Secondary Hochschild homology and differentials

Since A is commutative, the map \partial_1:A^{\otimes2}\otimes B\longrightarrow A is trivial. Therefore we have that HH_1(A,B,\varepsilon) is the quotient of A^{\otimes2}\otimes B by the relation

The map HH_1(A,B,\varepsilon)\longrightarrow\Omega_{\mathcal{T}|\mathbbm{k}}^1 sends the class of \otimes\begin{pmatrix} a & \alpha\\ 1 & b\\ \end{pmatrix} to ad(\alpha\otimes b). This map is well-defined because (Proof) maps to

In the other direction, the map \Omega_{\mathcal{T}|\mathbbm{k}}^1\longrightarrow HH_1(A,B,\varepsilon) sends ad(\alpha\otimes b) to the class of \otimes\begin{pmatrix} a & \alpha\\ 1 & b\\ \end{pmatrix}, which is indeed a cycle because ab\varepsilon(\alpha)-ba\varepsilon(\alpha)=0, relying on A being commutative. This is in fact a module homomorphism and we will now verify it respects the relations from Definition 3.1.

Observe that (?) is immediate.

For (?), notice that d(\alpha\beta\otimes bc)-b\varepsilon(\alpha)d(\beta\otimes c)-c\varepsilon(\beta)d(\alpha\otimes b) maps to

which is 0 by taking (Proof) with \gamma=1_B and a=1_A.

Finally for (?), we note that d(\gamma\otimes1)+d(\gamma\otimes1)-d(1\otimes\varepsilon(\gamma)) maps to

which is 0 from (Proof) when we take \alpha=\beta=1_B and a=b=c=1_A.

To complete the proof, we simply note that the two maps defined above are inverses of each other. That is,

and

The isomorphism follows.

First note that HC_0(A,B,\varepsilon)= HH_0(A,B,\varepsilon) by Example 2.7, and then by Example 2.4 and since A is commutative, we have that HH_0(A,B,\varepsilon)=\frac{A}{[A,A]}=A. Hence HC_0(A,B,\varepsilon)=A. Therefore, employing Theorem 2.8 in low dimension, we get that

Here B:A\longrightarrow A^{\otimes2}\otimes B is defined by

Then by exactness,

Finally we observe that the image of the map B has the form d(1\otimes a)+ad(1\otimes1)=d(1\otimes a), and this is exactly d(1\otimes A), which was what we wanted.

We first define the map \Omega_{\mathcal{T}|\mathbbm{k}}^1\longrightarrow J/(J^2+\hat{J}) by sending ad(\alpha\otimes b) to the class of \otimes\begin{pmatrix} a & \alpha\\ 1 & b\\ \end{pmatrix}-\otimes\begin{pmatrix} a\varepsilon(\alpha)b & 1\\ 1 & 1\\ \end{pmatrix}. Notice that this element is indeed in J. Next we aim to show that the map is well-defined and respects the relations found in Definition 3.1.

Of course, (?) is clear.

For (?), notice that d((\alpha\beta\otimes ab))-a\varepsilon(\alpha)d(\beta\otimes b)-b\varepsilon(\beta)d(\alpha\otimes a) maps to

Finally, d(\alpha\otimes1)+d(\alpha\otimes1)-d(1\otimes\varepsilon(\alpha)) maps to

and hence (?) is satisfied.

Next we define the map J/(J^2+\hat{J})\longrightarrow\Omega_{\mathcal{T}|\mathbbm{k}}^1 by sending the class of \otimes\begin{pmatrix} 1 & \alpha\\ 1 & a\\ \end{pmatrix}-\otimes\begin{pmatrix} a\varepsilon(\alpha) & 1\\ 1 & 1\\ \end{pmatrix} to 1d(\alpha\otimes a)-a\varepsilon(\alpha)d(1\otimes1)=d(\alpha\otimes a). Notice this is well-defined because elements in J^2, which are of the form

map to

by (?) from Definition 3.1 and Remark 3.2. Likewise, elements in \hat{J}, which are of the form

map to

by (?) from Definition 3.1 and Remark 3.2.

To conclude, we notice that these maps are inverses of each other. That is,

and

The isomorphism follows.

Unite Proposition 3.4 and Proposition 4.2.