ScalingStacks

5. Bimodule 22-representations

5.1. Differential algebras

5.1.1. 22-representations

Let AA be a differential algebra.

0P6T

Definition 5.1.1. A 22-representation on AA is the data of a differential (A,A)(A,A)-bimodule EE and of an endomorphism τ\tau of the (A,A)(A,A)-bimodule E⊗AEE\otimes_{A}E such that

τ2=0,d⁡(τ)=id⁡ and ​(E⊗τ)∘(τ⊗E)∘(E⊗τ)=(τ⊗E)∘(E⊗τ)∘(τ⊗E).\tau^{2}=0,\ d(\tau)=\operatorname{id}\nolimits\text{ and }(E\otimes\tau)\circ(\tau\otimes E)\circ(E\otimes\tau)=(\tau\otimes E)\circ(E\otimes\tau)\circ(\tau\otimes E).

We say that the 22-representation is right finite if EE is finitely generated and projective as a (non-differential) AoppA^{\operatorname{opp}\nolimits}-module.

Consider a 22-representation on AA. Note that E⊗A−E\otimes_{A}- is a differential endofunctor of A​−diffA\operatorname{\!-diff}\nolimits, and τ\tau defines an endomorphism of (E⊗A−)2(E\otimes_{A}-)^{2}. This gives a structure of 22-representation on A​−diffA\operatorname{\!-diff}\nolimits. It restricts to a 22-representation on (A¯)i(\bar{A})^{i} if EE is strictly perfect as a differential AA-module.

Note that there is a morphism of differential algebras

Hn→EndA⊗Aopp⁡(En),Ti↦En−i−1⊗τ⊗Ei−1.H_{n}\to\operatorname{End}\nolimits_{A\otimes A^{\operatorname{opp}\nolimits}}(E^{n}),\ T_{i}\mapsto E^{n-i-1}\otimes\tau\otimes E^{i-1}.

Let A′A^{\prime} be another differential algebra with a 22-representation (E′,τ′)(E^{\prime},\tau^{\prime}). We define a morphism of 22-representations from (A,E,τ)(A,E,\tau) to (A′,E′,τ′)(A^{\prime},E^{\prime},\tau^{\prime}) to be an (A′,A)(A^{\prime},A)-bimodule PP together with a closed isomorphism of (A′,A)(A^{\prime},A)-bimodules φ:P⊗AE→∼E′⊗A′P\varphi:P\otimes_{A}E\mathrel{\mathop{\kern 0.0pt\to}\limits^{\sim}}E^{\prime}\otimes_{A^{\prime}}P such that

(5.1.1) τ′​P∘E′​φ∘φ​E=E′​φ∘φ​E∘P​τ:P​E2→E′2​P.\tau^{\prime}P\circ E^{\prime}\varphi\circ\varphi E=E^{\prime}\varphi\circ\varphi E\circ P\tau:PE^{2}\to E^{\prime 2}P.

Note that such a pair (P,φ)(P,\varphi) gives rise to a morphism of 22-representations (P⊗A−,φ):(A−diff,E⊗A−,τ)→(A′−diff,E′⊗A′−,τ′)(P\otimes_{A}-,\varphi):(A\operatorname{\!-diff}\nolimits,E\otimes_{A}-,\tau)\to(A^{\prime}\operatorname{\!-diff}\nolimits,E^{\prime}\otimes_{A^{\prime}}-,\tau^{\prime}).

We obtain a differential 22-category of 22-representations on differential algebras.

The opposite 22-representation is the data (A′,E′,τ′)(A^{\prime},E^{\prime},\tau^{\prime}) where A′=AoppA^{\prime}=A^{\operatorname{opp}\nolimits}, E′=EE^{\prime}=E and τ′=τ\tau^{\prime}=\tau. Note that (A,E,τ)(A,E,\tau) coincides with its double dual.

Assume now the 22-representation is right finite. We have two morphisms of (A,A)(A,A)-bimodules η:A→E⊗AE∨\eta:A\to E\otimes_{A}E^{\vee} and ε:E∨⊗AE→A\varepsilon:E^{\vee}\otimes_{A}E\to A (unit and counit of adjunction). We have a morphism of (A,A)(A,A)-bimodules ρ:E∨​E→E​E∨\rho:E^{\vee}E\to EE^{\vee} defined as the composition

ρ:E∨​E→∙ηE∨​E​E​E∨→E∨​τ​E∨E∨​E​E​E∨→ε∙E​E∨.\rho:E^{\vee}E\xrightarrow{\bullet\eta}E^{\vee}EEE^{\vee}\xrightarrow{E^{\vee}\tau E^{\vee}}E^{\vee}EEE^{\vee}\xrightarrow{\varepsilon\bullet}EE^{\vee}.

There is a canonical isomorphism of differential algebras End⁡(E2)opp→∼End⁡((E∨)2)\operatorname{End}\nolimits(E^{2})^{\operatorname{opp}\nolimits}\mathrel{\mathop{\kern 0.0pt\to}\limits^{\sim}}\operatorname{End}\nolimits((E^{\vee})^{2}) and we still denote by τ\tau the endomorphism of (E∨)2(E^{\vee})^{2} corresponding to τ\tau.

We define the left dual 22-representation on AA with the bimodule E∨E^{\vee} and the endomorphism τ\tau.

5.2. Lax cocenter

Let BB be a differential algebra. A lax bi-22-representation on BB is the data of

  • •

    differential (B,B)(B,B)-bimodules Ei,jE_{i,j} for i,j≥0i,j\geq 0

  • •

    morphisms of differential algebras Hi⊗Hj→End⁡(Ei,j)H_{i}\otimes H_{j}\to\operatorname{End}\nolimits(E_{i,j})

  • •

    morphisms μ(i,j),(i′,j′):Ei,j​Ei′,j′→Ei+i′,j+j′\mu_{(i,j),(i^{\prime},j^{\prime})}:E_{i,j}E_{i^{\prime},j^{\prime}}\to E_{i+i^{\prime},j+j^{\prime}} satisfying properties (1) and (2) of §4.2.1.

Consider a lax bi-22-representation EE. Note that the functors (Ei,j⊗B−)(E_{i,j}\otimes_{B}-) provide a structure of lax bi-22-representation on B​−diffB\operatorname{\!-diff}\nolimits.

We define the differential algebra A=ΔE​(B)A=\Delta_{E}(B) as the quotient of the tensor algebra TB​(E0,1​E1,0)T_{B}(E_{0,1}E_{1,0}) by the two-sided ideal generated by ⨁i≥0Ki\bigoplus_{i\geq 0}K_{i}, where KiK_{i} is the kernel of the composition

(E0,1​E1,0)i→canEi,i→canEi,i/((Tr⊗1)​x−(1⊗Tr)​x)x∈Ei,i, 1≤r<i.(E_{0,1}E_{1,0})^{i}\xrightarrow{{\mathrm{can}}}E_{i,i}\xrightarrow{{\mathrm{can}}}E_{i,i}/((T_{r}\otimes 1)x-(1\otimes T_{r})x)_{x\in E_{i,i},\ 1\leq r<i}.

We have A0=BA^{0}=B and AA is generated by A0A^{0} and A1=(E0,1​E1,0)/K1A^{1}=(E_{0,1}E_{1,0})/K_{1} as an algebra.

Let (M,ς)(M,\varsigma) be an object of ΔE⊗B−(B−diff)\Delta_{E\otimes_{B}-}(B\operatorname{\!-diff}\nolimits). The action of TB​(E0,1​E1,0)T_{B}(E_{0,1}E_{1,0}) on MM vanishes on KiK_{i} for all ii, hence defines an action of AA on MM. This gives a fully faithful differential functor ΔE⊗B−(B−diff)→(ΔE(B))−diff\Delta_{E\otimes_{B}-}(B\operatorname{\!-diff}\nolimits)\to(\Delta_{E}(B))\operatorname{\!-diff}\nolimits. If the canonical injective morphism of differential (B,B)(B,B)-bimodules

(5.2.1) (E0,1​E1,0)i/Ki→Ei,i/((Tr⊗1)​x−(1⊗Tr)​x)x∈Ei,i, 1≤r<i(E_{0,1}E_{1,0})^{i}/K_{i}\to E_{i,i}/((T_{r}\otimes 1)x-(1\otimes T_{r})x)_{x\in E_{i,i},\ 1\leq r<i}

is a split injection for all i≥1i\geq 1, then the functor above is an isomorphism

ΔE⊗B−(B−diff)→∼(ΔE(B))−diff.\Delta_{E\otimes_{B}-}(B\operatorname{\!-diff}\nolimits)\mathrel{\mathop{\kern 0.0pt\to}\limits^{\sim}}(\Delta_{E}(B))\operatorname{\!-diff}\nolimits.

5.3. Diagonal action

5.3.1. Algebra

Let BB be a differential algebra endowed with two 22-representations (F1,τ1)(F_{1},\tau_{1}) and (E2,τ2)(E_{2},\tau_{2}) together with a closed morphism λ:F1​E2→E2​F1\lambda:F_{1}E_{2}\to E_{2}F_{1} such that the diagrams (4.2.1) commute.

We define the algebra A=Δλ′​(B)A=\Delta^{\prime}_{\lambda}(B) as the quotient of the tensor algebra TB​(F1​E2)T_{B}(F_{1}E_{2}) by the two-sided ideal generated by the image of the composition

F12​E22→τ1​E22−F12​τ2F12​E22→F1​λ​E2(F1​E2)2.F_{1}^{2}E_{2}^{2}\xrightarrow{\tau_{1}E_{2}^{2}-F_{1}^{2}\tau_{2}}F_{1}^{2}E_{2}^{2}\xrightarrow{F_{1}\lambda E_{2}}(F_{1}E_{2})^{2}.

We have A0=BA^{0}=B and A1=F1​E2A^{1}=F_{1}E_{2}.

Let B′B^{\prime} be a differential algebra endowed with two 22-representations (F1′,τ1′)(F^{\prime}_{1},\tau^{\prime}_{1}) and (E2′,τ2′)(E^{\prime}_{2},\tau^{\prime}_{2}) together with a closed morphism λ′:F1′​E2′→E2′​F1′\lambda^{\prime}:F^{\prime}_{1}E^{\prime}_{2}\to E^{\prime}_{2}F^{\prime}_{1} such that the analogs of the diagrams (4.2.1) commute. Let A′=Δλ′′​(B′)A^{\prime}=\Delta^{\prime}_{\lambda^{\prime}}(B^{\prime}). Let PP be a (B′,B)(B^{\prime},B)-bimodule and φ1:P​F1→∼F1′​P\varphi_{1}:PF_{1}\mathrel{\mathop{\kern 0.0pt\to}\limits^{\sim}}F^{\prime}_{1}P and φ2:P​E2→∼E2′​P\varphi_{2}:PE_{2}\mathrel{\mathop{\kern 0.0pt\to}\limits^{\sim}}E^{\prime}_{2}P be two closed isomorphisms of bimodules such that (P,φ1)(P,\varphi_{1}) and (P,φ2)(P,\varphi_{2}) are morphisms of 22-representations and such that

λ′​P∘F1′​φ2∘φ1​E2=E2′​φ1∘φ2​F1∘P​λ:P​F1​E2→E2′​F1′​P.\lambda^{\prime}P\circ F^{\prime}_{1}\varphi_{2}\circ\varphi_{1}E_{2}=E^{\prime}_{2}\varphi_{1}\circ\varphi_{2}F_{1}\circ P\lambda:PF_{1}E_{2}\to E^{\prime}_{2}F^{\prime}_{1}P.

The isomorphism F1′​φ2∘φ1​E2:P​F1​E2→∼F1′​E2′​PF^{\prime}_{1}\varphi_{2}\circ\varphi_{1}E_{2}:PF_{1}E_{2}\mathrel{\mathop{\kern 0.0pt\to}\limits^{\sim}}F^{\prime}_{1}E_{2}^{\prime}P induces an isomorphism of (B′,B)(B^{\prime},B)-bimodules f:P⊗BTB​(F1​E2)→∼TB′​(F1′​E2′)⊗B′Pf:P\otimes_{B}T_{B}(F_{1}E_{2})\mathrel{\mathop{\kern 0.0pt\to}\limits^{\sim}}T_{B^{\prime}}(F^{\prime}_{1}E^{\prime}_{2})\otimes_{B^{\prime}}P. This isomorphism ff endows the right TB​(F1​E2)T_{B}(F_{1}E_{2})-module P⊗BTB​(F1​E2)P\otimes_{B}T_{B}(F_{1}E_{2}) with a commuting left action of TB′​(F1′​E2′)T_{B^{\prime}}(F^{\prime}_{1}E^{\prime}_{2}). The isomorphism ff induces an isomorphism

P⊗BTB(F1E2)⊗TB​(F1​E2)A→∼A′⊗TB′​(F1′​E2′)TB′(F1′E2′)⊗B′P.P\otimes_{B}T_{B}(F_{1}E_{2})\otimes_{T_{B}(F_{1}E_{2})}A\mathrel{\mathop{\kern 0.0pt\to}\limits^{\sim}}A^{\prime}\otimes_{T_{B^{\prime}}(F^{\prime}_{1}E^{\prime}_{2})}T_{B^{\prime}}(F^{\prime}_{1}E^{\prime}_{2})\otimes_{B^{\prime}}P.

So, we obtain a structure of (A′,A)(A^{\prime},A)-bimodule on P⊗BAP\otimes_{B}A.

0P6U

Remark 5.3.1. The data of φ1\varphi_{1} and φ2\varphi_{2} and the relations they are required to satisfy are described graphically as:

[Uncaptioned image]

5.3.2. Left dual

Let BB be a differential algebra endowed with two 22-representations (E1,τ1)(E_{1},\tau_{1}) and (E2,τ2)(E_{2},\tau_{2}), the first of which is right finite.

We consider the data of σ∈Z​Hom⁡(E2​E1,E1​E2)\sigma\in Z\operatorname{Hom}\nolimits(E_{2}E_{1},E_{1}E_{2}) such that the diagrams (4.3.1) commute.

We define

(5.3.1) λ:E1∨​E2→∙η1E1∨​E2​E1​E1∨→E1∨​σ​E1∨E1∨​E1​E2​E1∨→ε1∙E2​E1∨.\lambda:E_{1}^{\vee}E_{2}\xrightarrow{\bullet\eta_{1}}E_{1}^{\vee}E_{2}E_{1}E_{1}^{\vee}\xrightarrow{E_{1}^{\vee}\sigma E_{1}^{\vee}}E_{1}^{\vee}E_{1}E_{2}E_{1}^{\vee}\xrightarrow{\varepsilon_{1}\bullet}E_{2}E_{1}^{\vee}.

Let A=Δσ​(B)=Δλ′​(B)A=\Delta_{\sigma}(B)=\Delta^{\prime}_{\lambda}(B). This is the graded quotient of the tensor algebra TB​(E1∨​E2)T_{B}(E_{1}^{\vee}E_{2}) by the ideal generated by the image of the composition

(E1∨)2​E22→τ1​E22−(E1∨)2​τ2(E1∨)2​E22→E1∨​λ​E2(E1∨​E2)2.(E_{1}^{\vee})^{2}E_{2}^{2}\xrightarrow{\tau_{1}E_{2}^{2}-(E_{1}^{\vee})^{2}\tau_{2}}(E_{1}^{\vee})^{2}E_{2}^{2}\xrightarrow{E_{1}^{\vee}\lambda E_{2}}(E_{1}^{\vee}E_{2})^{2}.

The algebra AA is generated by A0=BA^{0}=B and A1=E1∨​E2A^{1}=E_{1}^{\vee}E_{2}.

Let LL be a differential BB-module. The data of a structure of AA-module on LL extending the action of BB is the same as the data of a morphism of BB-modules ς:E1∨​E2⊗BL→L\varsigma:E_{1}^{\vee}E_{2}\otimes_{B}L\to L such that d⁡(ς)=0d(\varsigma)=0 and the following diagram commutes

(5.3.2) (E1∨)2​E22​L\textstyle{(E_{1}^{\vee})^{2}E_{2}^{2}L\ignorespaces\ignorespaces\ignorespaces\ignorespaces}E1∨​λ​E2\scriptstyle{E_{1}^{\vee}\lambda E_{2}}(E1∨​E2)2​L\textstyle{(E_{1}^{\vee}E_{2})^{2}L\ignorespaces\ignorespaces\ignorespaces\ignorespaces}E1∨​E2​ς\scriptstyle{E_{1}^{\vee}E_{2}\varsigma}E1∨​E2​L\textstyle{E_{1}^{\vee}E_{2}L\ignorespaces\ignorespaces\ignorespaces\ignorespaces}ς\scriptstyle{\varsigma}(E1∨)2​E22​L\textstyle{(E_{1}^{\vee})^{2}E_{2}^{2}L\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces}τ1​E22\scriptstyle{\tau_{1}E_{2}^{2}}(E1∨)2​τ2\scriptstyle{(E_{1}^{\vee})^{2}\tau_{2}}L\textstyle{L}(E1∨)2​E22​L\textstyle{(E_{1}^{\vee})^{2}E_{2}^{2}L\ignorespaces\ignorespaces\ignorespaces\ignorespaces}E1∨​λ​E2\scriptstyle{E_{1}^{\vee}\lambda E_{2}}(E1∨​E2)2​L\textstyle{(E_{1}^{\vee}E_{2})^{2}L\ignorespaces\ignorespaces\ignorespaces\ignorespaces}E1∨​E2​ς\scriptstyle{E_{1}^{\vee}E_{2}\varsigma}E1∨​E2​L\textstyle{E_{1}^{\vee}E_{2}L\ignorespaces\ignorespaces\ignorespaces\ignorespaces}ς\scriptstyle{\varsigma}

This gives us an identification (isomorphism of categories) between differential AA-modules and pairs consisting of a differential BB-module LL and a map ς\varsigma as above.

Consider the adjunction isomorphism

ϕ:HomB⁡(E2​L,E1​L)→∼HomB⁡(E1∨​E2​L,L)\phi:\operatorname{Hom}\nolimits_{B}(E_{2}L,E_{1}L)\mathrel{\mathop{\kern 0.0pt\to}\limits^{\sim}}\operatorname{Hom}\nolimits_{B}(E_{1}^{\vee}E_{2}L,L)

Let π∈Z​HomB⁡(E2​L,E1​L)\pi\in Z\operatorname{Hom}\nolimits_{B}(E_{2}L,E_{1}L) and let ς=ϕ⁡(π)∈Z​HomB⁡(E1∨​E2​L,L)\varsigma=\phi(\pi)\in Z\operatorname{Hom}\nolimits_{B}(E_{1}^{\vee}E_{2}L,L). The commutativity of the diagram (5.3.2) is equivalent to the commutativity of the diagram

(5.3.3) E22​L\textstyle{E_{2}^{2}L\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces}E2​π\scriptstyle{E_{2}\pi}τ2\scriptstyle{\tau_{2}}E2​E1​L\textstyle{E_{2}E_{1}L\ignorespaces\ignorespaces\ignorespaces\ignorespaces}σ\scriptstyle{\sigma}E1​E2​L\textstyle{E_{1}E_{2}L\ignorespaces\ignorespaces\ignorespaces\ignorespaces}E1​π\scriptstyle{E_{1}\pi}E12​L\textstyle{E_{1}^{2}L\ignorespaces\ignorespaces\ignorespaces\ignorespaces}τ1\scriptstyle{\tau_{1}}E22​L\textstyle{E_{2}^{2}L\ignorespaces\ignorespaces\ignorespaces\ignorespaces}E2​π\scriptstyle{E_{2}\pi}E2​E1​L\textstyle{E_{2}E_{1}L\ignorespaces\ignorespaces\ignorespaces\ignorespaces}σ\scriptstyle{\sigma}E1​E2​L\textstyle{E_{1}E_{2}L\ignorespaces\ignorespaces\ignorespaces\ignorespaces}E1​π\scriptstyle{E_{1}\pi}E12​L\textstyle{E_{1}^{2}L}

This gives us an identification (isomorphism of categories) between differential AA-modules and pairs [L,π][L,\pi] where LL is a differential BB-module, π∈Z​HomB⁡(E2​L,E1​L)\pi\in Z\operatorname{Hom}\nolimits_{B}(E_{2}L,E_{1}L) and the diagram (5.3.3) commutes. We have obtained the following lemma.

0P6V

Lemma 5.3.2. The construction (m,π)↦[m,π](m,\pi)\mapsto[m,\pi] defines an isomorphism of differential categories Φ:Δσ​(B​−diff)→(Δσ​B)​−diff\Phi:\Delta_{\sigma}(B\operatorname{\!-diff}\nolimits)\to(\Delta_{\sigma}B)\operatorname{\!-diff}\nolimits.

We will show that the structure of 22-representation on Δσ​(B​−diff)\Delta_{\sigma}(B\operatorname{\!-diff}\nolimits) comes from a structure of 22-representation on Δσ​B\Delta_{\sigma}B, when σ\sigma is invertible.

0P6W

Remark 5.3.3. The map ς\varsigma, the relations it is required to satisfy, and the relation ς=ϕ⁡(π)\varsigma=\phi(\pi) are described graphically as:

[Uncaptioned image]

5.3.3. Action

We define the closed morphism of (B,A)(B,A)-bimodules u:E2⊗BA→E1⊗BAu:E_{2}\otimes_{B}A\to E_{1}\otimes_{B}A as the adjoint to the multiplication map E1∨​E2⊗BA→AE_{1}^{\vee}E_{2}\otimes_{B}A\to A. We define EE as the cone of uu.

We define a morphism of (B,A)(B,A)-bimodules v:E2⊗BE→E1⊗BEv:E_{2}\otimes_{B}E\to E_{1}\otimes_{B}E by

v11:E22⊗BA→τ2⊗1E22⊗BA→E2η1∙E2​E1​E1∨​E2⊗BA→σ∙E1​E2​E1∨​E2⊗BA→E1​E2​mult.E1​E2⊗BAv_{11}:E_{2}^{2}\otimes_{B}A\xrightarrow{\tau_{2}\otimes 1}E_{2}^{2}\otimes_{B}A\xrightarrow{E_{2}\eta_{1}\bullet}E_{2}E_{1}E_{1}^{\vee}E_{2}\otimes_{B}A\xrightarrow{\sigma\bullet}E_{1}E_{2}E_{1}^{\vee}E_{2}\otimes_{B}A\xrightarrow{E_{1}E_{2}\mathrm{mult.}}E_{1}E_{2}\otimes_{B}A
v12:E2​E1⊗BA→σ⊗1E1​E2⊗BAv_{12}:E_{2}E_{1}\otimes_{B}A\xrightarrow{\sigma\otimes 1}E_{1}E_{2}\otimes_{B}A
v21=0v_{21}=0
v22:E2​E1⊗BA→σ⊗1E1​E2⊗BA→E1η1∙E12​E1∨​E2⊗BA→τ1∙E12​E1∨​E2⊗BA→E12​mult.E12⊗BAv_{22}:E_{2}E_{1}\otimes_{B}A\xrightarrow{\sigma\otimes 1}E_{1}E_{2}\otimes_{B}A\xrightarrow{E_{1}\eta_{1}\bullet}E_{1}^{2}E_{1}^{\vee}E_{2}\otimes_{B}A\xrightarrow{\tau_{1}\bullet}E_{1}^{2}E_{1}^{\vee}E_{2}\otimes_{B}A\xrightarrow{E_{1}^{2}\mathrm{mult.}}E_{1}^{2}\otimes_{B}A

Note that the morphism vv corresponds, by adjunction, to the morphism w:E1∨​E2⊗BE→Ew:E_{1}^{\vee}E_{2}\otimes_{B}E\to E defined as follows

w11:E1∨​E22⊗BA→E1∨​τ2⊗1E1∨​E22⊗BA→λ​E2⊗1E2​E1∨​E2⊗BA→E2​mult.E2⊗BAw_{11}:E_{1}^{\vee}E_{2}^{2}\otimes_{B}A\xrightarrow{E_{1}^{\vee}\tau_{2}\otimes 1}E_{1}^{\vee}E_{2}^{2}\otimes_{B}A\xrightarrow{\lambda E_{2}\otimes 1}E_{2}E_{1}^{\vee}E_{2}\otimes_{B}A\xrightarrow{E_{2}\mathrm{mult.}}E_{2}\otimes_{B}A
w12:E1∨​E2​E1⊗BA→E1∨σ∙E1∨​E1​E2⊗BA→ε1∙E2⊗BA,w21=0w_{12}:E_{1}^{\vee}E_{2}E_{1}\otimes_{B}A\xrightarrow{E_{1}^{\vee}\sigma\bullet}E_{1}^{\vee}E_{1}E_{2}\otimes_{B}A\xrightarrow{\varepsilon_{1}\bullet}E_{2}\otimes_{B}A,\ w_{21}=0
w22:E1∨​E2​E1⊗BA→E1∨​σ⊗1E1∨​E1​E2⊗BA→ρ1∙E1​E1∨​E2⊗BA→E1​mult.E1⊗BA.w_{22}:E_{1}^{\vee}E_{2}E_{1}\otimes_{B}A\xrightarrow{E_{1}^{\vee}\sigma\otimes 1}E_{1}^{\vee}E_{1}E_{2}\otimes_{B}A\xrightarrow{\rho_{1}\bullet}E_{1}E_{1}^{\vee}E_{2}\otimes_{B}A\xrightarrow{E_{1}\mathrm{mult.}}E_{1}\otimes_{B}A.
0P6X

Lemma 5.3.4. The pair [E,v][E,v] gives EE a structure of differential (A,A)(A,A)-bimodule via Lemma 5.3.2. Furthermore, there is an isomorphism of functors Φℰ→∼(E⊗A−)Φ:Δσ(B−diff)→A−diff\Phi{\mathcal{E}}\mathrel{\mathop{\kern 0.0pt\to}\limits^{\sim}}(E\otimes_{A}-)\Phi:\Delta_{\sigma}(B\operatorname{\!-diff}\nolimits)\to A\operatorname{\!-diff}\nolimits.

0P6Y

Proof. The vanishing of d​(v)11d(v)_{11} and d​(v)22d(v)_{22} follows from d⁡(τ1)=idd(\tau_{1})=\operatorname{id}\nolimits and d⁡(τ2)=idd(\tau_{2})=\operatorname{id}\nolimits. The vanishing of d​(v)12d(v)_{12} is clear. Finally, the vanishing of d​(v)21d(v)_{21} follows from the commutativity of the diagram (5.3.3). Since d⁡(v)=0d(v)=0, we have obtained a structure of differential (TB​(E1∨​E2),A)(T_{B}(E_{1}^{\vee}E_{2}),A)-bimodule on EE.

The object of Δσ​(B​−diff)\Delta_{\sigma}(B\operatorname{\!-diff}\nolimits) corresponding to AA via Lemma 5.3.2 is (A,u)(A,u). We have ℰ⁡(A,u)=(E,v){\mathcal{E}}(A,u)=(E,v), where ℰ{\mathcal{E}} is the endofunctor defining the 22-representation on Δσ​(B​−diff)\Delta_{\sigma}(B\operatorname{\!-diff}\nolimits). Since (E,v)(E,v) is an object of Δσ​(B​−diff)\Delta_{\sigma}(B\operatorname{\!-diff}\nolimits), it follows that the action of TB​(E1∨​E2)T_{B}(E_{1}^{\vee}E_{2}) on EE factors through an action of AA. So, EE has a structure of differential (A,A)(A,A)-bimodule and we have an isomorphism of functors Φℰ→∼(E⊗A−)Φ:Δσ(B−diff)→A−diff\Phi{\mathcal{E}}\mathrel{\mathop{\kern 0.0pt\to}\limits^{\sim}}(E\otimes_{A}-)\Phi:\Delta_{\sigma}(B\operatorname{\!-diff}\nolimits)\to A\operatorname{\!-diff}\nolimits. ∎

0P6Z

Remark 5.3.5. The maps vv and ww are described graphically as:

[Uncaptioned image]

We assume now that σ\sigma is invertible. We define τ\tau an endomorphism of (B,A)(B,A)-bimodules of E22⊗BA⊕E2​E1⊗BA⊕E1​E2⊗BA⊕E12⊗BAE_{2}^{2}\otimes_{B}A\oplus E_{2}E_{1}\otimes_{B}A\oplus E_{1}E_{2}\otimes_{B}A\oplus E_{1}^{2}\otimes_{B}A by

(5.3.4) τ=(τ2⊗100000σ−1⊗100000000τ1⊗1).\tau=\left(\begin{matrix}\tau_{2}\otimes 1&0&0&0\\ 0&0&\sigma^{-1}\otimes 1&0\\ 0&0&0&0\\ 0&0&0&\tau_{1}\otimes 1\end{matrix}\right).
0P70

Proposition 5.3.6. The pair (E,τ)(E,\tau) defines a 22-representation on AA and Φ\Phi induces a isomorphism of 22-representations Δσ​(B​−diff)→∼(Δσ​B)​−diff\Delta_{\sigma}(B\operatorname{\!-diff}\nolimits)\mathrel{\mathop{\kern 0.0pt\to}\limits^{\sim}}(\Delta_{\sigma}B)\operatorname{\!-diff}\nolimits. If E2E_{2} is right finite, then EE is right finite.

0P71

Proof. The fact that τ\tau defines an endomorphism of (A,A)(A,A)-bimodules of E2E^{2} satisfying the appropriate relations follows from the fact that it agrees with the endomorphism of ℰ2{\mathcal{E}}^{2} defining the 22-representation on Δσ​(B​−diff)\Delta_{\sigma}(B\operatorname{\!-diff}\nolimits). We deduce that (E,τ)(E,\tau) is a 22-representation on AA and Φ\Phi is a morphism of 22-representations.

Note that EE is finitely generated and projective as a (non-differential) AoppA^{\operatorname{opp}\nolimits}-module if E1E_{1} and E2E_{2} are finitely generated and projective BoppB^{\operatorname{opp}\nolimits}-modules. ∎

0P72

Remark 5.3.7. Consider three 22-representations (Ei,τi)1≤i≤3(E_{i},\tau_{i})_{1\leq i\leq 3} on a differential algebra BB together with closed morphisms σi​j:Ei​Ej→∼Ej​Ei\sigma_{ij}:E_{i}E_{j}\mathrel{\mathop{\kern 0.0pt\to}\limits^{\sim}}E_{j}E_{i} for i≠ji\neq j satisfying (4.3.5). Assume E1E_{1} and E2E_{2} are right finite. We will construct a triple tensor product 22-representation.

Define

λi​j:Ei∨​Ej→Ei∨​Ej​ηiEi∨​Ej​Ei​Ei∨→Ei∨​σj​i​Ei∨Ei∨​Ei​Ej​Ei∨→εi​Ej​Ei∨Ej​Ei∨\lambda_{ij}:E_{i}^{\vee}E_{j}\xrightarrow{E_{i}^{\vee}E_{j}\eta_{i}}E_{i}^{\vee}E_{j}E_{i}E_{i}^{\vee}\xrightarrow{E_{i}^{\vee}\sigma_{ji}E_{i}^{\vee}}E_{i}^{\vee}E_{i}E_{j}E_{i}^{\vee}\xrightarrow{\varepsilon_{i}E_{j}E_{i}^{\vee}}E_{j}E_{i}^{\vee}

and denote by σi​j∨:Ei∨​Ej∨→Ej∨​Ei∨\sigma_{ij}^{\vee}:E_{i}^{\vee}E_{j}^{\vee}\to E_{j}^{\vee}E_{i}^{\vee} the map adjoint to σi​j\sigma_{ij}.

Let A′=TB​(E1∨​E2⊕E2∨​E3⊕E1∨​E3)A^{\prime}=T_{B}(E_{1}^{\vee}E_{2}\oplus E_{2}^{\vee}E_{3}\oplus E_{1}^{\vee}E_{3}). There is a derivation ∂\partial of A′A^{\prime} whose restriction to B⊕E1∨​E2⊕E2∨​E3B\oplus E_{1}^{\vee}E_{2}\oplus E_{2}^{\vee}E_{3} is 00 and whose restriction to E1∨​E3E_{1}^{\vee}E_{3} is

∂:E1∨​E3→E1∨​η2​E3(E1∨​E2)​(E2∨​E3).\partial:E_{1}^{\vee}E_{3}\xrightarrow{E_{1}^{\vee}\eta_{2}E_{3}}(E_{1}^{\vee}E_{2})(E_{2}^{\vee}E_{3}).

Define A′′A^{\prime\prime} to be the differential algebra with underlying algebra A′A^{\prime} and with differential ∂+dA′\partial+d_{A^{\prime}}.

Let EE be the set of quadruples (i,j,k,l)(i,j,k,l) with i,j,k,l∈{1,2,3}i,j,k,l\in\{1,2,3\}, j−l≥i−k>0j-l\geq i-k>0 and (i,j,k,l)≠(2,3,1,2)(i,j,k,l)\neq(2,3,1,2). Given such a quadruple, we define

fi​j​k​l:El∨​Ek∨​Ei​Ej→σl​k∨​Ei​EjEk∨​El∨​Ei​Ej→Ek​λl​i​Ej(Ek∨​Ei)​(El∨​Ej)f_{ijkl}:E_{l}^{\vee}E_{k}^{\vee}E_{i}E_{j}\xrightarrow{\sigma_{lk}^{\vee}E_{i}E_{j}}E_{k}^{\vee}E_{l}^{\vee}E_{i}E_{j}\xrightarrow{E_{k}\lambda_{li}E_{j}}(E_{k}^{\vee}E_{i})(E_{l}^{\vee}E_{j})
gi​j​k​l:El∨​Ek∨​Ei​Ej→El∨​Ek∨​σi​jEl∨​Ek∨​Ej​Ei→El​λk​j​Ei(El∨​Ej)​(Ek∨​Ei)g_{ijkl}:E_{l}^{\vee}E_{k}^{\vee}E_{i}E_{j}\xrightarrow{E_{l}^{\vee}E_{k}^{\vee}\sigma_{ij}}E_{l}^{\vee}E_{k}^{\vee}E_{j}E_{i}\xrightarrow{E_{l}\lambda_{kj}E_{i}}(E_{l}^{\vee}E_{j})(E_{k}^{\vee}E_{i})
h3221:E1∨​E2∨​E3​E2→E1∨​λ23​E2E1∨​E3​E2∨​E2→E1∨​E3​ε2E1∨​E3h_{3221}:E_{1}^{\vee}E_{2}^{\vee}E_{3}E_{2}\xrightarrow{E_{1}^{\vee}\lambda_{23}E_{2}}E_{1}^{\vee}E_{3}E_{2}^{\vee}E_{2}\xrightarrow{E_{1}^{\vee}E_{3}\varepsilon_{2}}E_{1}^{\vee}E_{3}

where we put σr​r=τr\sigma_{rr}=\tau_{r} and σr​r∨=τr\sigma_{rr}^{\vee}=\tau_{r}. We define I′′I^{\prime\prime} to be the two-sided ideal generated by the images of fi​j​k​l+gi​j​k​l+δj​k​h3221f_{ijkl}+g_{ijkl}+\delta_{jk}h_{3221} for (i,j,k,l)∈E(i,j,k,l)\in E. We put A=A′′/I′′A=A^{\prime\prime}/I^{\prime\prime}.

As in §5.3.2, we have an isomorphism of differential categories Δ123​(B​−diff)→∼A​−diff\Delta_{123}(B\operatorname{\!-diff}\nolimits)\mathrel{\mathop{\kern 0.0pt\to}\limits^{\sim}}A\operatorname{\!-diff}\nolimits (cf §4.3.5).

We obtain a bimodule 22-representation on AA as in §4.3.5. We define the differential (B,A)(B,A)-bimodule

E=    E3⊗BA⊕E2⊗BA⊕E1⊗BA   π31        π32        π21         E={\lx@xy@svg{\hbox{\raise 2.5pt\hbox{\kern 71.13152pt\hbox{{\hbox{\kern-64.13002pt\raise 0.0pt\hbox{\hbox{\kern 0.0pt\raise-2.66666pt\hbox{$\textstyle{E_{3}\otimes_{B}A\oplus E_{2}\otimes_{B}A\oplus E_{1}\otimes_{B}A}$}}}}}\ignorespaces{}{}\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces{}{}{}{{}}{}\ignorespaces\ignorespaces\hbox{\hbox{\kern 0.0pt\raise 0.0pt\hbox{{}{}{}{{}{}{}}{}}}}\ignorespaces{}\ignorespaces{}{}{}{{}{}}\ignorespaces\ignorespaces{\hbox{\kern-8.99098pt\raise 37.54272pt\hbox{{}\hbox{\kern 0.0pt\raise 0.0pt\hbox{\hbox{\kern 3.0pt\hbox{\hbox{\kern 0.0pt\raise-1.00694pt\hbox{$\scriptstyle{\pi_{31}}$}}}\kern 3.0pt}}}}}}\ignorespaces{}{}{}{{}}{\hbox{\kern 71.13152pt\raise 8.53578pt\hbox{\hbox{\kern 0.0pt\raise 0.0pt\hbox{\lx@xy@tip{1}\lx@xy@tip{-1}}}}}}\ignorespaces\ignorespaces\hbox{\hbox{\kern 0.0pt\raise 0.0pt\hbox{{}{}{}{}{{}{}{}\lx@xy@spline@}{}}}}\ignorespaces{}\ignorespaces\hbox{\hbox{\kern 0.0pt\raise 0.0pt\hbox{{}{}{}{{}{}{}}{}}}}\ignorespaces{}\ignorespaces\ignorespaces{}{}\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces{}{}{}{{}}{}\ignorespaces\ignorespaces\hbox{\hbox{\kern 0.0pt\raise 0.0pt\hbox{{}{}{}{{}{}{}}{}}}}\ignorespaces{}\ignorespaces{}{}{}{{}{}}\ignorespaces\ignorespaces{\hbox{\kern-42.33748pt\raise 19.54272pt\hbox{{}\hbox{\kern 0.0pt\raise 0.0pt\hbox{\hbox{\kern 3.0pt\hbox{\hbox{\kern 0.0pt\raise-1.00694pt\hbox{$\scriptstyle{\ \pi_{32}}$}}}\kern 3.0pt}}}}}}\ignorespaces{}{}{}{{}}{\hbox{\kern 5.69052pt\raise 8.53578pt\hbox{\hbox{\kern 0.0pt\raise 0.0pt\hbox{\lx@xy@tip{1}\lx@xy@tip{-1}}}}}}\ignorespaces\ignorespaces\hbox{\hbox{\kern 0.0pt\raise 0.0pt\hbox{{}{}{}{}{{}{}{}\lx@xy@spline@}{}}}}\ignorespaces{}\ignorespaces\hbox{\hbox{\kern 0.0pt\raise 0.0pt\hbox{{}{}{}{{}{}{}}{}}}}\ignorespaces{}\ignorespaces\ignorespaces{}{}\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces{}{}{}{{}}{}\ignorespaces\ignorespaces\hbox{\hbox{\kern 0.0pt\raise 0.0pt\hbox{{}{}{}{{}{}{}}{}}}}\ignorespaces{}\ignorespaces{}{}{}{{}{}}\ignorespaces\ignorespaces{\hbox{\kern 25.32281pt\raise 19.54272pt\hbox{{}\hbox{\kern 0.0pt\raise 0.0pt\hbox{\hbox{\kern 3.0pt\hbox{\hbox{\kern 0.0pt\raise-1.00694pt\hbox{$\scriptstyle{\pi_{21}\ \ }$}}}\kern 3.0pt}}}}}}\ignorespaces{}{}{}{{}}{\hbox{\kern 68.28625pt\raise 8.53578pt\hbox{\hbox{\kern 0.0pt\raise 0.0pt\hbox{\lx@xy@tip{1}\lx@xy@tip{-1}}}}}}\ignorespaces\ignorespaces\hbox{\hbox{\kern 0.0pt\raise 0.0pt\hbox{{}{}{}{}{{}{}{}\lx@xy@spline@}{}}}}\ignorespaces{}\ignorespaces\hbox{\hbox{\kern 0.0pt\raise 0.0pt\hbox{{}{}{}{{}{}{}}{}}}}\ignorespaces{}\ignorespaces}}}}}

where

πi​j:Ei⊗BA→ηj​idEj​Ej∨​Ei⊗BA→Ej​multEj⊗BA.\pi_{ij}:E_{i}\otimes_{B}A\xrightarrow{\eta_{j}\operatorname{id}\nolimits}E_{j}E_{j}^{\vee}E_{i}\otimes_{B}A\xrightarrow{E_{j}\mathrm{mult}}E_{j}\otimes_{B}A.

We extend the left action of BB to an action of AA by letting the action maps

Ei∨​Ej⊗BE→EE_{i}^{\vee}E_{j}\otimes_{B}E\to E

for i<ji<j be given by

E1∨​E32⊗BA⊕E1∨​E3​E2⊗BA⊕E1∨​E3​E1⊗BA\textstyle{E_{1}^{\vee}E_{3}^{2}\otimes_{B}A\oplus E_{1}^{\vee}E_{3}E_{2}\otimes_{B}A\oplus E_{1}^{\vee}E_{3}E_{1}\otimes_{B}A}E3⊗BA⊕E2⊗BA⊕E1⊗BA\textstyle{E_{3}\otimes_{B}A\oplus E_{2}\otimes_{B}A\oplus E_{1}\otimes_{B}A}E3​ mult∘λ13​E3∘E1∨​τ3\scriptstyle{E_{3}\text{ mult}\circ\lambda_{13}E_{3}\circ E_{1}^{\vee}\tau_{3}}E1​mult∘ρ1​E3∘E1∨​σ31\scriptstyle{E_{1}\mathrm{mult}\circ\rho_{1}E_{3}\circ E_{1}^{\vee}\sigma_{31}}E3​ε1∘λ13​E1\scriptstyle{E_{3}\varepsilon_{1}\circ\lambda_{13}E_{1}\ \ }
E2∨​E32⊗BA⊕E2∨​E3​E2⊗BA⊕E2∨​E3​E1⊗BA\textstyle{E_{2}^{\vee}E_{3}^{2}\otimes_{B}A\oplus E_{2}^{\vee}E_{3}E_{2}\otimes_{B}A\oplus E_{2}^{\vee}E_{3}E_{1}\otimes_{B}A}E3⊗BA⊕E2⊗BA⊕E1⊗BA\textstyle{E_{3}\otimes_{B}A\oplus E_{2}\otimes_{B}A\oplus E_{1}\otimes_{B}A}E3​ mult∘λ23​E3∘E2∨​τ3\scriptstyle{E_{3}\text{ mult}\circ\lambda_{23}E_{3}\circ E_{2}^{\vee}\tau_{3}}E2​mult∘ρ2​E3∘E2∨​σ32\scriptstyle{\!\!E_{2}\mathrm{mult}\circ\rho_{2}E_{3}\circ E_{2}^{\vee}\sigma_{32}}E3​ε2∘λ23​E2\scriptstyle{E_{3}\varepsilon_{2}\circ\lambda_{23}E_{2}\!}E1​mult∘λ21​E3∘E2∨​σ31\scriptstyle{E_{1}\mathrm{mult}\circ\lambda_{21}E_{3}\circ E_{2}^{\vee}\sigma_{31}}
E1∨​E2​E3⊗BA⊕E1∨​E22⊗BA⊕E1∨​E2​E1⊗BA\textstyle{E_{1}^{\vee}E_{2}E_{3}\otimes_{B}A\oplus E_{1}^{\vee}E_{2}^{2}\otimes_{B}A\oplus E_{1}^{\vee}E_{2}E_{1}\otimes_{B}A}E3⊗BA⊕E2⊗BA⊕E1⊗BA\textstyle{E_{3}\otimes_{B}A\oplus E_{2}\otimes_{B}A\oplus E_{1}\otimes_{B}A}E3​mult∘λ13​E2∘E1∨​σ23\scriptstyle{E_{3}\mathrm{mult}\circ\lambda_{13}E_{2}\circ E_{1}^{\vee}\sigma_{23}}E2​ mult∘λ12​E2∘E1∨​τ2\scriptstyle{E_{2}\text{ mult}\circ\lambda_{12}E_{2}\circ E_{1}^{\vee}\tau_{2}\!\!}E2​ε1∘λ12​E1\scriptstyle{E_{2}\varepsilon_{1}\circ\lambda_{12}E_{1}\!}E1​mult∘ρ1​E2∘E1∨​σ21\scriptstyle{E_{1}\mathrm{mult}\circ\rho_{1}E_{2}\circ E_{1}^{\vee}\sigma_{21}}

Finally, we define the endomorphism τ\tau of E2E^{2} as in (4.3.6).

5.3.4. Tensor product case

Let A1A_{1} and A2A_{2} be two differential algebras equipped with structures of 22-representations (Ei,τi)(E_{i},\tau_{i}), i=1,2i=1,2.

Let B=A1⊗A2B=A_{1}\otimes A_{2}. It is endowed with commuting 22-representations (E1⊗A2,τ1⊗1)(E_{1}\otimes A_{2},\tau_{1}\otimes 1) and (A1⊗E2,1⊗τ2)(A_{1}\otimes E_{2},1\otimes\tau_{2}): the isomorphism σ\sigma is induced by the swap map E2⊗E1→∼E1⊗E2,a2⊗a1↦a1⊗a2E_{2}\otimes E_{1}\xrightarrow{\sim}E_{1}\otimes E_{2},\ a_{2}\otimes a_{1}\mapsto a_{1}\otimes a_{2}. The tensor product identifies (A1​−diff)⊗(A2​−diff)(A_{1}\operatorname{\!-diff}\nolimits)\otimes(A_{2}\operatorname{\!-diff}\nolimits) with a full subcategory of B​−diffB\operatorname{\!-diff}\nolimits.

Assume E1E_{1} is right finite. The map λ\lambda is an isomorphism. We put A1⊗○A2=Δλ′(B)A_{1}{\hskip 1.9919pt\otimes\hskip-10.81218pt\bigcirc}A_{2}=\Delta^{\prime}_{\lambda}(B). It is the quotient of the tensor algebra TA1⊗A2​(E1∨⊗E2)T_{A_{1}\otimes A_{2}}(E_{1}^{\vee}\otimes E_{2}) by the ideal generated by p​τ2​(q)−τ1​(p)​qp\tau_{2}(q)-\tau_{1}(p)q for p∈(E1∨)⊗2p\in(E_{1}^{\vee})^{\otimes 2} and q∈(E2)⊗2q\in(E_{2})^{\otimes 2}. The underlying differential module is

A=⨁i≥0(E1∨)i⊗HiE2i.A=\bigoplus_{i\geq 0}(E_{1}^{\vee})^{i}\otimes_{H_{i}}E_{2}^{i}.

The multiplication is defined by

((E1i)∨⊗HiE2i)⊗((E1j)∨⊗HjE2j)→(E1i+j)∨⊗Hi+jE2i+j,(a1⊗a2)⊗(b1⊗b2)↦(a1​b1)⊗(a2​b2).\bigl((E_{1}^{i})^{\vee}\otimes_{H_{i}}E_{2}^{i}\bigr)\otimes\bigl((E_{1}^{j})^{\vee}\otimes_{H_{j}}E_{2}^{j}\bigr)\to(E_{1}^{i+j})^{\vee}\otimes_{H_{i+j}}E_{2}^{i+j},\ (a_{1}\otimes a_{2})\otimes(b_{1}\otimes b_{2})\mapsto(a_{1}b_{1})\otimes(a_{2}b_{2}).

We have

E=    (⨁i≥0(E1∨)i⊗HiE2​E2i)⊕(⨁i≥0E1​(E1∨)i⊗HiE2i)   η1⊗1         .E={\lx@xy@svg{\hbox{\raise 2.5pt\hbox{\kern 106.07784pt\hbox{{\hbox{\kern-106.07784pt\raise 0.0pt\hbox{\hbox{\kern 0.0pt\raise 0.50006pt\hbox{$\textstyle{\bigl(\bigoplus_{i\geq 0}(E_{1}^{\vee})^{i}\otimes_{H_{i}}E_{2}E_{2}^{i}\bigr)\oplus\bigl(\bigoplus_{i\geq 0}E_{1}(E_{1}^{\vee})^{i}\otimes_{H_{i}}E_{2}^{i}\bigr)}$}}}}}\ignorespaces{}{}\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces{}{}{}{{}}{}\ignorespaces\ignorespaces\hbox{\hbox{\kern 0.0pt\raise 0.0pt\hbox{{}{}{}{{}{}{}}{}}}}\ignorespaces{}\ignorespaces{}{}{}{{}{}}\ignorespaces\ignorespaces{\hbox{\kern-17.8201pt\raise 23.31715pt\hbox{{}\hbox{\kern 0.0pt\raise 0.0pt\hbox{\hbox{\kern 3.0pt\hbox{\hbox{\kern 0.0pt\raise-1.57501pt\hbox{$\scriptstyle{\eta_{1}\otimes 1}$}}}\kern 3.0pt}}}}}}\ignorespaces{}{}{}{{}}{\hbox{\kern 68.28625pt\raise 11.38104pt\hbox{\hbox{\kern 0.0pt\raise 0.0pt\hbox{\lx@xy@tip{1}\lx@xy@tip{-1}}}}}}\ignorespaces\ignorespaces\hbox{\hbox{\kern 0.0pt\raise 0.0pt\hbox{{}{}{}{}{{}{}{}\lx@xy@spline@}{}}}}\ignorespaces{}\ignorespaces\hbox{\hbox{\kern 0.0pt\raise 0.0pt\hbox{{}{}{}{{}{}{}}{}}}}\ignorespaces{}\ignorespaces}}}}}.

The right action of AA on EE is given by right multiplication, while the left action of E1∨⊗E2E_{1}^{\vee}\otimes E_{2} on A1⊗E2⊕E1⊗A2⊂EA_{1}\otimes E_{2}\oplus E_{1}\otimes A_{2}\subset E is given by

(E1∨⊗E2)⊗A1⊗A2(A1⊗E2)→∼canE1∨⊗E22→1⊗τ2E1∨⊗E22(E_{1}^{\vee}\otimes E_{2})\otimes_{A_{1}\otimes A_{2}}(A_{1}\otimes E_{2})\xrightarrow[\sim]{{\mathrm{can}}}E_{1}^{\vee}\otimes E_{2}^{2}\xrightarrow{1\otimes\tau_{2}}E_{1}^{\vee}\otimes E_{2}^{2}
(E1∨⊗E2)⊗A1⊗A2(E1⊗A2)→∼canE1∨​E1⊗E2→(ε1,ρ1)A1⊗E2⊕E1​E1∨⊗E2.(E_{1}^{\vee}\otimes E_{2})\otimes_{A_{1}\otimes A_{2}}(E_{1}\otimes A_{2})\xrightarrow[\sim]{{\mathrm{can}}}E_{1}^{\vee}E_{1}\otimes E_{2}\xrightarrow{(\varepsilon_{1},\rho_{1})}A_{1}\otimes E_{2}\oplus E_{1}E_{1}^{\vee}\otimes E_{2}.

We have

E2=    (⨁(E1∨)i⊗HiE22​E2i)⊕(⨁E1​(E1∨)i⊗HiE2​E2i)⊕(⨁E1​(E1∨)i⊗HiE2​E2i)⊕(⨁E12​(E1∨)i⊗HiE2i)   η1​(E1∨)i⊗E22+i        η1​(E1∨)i⊗τ2​E2i        E1​η1​(E1∨)i⊗E21+i        (E1​ρ1∘η1​E1)​(E1∨)i⊗E21+i        1         .E^{2}={\lx@xy@svg{\hbox{\raise 2.5pt\hbox{\kern 203.8581pt\hbox{{\hbox{\kern-203.8581pt\raise 0.0pt\hbox{\hbox{\kern 0.0pt\raise 0.50006pt\hbox{$\textstyle{\bigl(\bigoplus(E_{1}^{\vee})^{i}\otimes_{H_{i}}E_{2}^{2}E_{2}^{i}\bigr)\oplus\bigl(\bigoplus E_{1}(E_{1}^{\vee})^{i}\otimes_{H_{i}}E_{2}E_{2}^{i}\bigr)\oplus\bigl(\bigoplus E_{1}(E_{1}^{\vee})^{i}\otimes_{H_{i}}E_{2}E_{2}^{i}\bigr)\oplus\bigl(\bigoplus E_{1}^{2}(E_{1}^{\vee})^{i}\otimes_{H_{i}}E_{2}^{i}\bigr)}$}}}}}\ignorespaces{}{}\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces{}{}{}{{}}{}\ignorespaces\ignorespaces\hbox{\hbox{\kern 0.0pt\raise 0.0pt\hbox{{}{}{}{{}{}{}}{}}}}\ignorespaces{}\ignorespaces{}{}{}{{}{}{}{}{}{}{}{}{}{}{}{}{}{}{}{}{}{}{}}\ignorespaces\ignorespaces{\hbox{\kern-129.51878pt\raise 23.4721pt\hbox{{}\hbox{\kern 0.0pt\raise 0.0pt\hbox{\hbox{\kern 3.0pt\hbox{\hbox{\kern 0.0pt\raise-2.30058pt\hbox{$\scriptstyle{\eta_{1}(E_{1}^{\vee})^{i}\otimes E_{2}^{2+i}}$}}}\kern 3.0pt}}}}}}\ignorespaces{}{}{}{{}}{\hbox{\kern-68.28625pt\raise 11.38104pt\hbox{\hbox{\kern 0.0pt\raise 0.0pt\hbox{\lx@xy@tip{1}\lx@xy@tip{-1}}}}}}\ignorespaces\ignorespaces\hbox{\hbox{\kern 0.0pt\raise 0.0pt\hbox{{}{}{}{}{{}{}{}\lx@xy@spline@}{}}}}\ignorespaces{}\ignorespaces\hbox{\hbox{\kern 0.0pt\raise 0.0pt\hbox{{}{}{}{{}{}{}}{}}}}\ignorespaces{}\ignorespaces\ignorespaces{}{}\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces{}{}{}{{}}{}\ignorespaces\ignorespaces\hbox{\hbox{\kern 0.0pt\raise 0.0pt\hbox{{}{}{}{{}{}{}}{}}}}\ignorespaces{}\ignorespaces{}{}{}{{}{}}\ignorespaces\ignorespaces{\hbox{\kern-96.32256pt\raise 42.43112pt\hbox{{}\hbox{\kern 0.0pt\raise 0.0pt\hbox{\hbox{\kern 3.0pt\hbox{\hbox{\kern 0.0pt\raise-2.30008pt\hbox{$\scriptstyle{\eta_{1}(E_{1}^{\vee})^{i}\otimes\tau_{2}E_{2}^{i}}$}}}\kern 3.0pt}}}}}}\ignorespaces{}{}{}{{}}{\hbox{\kern 56.90521pt\raise 11.38104pt\hbox{\hbox{\kern 0.0pt\raise 0.0pt\hbox{\lx@xy@tip{1}\lx@xy@tip{-1}}}}}}\ignorespaces\ignorespaces\hbox{\hbox{\kern 0.0pt\raise 0.0pt\hbox{{}{}{}{}{{}{}{}\lx@xy@spline@}{}}}}\ignorespaces{}\ignorespaces\hbox{\hbox{\kern 0.0pt\raise 0.0pt\hbox{{}{}{}{{}{}{}}{}}}}\ignorespaces{}\ignorespaces\ignorespaces{}{}\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces{}{}{}{{}}{}\ignorespaces\ignorespaces\hbox{\hbox{\kern 0.0pt\raise 0.0pt\hbox{{}{}{}{{}{}{}}{}}}}\ignorespaces{}\ignorespaces{}{}{}{{}{}{}{}{}{}{}{}{}{}{}{}{}{}{}{}{}{}}\ignorespaces\ignorespaces{\hbox{\kern 64.87904pt\raise 20.62685pt\hbox{{}\hbox{\kern 0.0pt\raise 0.0pt\hbox{\hbox{\kern 3.0pt\hbox{\hbox{\kern 0.0pt\raise-2.30058pt\hbox{$\scriptstyle{E_{1}\eta_{1}(E_{1}^{\vee})^{i}\otimes E_{2}^{1+i}}$}}}\kern 3.0pt}}}}}}\ignorespaces{}{}{}{{}}{\hbox{\kern 182.09668pt\raise 8.53578pt\hbox{\hbox{\kern 0.0pt\raise 0.0pt\hbox{\lx@xy@tip{1}\lx@xy@tip{-1}}}}}}\ignorespaces\ignorespaces\hbox{\hbox{\kern 0.0pt\raise 0.0pt\hbox{{}{}{}{}{{}{}{}\lx@xy@spline@}{}}}}\ignorespaces{}\ignorespaces\hbox{\hbox{\kern 0.0pt\raise 0.0pt\hbox{{}{}{}{{}{}{}}{}}}}\ignorespaces{}\ignorespaces\ignorespaces{}{}\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces{}{}{}{{}}{}\ignorespaces\ignorespaces\hbox{\hbox{\kern 0.0pt\raise 0.0pt\hbox{{}{}{}{{}{}{}}{}}}}\ignorespaces{}\ignorespaces{}{}{}{{}{}}\ignorespaces\ignorespaces{\hbox{\kern 22.2066pt\raise 42.43163pt\hbox{{}\hbox{\kern 0.0pt\raise 0.0pt\hbox{\hbox{\kern 3.0pt\hbox{\hbox{\kern 0.0pt\raise-2.30058pt\hbox{$\scriptstyle{(E_{1}\rho_{1}\circ\eta_{1}E_{1})(E_{1}^{\vee})^{i}\otimes E_{2}^{1+i}}$}}}\kern 3.0pt}}}}}}\ignorespaces{}{}{}{{}}{\hbox{\kern 182.09668pt\raise 11.38104pt\hbox{\hbox{\kern 0.0pt\raise 0.0pt\hbox{\lx@xy@tip{1}\lx@xy@tip{-1}}}}}}\ignorespaces\ignorespaces\hbox{\hbox{\kern 0.0pt\raise 0.0pt\hbox{{}{}{}{}{{}{}{}\lx@xy@spline@}{}}}}\ignorespaces{}\ignorespaces\hbox{\hbox{\kern 0.0pt\raise 0.0pt\hbox{{}{}{}{{}{}{}}{}}}}\ignorespaces{}\ignorespaces\ignorespaces{}{}\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces{}{}{}{{}}{}\ignorespaces\ignorespaces\hbox{\hbox{\kern 0.0pt\raise 0.0pt\hbox{{}{}{}{{}{}{}}{}}}}\ignorespaces{}\ignorespaces{}{}{}{{}{}}\ignorespaces\ignorespaces{\hbox{\kern 3.54272pt\raise 19.79134pt\hbox{{}\hbox{\kern 0.0pt\raise 0.0pt\hbox{\hbox{\kern 3.0pt\hbox{\hbox{\kern 0.0pt\raise-2.25555pt\hbox{$\scriptstyle{1}$}}}\kern 3.0pt}}}}}}\ignorespaces{}{}{}{{}}{\hbox{\kern 56.90521pt\raise 8.53578pt\hbox{\hbox{\kern 0.0pt\raise 0.0pt\hbox{\lx@xy@tip{1}\lx@xy@tip{-1}}}}}}\ignorespaces\ignorespaces\hbox{\hbox{\kern 0.0pt\raise 0.0pt\hbox{{}{}{}{}{{}{}{}\lx@xy@spline@}{}}}}\ignorespaces{}\ignorespaces\hbox{\hbox{\kern 0.0pt\raise 0.0pt\hbox{{}{}{}{{}{}{}}{}}}}\ignorespaces{}\ignorespaces}}}}}.

The endomorphism τ\tau of E2E^{2} is given on

((E1∨)i⊗HiE22​E2i)⊕(E1​(E1∨)i⊗HiE2​E2i)⊕(E1​(E1∨)i⊗HiE2​E2i)⊕(E12​(E1∨)i⊗HiE2i)\bigl((E_{1}^{\vee})^{i}\otimes_{H_{i}}E_{2}^{2}E_{2}^{i}\bigr)\oplus\bigl(E_{1}(E_{1}^{\vee})^{i}\otimes_{H_{i}}E_{2}E_{2}^{i}\bigr)\oplus\bigl(E_{1}(E_{1}^{\vee})^{i}\otimes_{H_{i}}E_{2}E_{2}^{i}\bigr)\oplus\bigl(E_{1}^{2}(E_{1}^{\vee})^{i}\otimes_{H_{i}}E_{2}^{i}\bigr)

by

τ=(1⊗τ2​E2i00000100000000τ1​(E1∨)i⊗1).\tau=\left(\begin{matrix}1\otimes\tau_{2}E_{2}^{i}&0&0&0\\ 0&0&1&0\\ 0&0&0&0\\ 0&0&0&\tau_{1}(E_{1}^{\vee})^{i}\otimes 1\end{matrix}\right).

This construction provides the differential 22-category of right finite 22-representations on differential algebras with a monoidal structure.

5.4. Dual diagonal action

5.4.1. Algebra

Let BB be a differential algebra endowed with two 22-representations (F1,τ1)(F_{1},\tau_{1}) and (E2,τ2)(E_{2},\tau_{2}) together with a closed morphism λ:F1​E2→E2​F1\lambda:F_{1}E_{2}\to E_{2}F_{1} such that the diagrams (4.2.1) commute.

We define the differential algebra

A=Δλ​(B)=⨁i≥0(E2i​F1i)/((Tr⊗1)​x−(1⊗Tr)​x)x∈E2i​F1i, 1≤r<i.A=\Delta_{\lambda}(B)=\bigoplus_{i\geq 0}(E_{2}^{i}F_{1}^{i})/((T_{r}\otimes 1)x-(1\otimes T_{r})x)_{x\in E_{2}^{i}F_{1}^{i},\ 1\leq r<i}.

Its multiplication is given by the maps μi,j=μ(i,i),(j,j)​E2i​F1i​E2j​F1j→E2i+j​F1i+j\mu_{i,j}=\mu_{(i,i),(j,j)}E_{2}^{i}F_{1}^{i}E_{2}^{j}F_{1}^{j}\to E_{2}^{i+j}F_{1}^{i+j} defined in §4.2.1.

Given MM a differential AA-module and given i≥1i\geq 1, we have differential BB-module maps ςi:E2i​F1i⊗BM→M\varsigma_{i}:E_{2}^{i}F_{1}^{i}\otimes_{B}M\to M. These make (M,(ςi)i)(M,(\varsigma_{i})_{i}) into an object of Δλ​(B​−diff)\Delta_{\lambda}(B\operatorname{\!-diff}\nolimits) and provides an isomorphism of differential categories Δλ​(B)​−diff→∼Δλ​(B​−diff)\Delta_{\lambda}(B)\operatorname{\!-diff}\nolimits\mathrel{\mathop{\kern 0.0pt\to}\limits^{\sim}}\Delta_{\lambda}(B\operatorname{\!-diff}\nolimits).

0P73

Remark 5.4.1. As in Remark 4.4.3, we obtain a lax bi-22-representation on BB by setting Ei,j=E2j​F1iE_{i,j}=E_{2}^{j}F_{1}^{i}. We have an injective morphism of differential algebras ΔE​(B)→Δλ​(B)\Delta_{E}(B)\to\Delta_{\lambda}(B).

Assume the morphisms (5.2.1) are isomorphisms for all ii (this holds for example if λ\lambda is an isomorphism). Then we have a canonical isomorphism ΔE​(B)→∼Δλ​(B)\Delta_{E}(B)\mathrel{\mathop{\kern 0.0pt\to}\limits^{\sim}}\Delta_{\lambda}(B). The algebra Δλ​(B)\Delta_{\lambda}(B) is generated by BB and E2​F1E_{2}F_{1}.

The map λ\lambda extends (uniquely) to a morphism of algebras Δλ′​(B)→Δλ​(B)\Delta^{\prime}_{\lambda}(B)\to\Delta_{\lambda}(B) that is the identity on BB. If λ\lambda is an isomorphism, then this map is an isomorphism Δλ′​(B)→∼Δλ​(B)\Delta^{\prime}_{\lambda}(B)\mathrel{\mathop{\kern 0.0pt\to}\limits^{\sim}}\Delta_{\lambda}(B).

5.4.2. Left dual

We assume now that F1F_{1} is left finite and we put E1=∨F1E_{1}={{}^{\vee}F}_{1}. Consider σ∈Z​Hom⁡(E2​E1,E1​E2)\sigma\in Z\operatorname{Hom}\nolimits(E_{2}E_{1},E_{1}E_{2}) defined as in (4.4.1).

Let π:E2⊗BA→E1⊗BA\pi:E_{2}\otimes_{B}A\to E_{1}\otimes_{B}A be the closed morphism of (B,A)(B,A)-bimodules given as a composition

π:E2⊗BA→E2​η1E2​E1​F1⊗BA→σ​F1E1​E2​E1∨⊗BA→E1​multE1⊗BA.\pi:E_{2}\otimes_{B}A\xrightarrow{E_{2}\eta_{1}}E_{2}E_{1}F_{1}\otimes_{B}A\xrightarrow{\sigma F_{1}}E_{1}E_{2}E_{1}^{\vee}\otimes_{B}A\xrightarrow{E_{1}\mathrm{mult}}E_{1}\otimes_{B}A.

We put E=cone⁡(π)E=\operatorname{cone}\nolimits(\pi). Given i≥1i\geq 1, we define a morphism of (B,B)(B,B)-bimodules ςi:E2i​F1i​E→E\varsigma_{i}:E_{2}^{i}F_{1}^{i}E\to E

ςi=(E2mult∘λ(1⋯2i+1)∑r=1iE2mult∘E2iF1i−1ε1∘λ(1⋯r)(2i⋯i+r)0E1mult∘λ(1⋯2i+1))\varsigma_{i}=\left(\begin{matrix}E_{2}\mathrm{mult}\circ\lambda_{(1\cdots 2i+1)}&\sum_{r=1}^{i}E_{2}\mathrm{mult}\circ E_{2}^{i}F_{1}^{i-1}\varepsilon_{1}\circ\lambda_{(1\cdots r)(2i\cdots i+r)}\\ 0&E_{1}\mathrm{mult}\circ\lambda_{(1\cdots 2i+1)}\end{matrix}\right)

The following lemma is a consequence of Lemmas 4.3.5 and 4.3.7 applied to m=Am=A.

0P74

Lemma 5.4.2. The ςi\varsigma_{i}’s define a left action of AA on EE, giving EE a structure of differential (A,A)(A,A)-bimodule.

Note that the isomorphism of differential categories Δλ​(B)​−diff→∼Δλ​(B​−diff)\Delta_{\lambda}(B)\operatorname{\!-diff}\nolimits\mathrel{\mathop{\kern 0.0pt\to}\limits^{\sim}}\Delta_{\lambda}(B\operatorname{\!-diff}\nolimits) commutes with EE.

Assume now σ\sigma is an isomorphism. We define τ\tau a (B,A)(B,A)-bimodule endomorphism of E2E^{2} as in (5.3.4).

Theorem 4.4.15 has the following consequence.

0P75

Theorem 5.4.3. The data (E,τ)(E,\tau) defines a 22-representation on Δλ​(B)\Delta_{\lambda}(B).

Note that we have an isomorphism of 22-representations Δλ​(B)​−diff→∼Δλ​(B​−diff)\Delta_{\lambda}(B)\operatorname{\!-diff}\nolimits\mathrel{\mathop{\kern 0.0pt\to}\limits^{\sim}}\Delta_{\lambda}(B\operatorname{\!-diff}\nolimits).

Consider the (Δσ​(B),Δλ​(B))(\Delta_{\sigma}(B),\Delta_{\lambda}(B))-bimodule Δλ​(B)\Delta_{\lambda}(B), where the right action is given by multiplication and the left action by multiplication preceded by the morphism of algebras Δσ​(B)=Δλ′​(B)→Δλ​(B)\Delta_{\sigma}(B)=\Delta^{\prime}_{\lambda}(B)\to\Delta_{\lambda}(B). It follows from Proposition 4.4.16 that this bimodule induces a morphism of 22-representations from Δλ​(B)\Delta_{\lambda}(B) to Δσ​(B)\Delta_{\sigma}(B).

5.5. Differential categories

5.5.1. Bimodule 22-representations

All the definitions and constructions of §5.1–5.4 extend from the setting of differential algebras to that of differential categories. We will describe this explicitly.

We view the monoidal category 𝒰{\mathcal{U}} as a 22-category with one object ∗\ast.

0P76

Definition 5.5.1. A bimodule 22-representation is the data of a 22-functor Υ:𝒰→Bimod\Upsilon:{\mathcal{U}}\to\mathrm{Bimod}.

It is right finite if Υ⁡(e)\Upsilon(e) is right finite.

We say that Υ\Upsilon is a bimodule 22-representation on Υ⁡(∗)\Upsilon(\ast).

Bimodule 22-representations form a differential 22-category.

Let 𝒞{\mathcal{C}} be a differential category. There are equivalences of differential 22-categories between

  • •

    the 22-category of bimodule 22-representations Υ\Upsilon on 𝒞{\mathcal{C}}

  • •

    the 22-category with objects differential functors M:𝒞×𝒞opp×𝒰→k​−diffM:{\mathcal{C}}\times{\mathcal{C}}^{\operatorname{opp}\nolimits}\times{\mathcal{U}}\to k\operatorname{\!-diff}\nolimits together with

    • –

      isomorphisms μm,n:M⁡(c,−,em)⊗𝒞M⁡(−,c′,en)→∼M⁡(c,c′,en+m)\mu_{m,n}:M(c,-,e^{m})\otimes_{{\mathcal{C}}}M(-,c^{\prime},e^{n})\mathrel{\mathop{\kern 0.0pt\to}\limits^{\sim}}M(c,c^{\prime},e^{n+m}) functorial in cc and c′c^{\prime}, compatible with the canonical morphism End⁡(em)⊗End⁡(en)→End⁡(en+m)\operatorname{End}\nolimits(e^{m})\otimes\operatorname{End}\nolimits(e^{n})\to\operatorname{End}\nolimits(e^{n+m}) and satisfying μl,n+m∘(id⊗μm,n)=μm+l,n∘(μl,m⊗id)\mu_{l,n+m}\circ(\operatorname{id}\nolimits\otimes\mu_{m,n})=\mu_{m+l,n}\circ(\mu_{l,m}\otimes\operatorname{id}\nolimits)

    • –

      an isomorphism μ0:M⁡(−,−,e0)→∼Id\mu_{0}:M(-,-,e^{0})\mathrel{\mathop{\kern 0.0pt\to}\limits^{\sim}}\operatorname{Id}\nolimits such that μm,0=mult∘(M⁡(c,−,em)⊗μ0)\mu_{m,0}=\mathrm{mult}\circ(M(c,-,e^{m})\otimes\mu_{0}) and μ0,m=mult∘(μ0⊗M⁡(−,c,em))\mu_{0,m}=\mathrm{mult}\circ(\mu_{0}\otimes M(-,c,e^{m}))

  • •

    the 22-category of pairs (E,τ)(E,\tau) where EE is a (𝒞,𝒞)({\mathcal{C}},{\mathcal{C}})-bimodule and τ∈End⁡(E2)\tau\in\operatorname{End}\nolimits(E^{2}) satisfies (4.1.1).

The category ℋ​o​m​((𝒞,E,τ),(𝒞′,E′,τ′)){\mathcal{H}}{om}(({\mathcal{C}},E,\tau),({\mathcal{C}}^{\prime},E^{\prime},\tau^{\prime})) of 11-arrows in the third 22-category above has objects pairs (P,φ)(P,\varphi) where PP is a (𝒞′,𝒞)({\mathcal{C}}^{\prime},{\mathcal{C}})-bimodule and φ:P⊗𝒞E→∼E′⊗𝒞′P′\varphi:P\otimes_{\mathcal{C}}E\mathrel{\mathop{\kern 0.0pt\to}\limits^{\sim}}E^{\prime}\otimes_{{\mathcal{C}}^{\prime}}P^{\prime} is a closed isomorphism of (𝒞′,𝒞)({\mathcal{C}}^{\prime},{\mathcal{C}})-bimodules satisfying (5.1.1). We leave it to the reader to describe 11-arrows in the second 22-category above. In these 22-categories, the 22-arrows are morphisms of (non-differential) bimodules or functors compatible with the additional structure.

The equivalences are given by

Υ↦(M:(c1,c2,en)↦Υ(en)(c1,c2)),M↦(E=M(−,−,e),τ=M(−,−,τ))\Upsilon\mapsto(M:(c_{1},c_{2},e^{n})\mapsto\Upsilon(e^{n})(c_{1},c_{2})),\ M\mapsto(E=M(-,-,e),\tau=M(-,-,\tau))
E↦(Υ:en↦En).E\mapsto(\Upsilon:e^{n}\mapsto E^{n}).

We will use the terminology “bimodule 22-representation” for either one of those three equivalent structures.

Note that a 22-representation Υ:𝒰→End⁡(𝒞)\Upsilon:{\mathcal{U}}\to\operatorname{End}\nolimits({\mathcal{C}}) gives rise to a bimodule 22-representation MM on 𝒞{\mathcal{C}} given by M⁡(c1,c2,en)=Hom𝒞⁡(c2,Υ∘rev⁡(en)​(c1))M(c_{1},c_{2},e^{n})=\operatorname{Hom}\nolimits_{\mathcal{C}}(c_{2},\Upsilon\circ\mathrm{rev}(e^{n})(c_{1})) (cf §2.2.3). Note also that a bimodule 22-representation MM on a differential category 𝒞{\mathcal{C}} gives rise to a 22-representation Υ:𝒰→End⁡(𝒞​−diff)\Upsilon:{\mathcal{U}}\to\operatorname{End}\nolimits({\mathcal{C}}\operatorname{\!-diff}\nolimits) given by Υ(en)=M(−,−,en)⊗𝒞−\Upsilon(e^{n})=M(-,-,e^{n})\otimes_{\mathcal{C}}-.

5.5.2. Diagonal action

A bimodule lax bi-22-representation is a lax differential 22-functor Υ:𝒰⊗𝒰→Bimod\Upsilon:{\mathcal{U}}\otimes{\mathcal{U}}\to\mathrm{Bimod}. We say it is a bimodule lax bi-22-representation on Υ(∗⊗∗)\Upsilon(\ast\otimes\ast).

A bimodule lax bi-22-representation on 𝒞{\mathcal{C}} is the same as the data of

  • •

    (𝒞,𝒞)({\mathcal{C}},{\mathcal{C}})-bimodules Ei,jE_{i,j} for i,j≥0i,j\geq 0

  • •

    morphisms of differential algebras Hi⊗Hj→End⁡(Ei,j)H_{i}\otimes H_{j}\to\operatorname{End}\nolimits(E_{i,j})

  • •

    morphisms μ(i,j),(i′,j′):Ei,j​Ei′,j′→Ei+i′,j+j′\mu_{(i,j),(i^{\prime},j^{\prime})}:E_{i,j}E_{i^{\prime},j^{\prime}}\to E_{i+i^{\prime},j+j^{\prime}} satisfying properties (1) and (2) of §4.2.1.

We define the differential category ΔE​(𝒞)\Delta_{E}({\mathcal{C}}) as the additive category quotient of T𝒞​(E0,1​E1,0)T_{{\mathcal{C}}}(E_{0,1}E_{1,0}) by the ideal of maps generated by the kernels of the compositions

(E0,1​E1,0)i​(c1,c2)→canEi,i​(c1,c2)→canEi,i​(c1,c2)/((Tr⊗1)​x−(1⊗Tr)​x)x∈Ei,i, 1≤r<i.(E_{0,1}E_{1,0})^{i}(c_{1},c_{2})\xrightarrow{{\mathrm{can}}}E_{i,i}(c_{1},c_{2})\xrightarrow{{\mathrm{can}}}E_{i,i}(c_{1},c_{2})/((T_{r}\otimes 1)x-(1\otimes T_{r})x)_{x\in E_{i,i},\ 1\leq r<i}.

Assume now 𝒞{\mathcal{C}} is a differential category endowed with two structures (F1,τ1)(F_{1},\tau_{1}) and (E2,τ2)(E_{2},\tau_{2}) of bimodule 22-representations together with a closed morphism λ:F1​E2→E2​F1\lambda:F_{1}E_{2}\to E_{2}F_{1} such that the diagrams (4.2.1) commute.

We define the differential category Δλ′​(𝒞)\Delta^{\prime}_{\lambda}({\mathcal{C}}) as the additive category quotient of T𝒞​(F1​E2)T_{\mathcal{C}}(F_{1}E_{2}) by the ideal of maps generated by the image of the composition

F12​E22​(c1,c2)→τ1​E22−F12​τ2F12​E22​(c1,c2)→F1​λ​E2(F1​E2)2​(c1,c2).F_{1}^{2}E_{2}^{2}(c_{1},c_{2})\xrightarrow{\tau_{1}E_{2}^{2}-F_{1}^{2}\tau_{2}}F_{1}^{2}E_{2}^{2}(c_{1},c_{2})\xrightarrow{F_{1}\lambda E_{2}}(F_{1}E_{2})^{2}(c_{1},c_{2}).

We have a differential category 𝒞′=⨁i≥0E2i​F1i{\mathcal{C}}^{\prime}=\bigoplus_{i\geq 0}E_{2}^{i}F_{1}^{i}. Its objects are those of 𝒞{\mathcal{C}} and Hom𝒞′⁡(c1,c2)=⨁i≥0E2i​F1i​(c1,c2)\operatorname{Hom}\nolimits_{{\mathcal{C}}^{\prime}}(c_{1},c_{2})=\bigoplus_{i\geq 0}E_{2}^{i}F_{1}^{i}(c_{1},c_{2}). The multiplication is induced by the maps μi,j\mu_{i,j}. We define the differential category Δλ​(𝒞)\Delta_{\lambda}({\mathcal{C}}) as the additive category quotient of ⨁i≥0E2i​F1i\bigoplus_{i\geq 0}E_{2}^{i}F_{1}^{i} by the ideal of maps generated by the images of Tr⊗1−1⊗Tr:E2i​F1i→E2i​F1iT_{r}\otimes 1-1\otimes T_{r}:E_{2}^{i}F_{1}^{i}\to E_{2}^{i}F_{1}^{i} for 1≤r<i1\leq r<i.

Assume now 𝒞{\mathcal{C}} is a differential category endowed with two structures (E1,τ1)(E_{1},\tau_{1}) and (E2,τ2)(E_{2},\tau_{2}) of bimodule 22-representations, the first of which is right finite. Consider σ:E2​E1→E1​E2\sigma:E_{2}E_{1}\to E_{1}E_{2} closed such that the diagrams (4.3.1) commute. We define λ:E1∨​E2→E2​E1∨\lambda:E_{1}^{\vee}E_{2}\to E_{2}E_{1}^{\vee} as in (5.3.1).

∙\bullet\ We put Δσ​(𝒞)=Δλ′​(𝒞)\Delta_{\sigma}({\mathcal{C}})=\Delta^{\prime}_{\lambda}({\mathcal{C}}). As in §5.3.3, we define a (Δσ​𝒞,𝒞)(\Delta_{\sigma}{\mathcal{C}},{\mathcal{C}})-bimodule EE and extend it to a (Δσ​𝒞,Δσ​𝒞)(\Delta_{\sigma}{\mathcal{C}},\Delta_{\sigma}{\mathcal{C}})-bimodule. Assume finally that σ\sigma is invertible. We construct in addition an endomorphism τ\tau of E2E^{2}. We obtain a bimodule 22-representation on Δσ​𝒞\Delta_{\sigma}{\mathcal{C}} and an isomorphism of 22-representations Δσ​(𝒞​−diff)→∼Δσ​(𝒞)​−diff\Delta_{\sigma}({\mathcal{C}}\operatorname{\!-diff}\nolimits)\mathrel{\mathop{\kern 0.0pt\to}\limits^{\sim}}\Delta_{\sigma}({\mathcal{C}})\operatorname{\!-diff}\nolimits. The 22-representation is right finite if E2E_{2} is right finite.

As in §5.3.4, we have a monoidal structure on the differential 22-category of right finite bimodule 22-representations.

∙\bullet\ We drop now the assumption that σ\sigma is invertible. We define as in §5.4.2 a (Δλ​𝒞,Δλ​𝒞)(\Delta_{\lambda}{\mathcal{C}},\Delta_{\lambda}{\mathcal{C}})-bimodule EE. Assume σ\sigma is invertible. We obtain an endomorphism τ\tau of E2E^{2} and a bimodule 22-representation on Δλ​𝒞\Delta_{\lambda}{\mathcal{C}}.

5.6. Pointed categories

Let 𝒱{\mathcal{V}} be a differential pointed category. A bimodule 22-represesentation on 𝒱{\mathcal{V}} is the data of a strict monoidal differential pointed functor from the 22-category with one object given by 𝒰∙{\mathcal{U}}^{\bullet} to Bimod∙\mathrm{Bimod}^{\bullet}. Note that a bimodule 22-representation on 𝒱{\mathcal{V}} gives rise to a bimodule 22-representation on k⁡[𝒱]k[{\mathcal{V}}].

A bimodule lax bi-22-representation is a lax differential pointed 22-functor Υ:𝒰∙∧𝒰∙→Bimod∙\Upsilon:{\mathcal{U}}^{\bullet}\wedge{\mathcal{U}}^{\bullet}\to\mathrm{Bimod}^{\bullet}. We say it is a bimodule lax bi-22-representation on Υ(∗∧∗)\Upsilon(\ast\wedge\ast).

A bimodule lax bi-22-representation on 𝒱{\mathcal{V}} is the same as the data of

  • •

    (𝒱,𝒱)({\mathcal{V}},{\mathcal{V}})-bimodules Ei,jE_{i,j} for i,j≥0i,j\geq 0

  • •

    morphisms of differential pointed algebras Hi∧Hj→End⁡(Ei,j)H_{i}\wedge H_{j}\to\operatorname{End}\nolimits(E_{i,j})

  • •

    morphisms μ(i,j),(i′,j′):Ei,j​Ei′,j′→Ei+i′,j+j′\mu_{(i,j),(i^{\prime},j^{\prime})}:E_{i,j}E_{i^{\prime},j^{\prime}}\to E_{i+i^{\prime},j+j^{\prime}} satisfying properties (1) and (2) of §4.2.1.

We define the differential pointed category ΔE​(𝒱)\Delta_{E}({\mathcal{V}}) as the quotient of T𝒱​(E0,1​E1,0)T_{{\mathcal{V}}}(E_{0,1}E_{1,0}) by the equivalence relation generated by f∼f′f\sim f^{\prime} if (f,f′)(f,f^{\prime}) is in the equalizer of a composition

(E0,1​E1,0)i​(c1,c2)→canEi,i​(c1,c2)→canEi,i​(c1,c2)/((Tr∧1)​x∼(1∧Tr)​x)x∈Ei,i, 1≤r<i.(E_{0,1}E_{1,0})^{i}(c_{1},c_{2})\xrightarrow{{\mathrm{can}}}E_{i,i}(c_{1},c_{2})\xrightarrow{{\mathrm{can}}}E_{i,i}(c_{1},c_{2})/((T_{r}\wedge 1)x\sim(1\wedge T_{r})x)_{x\in E_{i,i},\ 1\leq r<i}.

Consider a differential pointed category 𝒱{\mathcal{V}} endowed with two bimodule 22-representations (F1,τ1)(F_{1},\tau_{1}) and (E2,τ2)(E_{2},\tau_{2}) and a closed morphism λ:F1​E2→E2​F1\lambda:F_{1}E_{2}\to E_{2}F_{1} such that the diagrams (4.2.1) commute.

We define the differential pointed category Δλ′​(𝒱)\Delta^{\prime}_{\lambda}({\mathcal{V}}) as the quotient of T𝒱​(F1​E2)T_{{\mathcal{V}}}(F_{1}E_{2}) by the equivalence relation generated by

(F1​λ​E2)∘(τ1​E22)​(f)∼(F1​λ​E2)∘(F12​τ2)​(f)​ for ​f∈F12​E22​(c1,c2)​ and ​c1,c2∈𝒱.(F_{1}\lambda E_{2})\circ(\tau_{1}E_{2}^{2})(f)\sim(F_{1}\lambda E_{2})\circ(F_{1}^{2}\tau_{2})(f)\text{ for }f\in F_{1}^{2}E_{2}^{2}(c_{1},c_{2})\text{ and }c_{1},c_{2}\in{\mathcal{V}}.

We define the differential pointed category Δλ​(𝒱)\Delta_{\lambda}({\mathcal{V}}). We consider first the differential pointed category with same objects as 𝒱{\mathcal{V}} and pointed set of maps v1→v2v_{1}\to v_{2} given by ⋁i≥0E2i​F1i​(v1,v2)\bigvee_{i\geq 0}E_{2}^{i}F_{1}^{i}(v_{1},v_{2}). The category Δλ​(𝒱)\Delta_{\lambda}({\mathcal{V}}) is the quotient of that category by the equivalence relation generated by (Tr∧1)​(f)∼(1∧Tr)​(f)(T_{r}\wedge 1)(f)\sim(1\wedge T_{r})(f) for f∈E2i​F1if\in E_{2}^{i}F_{1}^{i} and 1≤r<i1\leq r<i.

Note that there is a canonical isomorphism of differential categories for ?∈{∅,′}?\in\{\emptyset,\prime\}

k⁡[Δλ?​(𝒱)]→∼Δλ?​(k⁡[𝒱])k[\Delta^{?}_{\lambda}({\mathcal{V}})]\mathrel{\mathop{\kern 0.0pt\to}\limits^{\sim}}\Delta^{?}_{\lambda}(k[{\mathcal{V}}])

5.7. Douglas-Manolescu’s algebra-modules

Let us recall some aspects of Douglas-Manolescu’s theory [DouMa].

Note that Douglas and Manolescu work in the differential graded setting, and we translate their constructions to the differential setting.

Their nil-Coxeter 22-algebra [DouMa, §2.2] can be viewed as the same data as our monoidal category 𝒰{\mathcal{U}} (cf [DouMa, Remark 2.4]). A bottom-algebra module [DouMa, §2.4] for the nil-Coxeter 22-algebra is the same data as a lax bimodule 22-representation on a differential algebra AA, where a lax bimodule 22-representation on AA is defined to be a lax 22-functor Υ:𝒰→Bimod\Upsilon:{\mathcal{U}}\to\mathrm{Bimod} with Υ⁡(1)\Upsilon(1) the differential category with one object whose endomorphism ring is AA. They also consider top-algebra modules, where 𝒰{\mathcal{U}} above is replaced by 𝒰opp{\mathcal{U}}^{\operatorname{opp}\nolimits}. Using the isomorphism 𝒰→∼𝒰opp{\mathcal{U}}\mathrel{\mathop{\kern 0.0pt\to}\limits^{\sim}}{\mathcal{U}}^{\operatorname{opp}\nolimits} (§4.1.1), a top-algebra module can be viewed as a bottom-algebra module, hence as a lax bimodule 22-representation.

Douglas and Manolescu define a tensor product of a top algebra-module and a bottom algebra-module [DouMa, Definition 2.11]. This corresponds to our construction of a differential algebra AA as a tensor product ⊗⁣○{\hskip 1.9919pt\otimes\hskip-10.81218pt\bigcirc}. Note that they do not endow this tensor product with any algebra-module structure.

Original mathematics by the credited authors. Source collection and HTML conversion remain in progress.

Andrew Manion, Raphael Rouquier

Original source: arXiv:2009.09627v2