes c^{1}\otimes\cdots\otimes c^{m+n-1})
=\displaystyle= (ψfxψgx)(ac11c12c1m+n1S(c21c22c2m+n1)b)\displaystyle(\psi^{*}f_{x}\circ\psi^{*}g_{x})(\sum a\otimes c_{1}^{1}\otimes c_{1}^{2}\otimes\cdots\otimes c_{1}^{m+n-1}\otimes S(c_{2}^{1}c_{2}^{2}\cdots c_{2}^{m+n-1})b)
=\displaystyle= i=1m(1)(n1)(i1)fxψm(ac11c1i1gxψn(1c1ic1i+n11)\displaystyle\sum\sum_{i=1}^{m}(-1)^{(n-1)(i-1)}f_{x}\psi_{m}(a\otimes c_{1}^{1}\otimes\cdots\otimes c_{1}^{i-1}\otimes g_{x}\psi_{n}(1\otimes c_{1}^{i}\otimes\cdots\otimes c_{1}^{i+n-1}\otimes 1)
c1i+nc1m+n1S(c21c22c2m+n1)b)\displaystyle\otimes c_{1}^{i+n}\otimes\cdots\otimes c_{1}^{m+n-1}\otimes S(c_{2}^{1}c_{2}^{2}\cdots c_{2}^{m+n-1})b)
=\displaystyle= i=1m(1)(n1)(i1)fxψm(ac11c1i1\displaystyle\sum\sum_{i=1}^{m}(-1)^{(n-1)(i-1)}f_{x}\psi_{m}(a\otimes c_{1}^{1}\otimes\cdots\otimes c_{1}^{i-1}
gx(1c2ic2i+1c2i+n1A1c1ic1i+1c1i+n1)\displaystyle\otimes\sum g_{x}(1\otimes c_{2}^{i}c_{2}^{i+1}\cdots c_{2}^{i+n-1}\otimes_{A}1\otimes c_{1}^{i}\otimes c_{1}^{i+1}\otimes\cdots\otimes c_{1}^{i+n-1})
c1i+nc1m+n1S(c21c2i1c3ic3i+n1c2i+nc2m+n1)b)\displaystyle\otimes c_{1}^{i+n}\otimes\cdots\otimes c_{1}^{m+n-1}\otimes S(c_{2}^{1}\cdots c_{2}^{i-1}c_{3}^{i}\cdots c_{3}^{i+n-1}c_{2}^{i+n}\cdots c_{2}^{m+n-1})b)
=\displaystyle= i=1m(1)(n1)(i1)fx(ac21c22c2i1c2c2i+nc2m+n1S(c31c32c3m+n1)b\displaystyle\sum\sum_{i=1}^{m}(-1)^{(n-1)(i-1)}f_{x}(a\otimes c_{2}^{1}c_{2}^{2}\cdots c_{2}^{i-1}c_{2}^{*}c_{2}^{i+n}\cdots c_{2}^{m+n-1}S(c_{3}^{1}c_{3}^{2}\cdots c_{3}^{m+n-1})b
A1c11c12c1i1c1c1i+nc1m+n1)\displaystyle\otimes_{A}1\otimes c_{1}^{1}\otimes c_{1}^{2}\otimes\cdots\otimes c_{1}^{i-1}\otimes c_{1}^{*}\otimes c_{1}^{i+n}\otimes\cdots\otimes c_{1}^{m+n-1})

where

Δ(c)\displaystyle\Delta(c) =c1c2,Δ(2)(c)=c1c2c3,Δ(c)=c1c2 and\displaystyle=\sum c_{1}\otimes c_{2},\Delta^{(2)}(c)=\sum c_{1}\otimes c_{2}\otimes c_{3},\Delta(c^{*})=\sum c_{1}^{*}\otimes c_{2}^{*}\text{ and }
c\displaystyle c^{*} =gx(1c2ic2i+1c2i+n1A1c1ic1i+1c1i+n1).\displaystyle=\sum g_{x}(1\otimes c_{2}^{i}c_{2}^{i+1}\cdots c_{2}^{i+n-1}\otimes_{A}1\otimes c_{1}^{i}\otimes c_{1}^{i+1}\otimes\cdots\otimes c_{1}^{i+n-1}).

This is the general expression of the Gerstenhaber bracket on Hochschild cohomology of AA. Next, we start with the following theorem [14, Theorem 9.4.5] to construct an embedding from H(A,k){}^{*}(A,k) into HH(A){}^{*}(A).

Theorem 5.2.

Let AA be a Hopf algebra over k with bijective antipode. Then

HH(A)H(A,Aad).\operatorname{HH}^{*}(A)\cong\operatorname{H}^{*}(A,A^{ad}).

In this theorem AadA^{ad} is an AA-module AA under left adjoint action, given by
ab=a1bS(a2)a\cdot b=\sum a_{1}bS(a_{2}) for all a,bAa,b\in A. To find explicit isomorphism between HH(A){}^{*}(A) and H(A,Aad){}^{*}(A,A^{ad}), we give the Eckmann-Shapiro lemma.

Lemma 5.3 (Eckmann-Shapiro).

Let AA be a ring and let BB be a subring of AA such that A is projective as a right BB-module. Let MM be an AA-module and NN be a BB-module. Then

ExtBn(N,M)ExtAn(ABN,M).\emph{Ext}_{B}^{n}(N,M)\cong\emph{Ext}_{A}^{n}(A\otimes_{B}N,M).
Proof.

Let PNP_{\bullet}\to N be a BB projective resolution of NN. Then ABPnA\otimes_{B}P_{n} is projective as A-module so that ABPABNA\otimes_{B}P_{\bullet}\to A\otimes_{B}N is a projective resolution of ABNA\otimes_{B}N as an AA-module. Let

σ:HomB(Pn,M)HomA(ABPn,M) defined by σ(f)(aBp)=af(p),\sigma:\text{Hom}_{B}(P_{n},M)\to\text{Hom}_{A}(A\otimes_{B}P_{n},M)\text{ defined by }\sigma(f)(a\otimes_{B}p)=af(p),
τ:HomA(ABPn,M)HomB(Pn,M) defined by τ(g)(p)=g(1Bp)\tau:\text{Hom}_{A}(A\otimes_{B}P_{n},M)\to\text{Hom}_{B}(P_{n},M)\text{ defined by }\tau(g)(p)=g(1\otimes_{B}p)

where aA,pPn,fHomB(Pn,M),gHomA(ABPn,M).a\in A,p\in P_{n},f\in\text{Hom}_{B}(P_{n},M),g\in\text{Hom}_{A}(A\otimes_{B}P_{n},M). Since σ\sigma and τ\tau are inverse of each other and they are homomorphisms, HomA(ABPn,M)HomB(Pn,M)\text{Hom}_{A}(A\otimes_{B}P_{n},M)\cong\text{Hom}_{B}(P_{n},M). ∎

If we replace AA with AeA^{e}, BB with AA and take M=A,N=kM=A,N=k in the Eckmann-Shapiro lemma, we have the isomorphism Ext(AeAk,A)Aen{}_{A^{e}}^{n}(A^{e}\otimes_{A}k,A)\congExt(k,Aad)An{}_{A}^{n}(k,A^{ad}). We also know that AAeAkA\cong A^{e}\otimes_{A}k [14, Lemma 9.4.2] and the isomorphism is given by ρ(a)=a11\rho(a)=a\otimes 1\otimes 1 for all aAa\in A. Therefore Ext(A,A)Aen{}_{A^{e}}^{n}(A,A)\congExt(AeAk,A)Aen{}_{A^{e}}^{n}(A^{e}\otimes_{A}k,A)\congExt(k,Aad)An{}_{A}^{n}(k,A^{ad}).

We already have the Gerstenhaber bracket [,]X[,]_{X} on Ext(AeAk,A)Aen{}_{A^{e}}^{n}(A^{e}\otimes_{A}k,A). Hence we can use the isomorphisms σ\sigma and τ\tau in Eckmann-Shapiro Lemma and find the bracket expression on H(A,Aad){}^{*}(A,A^{ad}). Now let f~\tilde{f}\inHom(Pm,Aad)A{}_{A}(P_{m},A^{ad}) and g~\tilde{g}\inHom(Pn,Aad)A{}_{A}(P_{n},A^{ad}). Then [f~,g~]P[\tilde{f},\tilde{g}]_{P}\inHom(Pm+n1,Aad)A{}_{A}(P_{m+n-1},A^{ad}) and we have

[f~,g~]P\displaystyle[\tilde{f},\tilde{g}]_{P} =τ[σ(f~),σ(g~)]X\displaystyle=\tau[\sigma(\tilde{f}),\sigma(\tilde{g})]_{X}
=τ((ψ(σ(f~))ψ(σ(g~)))θ)(1)(m1)(n1)τ((ψ(σ(g~))ψ(σ(f~)))θ).\displaystyle=\tau((\psi^{*}(\sigma(\tilde{f}))\circ\psi^{*}(\sigma(\tilde{g})))\theta)-(-1)^{(m-1)(n-1)}\tau((\psi^{*}(\sigma(\tilde{g}))\circ\psi^{*}(\sigma(\tilde{f})))\theta).

For simplification we define

f~Pg~:=τ((ψ(σ(f~))ψ(σ(g~)))θ).\tilde{f}\circ_{P}\tilde{g}:=\tau((\psi^{*}(\sigma(\tilde{f}))\circ\psi^{*}(\sigma(\tilde{g})))\theta).

Then by using previous circle product formula we obtain:

f~\displaystyle\tilde{f} Pg~(1c1c2cm+n1)\displaystyle\circ_{P}\tilde{g}(1\otimes c^{1}\otimes c^{2}\otimes\cdots\otimes c^{m+n-1})
=\displaystyle= τ((ψ(σ(f~))ψ(σ(g~)))θ)(1c1c2cm+n1)\displaystyle\tau((\psi^{*}(\sigma(\tilde{f}))\circ\psi^{*}(\sigma(\tilde{g})))\theta)(1\otimes c^{1}\otimes c^{2}\otimes\cdots\otimes c^{m+n-1})
=\displaystyle= (ψ(σ(f~))ψ(σ(g~)))θ((11)A1c1c2cm+n1)\displaystyle(\psi^{*}(\sigma(\tilde{f}))\circ\psi^{*}(\sigma(\tilde{g})))\theta((1\otimes 1)\otimes_{A}1\otimes c^{1}\otimes c^{2}\otimes\cdots\otimes c^{m+n-1})
=\displaystyle= i=1m(1)(n1)(i1)σ(f~)(1c21c22c2i1c2c2i+nc2m+n1S(c31c32c3m+n1)\displaystyle\sum\sum_{i=1}^{m}(-1)^{(n-1)(i-1)}\sigma(\tilde{f})(1\otimes c_{2}^{1}c_{2}^{2}\cdots c_{2}^{i-1}c_{2}^{*}c_{2}^{i+n}\cdots c_{2}^{m+n-1}S(c_{3}^{1}c_{3}^{2}\cdots c_{3}^{m+n-1})
A1c11c12c1i1c1c1i+nc1m+n1)\displaystyle\otimes_{A}1\otimes c_{1}^{1}\otimes c_{1}^{2}\otimes\cdots\otimes c_{1}^{i-1}\otimes c_{1}^{*}\otimes c_{1}^{i+n}\otimes\cdots\otimes c_{1}^{m+n-1})
=\displaystyle= i=1m(1)(n1)(i1)f~(1c11c12c1i1c1c1i+nc1m+n1)\displaystyle\sum\sum_{i=1}^{m}(-1)^{(n-1)(i-1)}\tilde{f}(1\otimes c_{1}^{1}\otimes c_{1}^{2}\otimes\cdots\otimes c_{1}^{i-1}\otimes c_{1}^{*}\otimes c_{1}^{i+n}\otimes\cdots\otimes c_{1}^{m+n-1})
c21c22c2i1c2c2i+nc2m+n1S(c31c32c3m+n1))\displaystyle c_{2}^{1}c_{2}^{2}\cdots c_{2}^{i-1}c_{2}^{*}c_{2}^{i+n}\cdots c_{2}^{m+n-1}S(c_{3}^{1}c_{3}^{2}\cdots c_{3}^{m+n-1}))

with Δ(c)=c1c2\Delta(c^{*})=\sum c_{1}^{*}\otimes c_{2}^{*} and

c\displaystyle c^{*} =σ(g~)(1c2ic2i+1c2i+n1A1c1ic1i+1c1i+n1)\displaystyle=\sum\sigma(\tilde{g})(1\otimes c_{2}^{i}c_{2}^{i+1}\cdots c_{2}^{i+n-1}\otimes_{A}1\otimes c_{1}^{i}\otimes c_{1}^{i+1}\otimes\cdots\otimes c_{1}^{i+n-1})
=(1c2ic2i+1c2i+n1)g~(1c1ic1i+1c1i+n1)\displaystyle=\sum(1\otimes c_{2}^{i}c_{2}^{i+1}\cdots c_{2}^{i+n-1})\tilde{g}(1\otimes c_{1}^{i}\otimes c_{1}^{i+1}\otimes\cdots\otimes c_{1}^{i+n-1})
=g~(1c1ic1i+1c1i+n1)c2ic2i+1c2i+n1.\displaystyle=\sum\tilde{g}(1\otimes c_{1}^{i}\otimes c_{1}^{i+1}\otimes\cdots\otimes c_{1}^{i+n-1})c_{2}^{i}c_{2}^{i+1}\cdots c_{2}^{i+n-1}.

We now have the Lie bracket [,]P[,]_{P} on H(A,Aad){}^{*}(A,A^{ad}). Next, we embed H(A,k){}^{*}(A,k) into H(A,Aad){}^{*}(A,A^{ad}) [14, Corollary 9.4.7] via the unit map

η:HomA(P,k)HomA(P,Aad).\eta_{*}:\text{Hom}_{A}(P_{\bullet},k)\to\text{Hom}_{A}(P_{\bullet},A^{ad}).

Let fHomA(Pm,k)f\in\text{Hom}_{A}(P_{m},k) and gHomA(Pn,k)g\in\text{Hom}_{A}(P_{n},k). Then by using counit map

ε:HomA(P,A)HomA(P,k),\varepsilon_{*}:\text{Hom}_{A}(P_{\bullet},A)\to\text{Hom}_{A}(P_{\bullet},k),

η\eta_{*} and bracket on H(A,Aad){}^{*}(A,A^{ad}), we derive the formula for [f,g]HomA(Pm+n1,k)[f,g]\in\text{Hom}_{A}(P_{m+n-1},k):

[f,g]=ε[η(f),η(g)]P=ε(η(f)Pη(g))(1)(m1)(n1)ε(η(g)Pη(f))[f,g]=\varepsilon_{*}[\eta_{*}(f),\eta_{*}(g)]_{P}=\varepsilon_{*}(\eta_{*}(f)\circ_{P}\eta_{*}(g))-(-1)^{(m-1)(n-1)}\varepsilon_{*}(\eta_{*}(g)\circ_{P}\eta_{*}(f))

where

ε\displaystyle\varepsilon_{*} ((η(f)Pη(g))(1c1c2cm+n1))\displaystyle((\eta_{*}(f)\circ_{P}\eta_{*}(g))(1\otimes c^{1}\otimes c^{2}\otimes\cdots\otimes c^{m+n-1}))
=\displaystyle= ε(i=1m(1)(n1)(i1)η(f(1c11c12c1i1c1c1i+nc1m+n1))\displaystyle\varepsilon(\sum\sum_{i=1}^{m}(-1)^{(n-1)(i-1)}\eta(f(1\otimes c_{1}^{1}\otimes c_{1}^{2}\otimes\cdots\otimes c_{1}^{i-1}\otimes c_{1}^{*}\otimes c_{1}^{i+n}\otimes\cdots\otimes c_{1}^{m+n-1}))
c21c22c2i1c2c2i+nc2m+n1S(c31c32c3m+n1))\displaystyle c_{2}^{1}c_{2}^{2}\cdots c_{2}^{i-1}c_{2}^{*}c_{2}^{i+n}\cdots c_{2}^{m+n-1}S(c_{3}^{1}c_{3}^{2}\cdots c_{3}^{m+n-1}))

with

Δ(c)\displaystyle\Delta(c^{*}) =c1c2 and\displaystyle=\sum c_{1}^{*}\otimes c_{2}^{*}\text{ and }
c\displaystyle c^{*} =η(g(1c1ic1i+1c1i+n1))c2ic2i+1c2i+n1.\displaystyle=\sum\eta(g(1\otimes c_{1}^{i}\otimes c_{1}^{i+1}\otimes\cdots\otimes c_{1}^{i+n-1}))c_{2}^{i}c_{2}^{i+1}\cdots c_{2}^{i+n-1}.

Therefore, the last formula is a general expression of the Gerstenhaber bracket on a Hopf algebra cohomology which is indeed inherited from the formula of the bracket on Hochschild cohomology.

Acknowledgement

The author would like to thank S. Witherspoon for her precious time, suggestions and support.

References

  • [1] S. M. Burciu and S. Witherspoon, Hochschild cohomology of smash products and rank one Hopf algebras, Biblioteca de la Revista Matematica Iberoamericana Actas del ”XVI Coloquio Latinoamericano de Algebra” (2005), 153-170.
  • [2] M.A. Farinati, A. Solotar, G-structure on the cohomology of Hopf algebras, Proceedings of the American Mathematical Society 132 (2004), no. 10, 2859–2865.
  • [3] M. Gerstenhaber, The cohomology structure of an associative ring, Ann. of Math. (2) (1963), 78:267-288.
  • [4] L. Grimley, V. C. Nguyen and S. Witherspoon, Gerstenhaber brackets on Hochschild cohomology of twisted tensor products, J. Noncommutative Geometry 11 (2017), no. 4, 1351–1379.
  • [5] R. Hermann, Monoidal Categories and the Gerstenhaber Bracket in Hochschild Cohomology, American Mathematical Society, 2016.
  • [6] G. Hochschild, On the cohomology groups of an associative algebra, Ann. Math. (2) (1945), 58-67.
  • [7] C. Negron and S. Witherspoon, An alternate approach to the Lie bracket on Hochschild cohomology, Homology, Homotopy and Applications 18 (2016), no.1, 265-285.
  • [8] V. C. Nguyen, Tate and Tate-Hochschild cohomology for finite dimensional Hopf algebras, J. Pure Appl. Algebra, 217 (2013), 1967-1979.
  • [9] S. Sánchez-Flores, The Lie structure on the Hochschild cohomology of a modular group algebra, J. Pure Appl. Algebra 216(3) (2012),718–733.
  • [10] D. Ştefan, Hochschild cohomology on Hopf Galois extensions, J. Pure Appl. Algebra, 103 (1995), pp. 221–233.
  • [11] R. Taillefer, Injective Hopf bimodules, Cohomologies of Infinite Dimensional Hopf Algebras and Graded-commutativity of the Yoneda Product, Journal of Algebra. 276 (2004), no.1, 259-279.
  • [12] Y. Volkov, Gerstenhaber bracket on the Hochschild cohomology via an arbitrary resolution, Proc. Edinburgh Math. Soc. (2) 62 (3) (2019), 817-836.
  • [13] Y. Volkov, S. Witherspoon, Graded Lie structure on cohomology of some exact monoidal categories, arXiv:2004.06225.
  • [14] S. Witherspoon, Hochschild Cohomology for Algebras, volume 204. American Mathematical Society, 2019.

Department of Mathematics, Texas A&M University, College Station, Texas 77843, USA

E-mail address: [email protected]