Let \(R\) be a commutative ring and let \(A\) be a differential graded \(R\)-algebra with multiplication \(\mu \). How does its multiplication give a multiplication in the derived category? Writing \(\gamma \colon \Ch (R)\to \D (R)\) for the localization functor, we would like to form the composite \[ \gamma A\otimes _R^{\bL }\gamma A \longrightarrow \gamma (A\otimes _R A) \xrightarrow {\gamma (\mu )}\gamma A. \] For \(A\) bounded below, we can describe the first map using Definition 6.6.25. Choose a projective resolution \(P\to A\). Then \(P\otimes _R P\) represents the derived tensor product, and the resolution map induces \(P\otimes _R P\to A\otimes _R A\). We will show that these maps, for arbitrary pairs of complexes, equip \(\gamma \) with a lax symmetric monoidal structure. It then follows that the displayed multiplication makes \(\gamma A\) an associative algebra in \(\D (R)\), and a commutative algebra if \(A\) is commutative (Corollary 21.2.3).
More generally, suppose we wish to invert a class of morphisms \(W\) in a symmetric monoidal \(\infty \)-category \(C\). If tensoring with any object preserves \(W\), the tensor product descends to \(C[W^{-1}]\) and the localization functor is symmetric monoidal (Proposition 21.1.1). For chain complexes this hypothesis can fail even when tensoring with a module concentrated in degree zero, as we saw for \(\Z /2\) in Example 6.6.24. Following Nikolaus and Scholze (2018), Appendix A, we instead invert the tuples of weak equivalences in each fiber of \(C^\otimes \to \Span (\Fin )\). When the tensor products admit absolute left derived functors compatible with composition, these derived functors give the transport along spans in the localized \(\infty \)-category. In this way we obtain both the symmetric monoidal structure on \(C[W^{-1}]\) and the lax symmetric monoidal localization functor to it (Theorem 21.1.6).
We apply this construction to unbounded chain complexes. The resulting tensor product on \(\D (R)\) extends the bounded below derived tensor product of Chapter 6 and preserves colimits separately in each variable (Proposition 21.2.2, Proposition 21.2.4). Its unit \(R[0]\) is a compact generator. The compact objects are the perfect complexes, and we will see that their Ind-completion recovers \(\D (R)\) together with its tensor product (Corollary 21.2.6). These are the facts about \(\D (R)\) used in Corollary 8.2.4 to identify it, by Morita theory, with the \(\infty \)-category of modules over the Eilenberg–MacLane ring spectrum of \(R\).
Sections
Monoidal Dwyer–Kan localizations
We first consider the case in which tensoring with any object of \(C\) preserves the morphisms to be inverted.
The symmetric monoidal derived category
### Part III: Background on ∞-category theory.