Vanishing lines for modules over the motivic Steenrod algebra
Drew Heard, Achim Krause

TL;DR
This paper investigates conditions under which modules over specific subalgebras of the motivic Steenrod algebra exhibit freeness and possess a vanishing line, using Margolis homology to establish these criteria.
Contribution
It introduces new criteria based on Margolis homology for determining freeness and vanishing lines in modules over motivic Steenrod algebra subalgebras.
Findings
Criteria for freeness of modules established
Vanishing lines characterized via Margolis homology
Results applicable to modules over motivic Steenrod algebra at prime 2
Abstract
We study criteria for freeness and for the existence of a vanishing line for modules over certain Hopf subalgebras of the motivic Steenrod algebra over at the prime 2. These turn out to be determined by the vanishing of certain Margolis homology groups in the quotient Hopf algebra .
Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
Taxonomy
TopicsHomotopy and Cohomology in Algebraic Topology · Algebraic structures and combinatorial models · Advanced Topics in Algebra
\newalphalph\myfnsymbolmult
[mult]
Vanishing lines for modules over the motivic Steenrod algebra
Drew Heard
Fachbereich Mathematik der Universität Hamburg, Bundesstraße 55, 20146 Hamburg, Germany
and
Achim Krause
Max-Planck-Institute for Mathematics, Vivatsgasse 7, 53111 Bonn, Germany
Abstract.
We study criteria for freeness and for the existence of a vanishing line for modules over certain Hopf subalgebras of the motivic Steenrod algebra over at the prime 2. These turn out to be determined by the vanishing of certain Margolis homology groups in the quotient Hopf algebra .
keywords:
cohomology of the Steenrod algebra, vanishing line, motivic homotopy theory.
1991 Mathematics Subject Classification:
14F42,55S10
1. Introduction
Let be the classical mod Steenrod algebra. Criteria for freeness and for the existence of a vanishing line for modules over Hopf subalgebras of have been obtained by Adams–Margolis [AM74, AM71], Moore–Peterson [MP72], Anderson–Davis [AD73], and Miller–Wilkerson [MW81], with special cases also considered by other authors. These results are of great importance; for example, they are used critically in the proof of the nilpotence and periodicity theorems of Devinatz, Hopkins, and Smith [DHS88, HS98].
Now let be the motivic Steenrod algebra. We work over (although the results also hold more generally for algebraically closed fields, see Remark 5.5) and at , since the odd-primary motivic Steenrod algebra is a base-changed version of the classical odd-primary Steenrod algebra. By work of Voevodsky [Voe03b], as an algebra, the dual of is given by
[TABLE]
where is the motivic cohomology of a point (see Section 2 for more details on and .
Remark 1.1**.**
We remind the reader that motivic objects are bigraded; they have internal degrees (the topological degree) and (the weight). However, the weight will not play a major role in this paper, and we will usually suppress it from the notation, considering motivic objects as singly graded.
As shown in [GIR16] and [GI15] the most naive adaptation of the classical results on freeness and vanishing lines cannot hold for . Indeed, note the following, where we refer the reader to [Mar83, Ch. 19 and Ch. 20] for the definition of Margolis homology groups:
- (1)
Let be the Hopf subalgebra of the 2-primary classical Steenrod algebra generated by and , and its motivic analog. Then an -module is projective (equivalently, free, since is a connective Hopf algebra over a field, see [Mar83, Prop. 11.2.2]) if and only if the Margolis homology groups and vanish. In [GIR16, Ex. 4.6] the authors give an example of an -module with vanishing and that is not projective. Furthermore, in Prop. 4.7 of *loc. cit. * they show that a third Margolis homology group must vanish, as well as requiring to be a free -module, in order for to be a projective -module (note that the latter condition is automatic classically, since we work over ). 2. (2)
Let be the Hopf subalgebra of the 2-primary classical Steenrod algebra generated by , and its motivic analog. Adams [Ada66, Prop. 2.5] proves that if is a bounded-below -module which is free over , then has a vanishing line of slope 1/2 (that is, when drawn in the plane, the groups vanish above a line of slope 1/2). Motivically, Guillou and Isaksen [GI15, Prop. 7.2] prove that if is a bounded-below -module which is free over , then above a line of slope , all the elements of are -local (here again, the slope refers to only the plane. They also observed that there is a vanishing line, but now of slope ; that is, whenever for some constant .
The goal of this paper is to explain results such as these. The key observation, in a sense already noted in [GIR16], is that it is enough to work with the quotient Hopf algebra . We shall see that is a (bigraded) 2-primary version of the classical odd-primary Steenrod algebra. This already demystifies the above results partially; for odd primes, freeness of an -module does require the vanishing of three Margolis homology groups. We shall also see how to explain the vanishing line of slope 1, as well as the -local part, see 5.7. In future work with Tobias Barthel we will use our main theorem to investigate periodicity and vanishing line results in motivic homotopy theory akin to this example.
Classically Adams and Margolis [AM74] have shown that all Hopf subalgebras of the odd-primary are the obvious ones; namely the duals of the quotients
[TABLE]
where and , and the sequences and must satisfy certain conditions, see Condition 2.1 and 2.2 of [AM74]. The pair is called the profile function of the Hopf subalgebra. The first step in our work is to construct certain Hopf subalgebras of by a similar method. There is an added complexity arising from the fact that is not a field. Using the notion of a free profile function , originally introduced in the -equivariant Steenrod algebra by Ricka [Ric15], we construct certain quotient Hopf algebras that are free as -modules. In fact every quotient Hopf algebra of that is free as an -module is of this form.
Theorem A**.**
A quotient Hopf algebra of is free as an -module if and only if it is of the form for a profile function satisfying certain numerical conditions.
For a precise version of the theorem, see 3.7.
In order to state our main result, let for be dual to and be dual to . We shall see that these also make sense in the quotient , and there they satisfy for and , so we can define Margolis homology groups and for an -module . For the following, compare the main theorem of [MW81]. Note that we write for the topological degree. The following is given as 5.4.
Theorem B**.**
Let be a Hopf subalgebra of which is free as an -module. Let be a -module that is free and of finite type over .
- (i)
If for all with and for all , then is -free. 2. (ii)
Let be an integer. If for all such that and and if for all such that , then for all such that with depending only on and the connectivity of .
We note that if is a -module, which is free and of finite type over , such that has the property that acts trivially on , then if and only if see 5.3.
This paper is organized as follows: we first give an introduction to the motivic Steenrod algebra. We then construct certain Hopf subalgebras of . The proof of our main theorem is carried out in Section 5, and relies on a reduction to , which is studied in more detail in Section 4.
Acknowledgments
The first author is grateful to the Max Planck Institute for Mathematics and the Universität Hamburg for hospitality, and to the Max Planck Institute and DFG Schwerpunktprogramm SPP 1786 for financial support. The second author is grateful to the Max Planck Institute for Mathematics and the International Max Planck Research School for hospitality and financial support. The authors thank Dan Isaksen and the anonymous reviewer for helpful comments on earlier versions of this document.
2. Background
Let be the homotopy category of motivic spectra over , and let . We denote by the motivic cohomology of a point, where has bidegree . Let denote the motivic Steenrod algebra, and its dual. The properties of and have been determined by Voevodsky in [Voe03b, Voe03a], We give a summary of what we need here - see in particular [Voe03b, Prop. 10.2, Thm. 12.6, and Lem. 12.11], with some typos in the Adem relations corrected in [Rio12, Thm. 4.5.1]. In order to state these results, we note that a sequence of integers is called admissible if either or and for all .
Theorem 2.1** (Voevodsky).**
- (i)
The motivic Steenrod algebra is the associative algebra over generated by and for subject to the following relations, where :
[TABLE]
where
[TABLE] 2. (ii)
The motivic Steenrod algebra is a free -module on the admissible monomials . 3. (iii)
The dual motivic Steenrod algebra is given as an algebra by
[TABLE]
where the bigrading is given by , and . 4. (iv)
The coproduct in is given by
[TABLE]
Remark 2.2**.**
- (i)
can be identified with the Bockstein associated to the short exact sequence
[TABLE]
Note that the Adem relations show that for all , where . 2. (ii)
Note that is given the homological grading, so acts by bidegree on it.
From the description of the dual, we immediately see:
Lemma 2.3**.**
After inverting , there is an isomorphism of Hopf algebras , where is the classical dual Steenrod algebra. The isomorphism is given by
[TABLE]
Dually, there is an isomorphism .
Definition 2.4**.**
We will refer to the embedding as the embedding into the classical dual Steenrod algebra, since its codomain can be thought of as the classical dual Steenrod algebra over the graded field . Similarly, we refer to the embedding as the embedding into the classical Steenrod algebra.
The following is straightforward and can be proved in the same way as the classical odd-primary case.
Lemma 2.5**.**
Let be a sequence of ones and zeroes, almost all zero, and a sequence of nonnegative integers, almost all zero. As an -module, is free with basis
[TABLE]
Definition 2.6**.**
Let be the dual monomial basis of . Define
[TABLE]
and
[TABLE]
Additionally let be dual to , i.e.,
[TABLE]
where the 1 is in the -th spot. Finally, for , define , where the is in the -th position, to be the class dual to (note that the indexing on starts at position 0, and that on starts at position 1).
The indexing is chosen such that the motivic maps to the classical one under the natural isomorphism . More precisely, we have:
Lemma 2.7**.**
Under the embedding into the classical Steenrod algebra, goes to a -multiple of the classical Milnor basis element .
Proof.
Dually, this corresponds to the fact that monomials in the motivic dual Steenrod algebra go to -multiples of the corresponding monomials in the classical dual Steenrod algebra, which is immediate from the definition of the embedding. ∎
Remark 2.8**.**
This enables us to recover multiplicative relations between the from the corresponding relations between the , since the powers of that occur are determined by weight. For example, the -motivic Adem relations can be deduced from this.
Furthermore, one can also use this to deduce that the classical recursion formula for the given by
[TABLE]
also holds in the motivic Steenrod algebra. This also appears in [Kyl17, Eq. (2)] over more general base fields (in general, the result requires a correction term when ).
Lemma 2.9**.**
We have and whenever .
Proof.
This also follows immediately from the corresponding relations in the classical Steenrod algebra. ∎
Remark 2.10**.**
For , is nonzero. For example, at , , and (note that non-motivically, we have ). Note though that ; we shall see in Section 4 that this type of result holds for all with .
3. Some quotient Hopf algebras of
In this section, we introduce certain quotient Hopf algebras of , and identify conditions that ensure that they are free as -modules. The results are similar to the classification of quotient Hopf algebras of the classical odd-primary Steenrod algebra, although of course classically we do not need to worry about freeness.
To begin, let be a function from the set to the set , and let be a function from the set to the set . We call the pair a profile function.
Definition 3.1**.**
The quotient algebra is the quotient of by the relations
[TABLE]
with the convention that if or , then we impose no relation. We let denote the ideal of generated by these relations.
We point out two interesting properties in contrast to the classical Steenrod algebra: Firstly, need not be free as an module, and, secondly, the relation in implies that two sequences and can give rise to isomorphic quotient algebras of .
We deal with the latter issue first, by defining a partial order on profile functions by if we have and for all . This leads to the following definition, originally due to Ricka [Ric15, Def. 5.9].
Definition 3.2**.**
A profile function is minimal if it is minimal among profile functions such that .
The following is then proved in the same way as [Ric15, Lem. 5.10].
Lemma 3.3**.**
A profile function is minimal if and only if for all , is equivalent to .
We now introduce conditions that ensure that is a quotient Hopf algebra of .
Condition 3.1**.**
For all we have or .
Condition 3.2**.**
For all we have or .
For the following compare [Ric15, Prop. 5.13] or the odd-primary case of [AM74].
Proposition 3.4**.**
Let be a minimal profile functions satisfying Conditions 3.1 and 3.2. Then the quotient algebra is a quotient Hopf algebra of .
Proof.
We simply need to check that the coproduct passes through to the quotient. To that end, note that we have
[TABLE]
and
[TABLE]
Assuming Condition 3.1, then either or are in . By definition is in , whilst if we assume Condition 3.2 then either or are in . Note that the latter condition is equivalent to by minimality. It follows that passes to the quotient as required. ∎
Recall that the classical Steenrod algebra is a Hopf algebra of finite type; that is, it is finitely generated as an -module in each degree. Since we are working over a field it is in addition automatically free as an -module in each degree. For such Hopf algebras, the dual inherits the structure of a Hopf algebra. One then easily checks that there is a one-to-one correspondence between quotient Hopf algebras of the dual Steenrod algebra, and Hopf subalgebras of the Steenrod algebra. Indeed, as noted in the introduction, the classification of Hopf subalgebras of is proved via the classification of all quotient Hopf algebras of .
The situation in the motivic Steenrod algebra is more complicated since we work over the commutative ring . Since is a graded principal ideal domain and is free as an -module, any Hopf subalgebra of is automatically free as an -module, and hence the dual quotient Hopf algebra of is a free -module. However, if we start with a quotient Hopf algebra of , then, in general, the best we can conclude is that the dual has an -free cokernel. For this reason, we restrict ourselves to those quotient Hopf algebras which are free as -modules. Once again, our work is inspired by that of Ricka, although his result (in the context of the -equivariant Steenrod algebra), is more complicated.
Definition 3.5**.**
A profile function is free if for all we have
[TABLE]
Remark 3.6**.**
- (1)
One can easily check that if is a free profile function satisfying Condition 3.1, then Condition 3.2 is automatically satisfied. 2. (2)
For a simple example of a profile function that is not free, but still satisfies Conditions 3.1 and 3.2, one can take .
Although we cannot give a complete classification of quotient Hopf algebras of , the free profile functions do give a complete classification of those that are free as -modules.
Theorem 3.7**.**
A quotient Hopf algebra of is free as an -module if and only if it is of the form for a free profile function satisfying Conditions 3.1 and 3.2.
Proof.
Let be a quotient map, and its kernel. is free as an -module if and only if agrees with the kernel of , or equivalently, if agrees with the intersection of and the kernel of .
Since is periodic, we can recover this quotient map from its weight [math] part. Under the isomorphism of 2.3, this weight [math] part gives a Hopf algebra quotient map out of the classical dual Steenrod algebra.
Adams and Margolis [AM74] showed that any Hopf algebra quotient of the classical dual Steenrod algebra is of the form . Here ranges over integers , takes values in nonnegative integers or , and for each we have or .
From the description of the isomorphism we see that the kernel of is generated as an ideal in by and .
This yields that the intersection of this kernel with is generated as an ideal in by and , and so is of the form for and . Condition 3.1 and Condition 3.2 for of this form are equivalent to the classical condition for , and is obviously a free profile function. Vice-versa, any free profile function is determined by an in this way. ∎
Proposition 3.8**.**
Let be a free profile function satisfying Conditions 3.1 and 3.2. Let range over sequences of ones and zeroes, almost all zero, with , and let range over sequences of nonnegative integers, almost all zero, with . Then, as an -module, is free with basis
[TABLE]
Proof.
We can again consider as the image of in with . Then the statement follows from the corresponding classical statement, which is that has basis the monomials
[TABLE]
for . ∎
We now introduce a useful family of Hopf subalgebras of .
Example 3.9**.**
Let be the quotient Hopf algebra of defined by the minimal profile function and . Since for and both are 0 otherwise, the profile function is free, and is a finitely generated free -module. Of course, one easily sees that
[TABLE]
It is dual to the subalgebra of generated by for .
4. The Hopf algebra
Let denote the quotient of the motivic Hopf algebra by ; note that this is a Hopf algebra over the field . In the introduction we claimed that this is a 2-primary version of the classical odd-primary Steenrod algebra. Indeed, one can check that in the Adem relations are the same as those in the classical odd-primary Steenrod algebra, under the identification . However, we do not need this result, rather we work with the dual . We start with a general result about -modules.
Lemma 4.1**.**
If is an -module, which is free as an -module, then there is an equivalence of -modules between and .
Proof.
First note that
[TABLE]
There is a short exact sequence
[TABLE]
and applying the exact functor (since is -free) we get a short exact sequence
[TABLE]
It follows that, as -modules, , and it is easy to see that this is an isomorphism of -modules. ∎
Note that since does not appear in the formula for the comultiplication of , the quotient has the structure of a Hopf algebra with comultiplication given by
[TABLE]
We then have the following.
Corollary 4.2**.**
As a Hopf algebra, the dual is isomorphic to
[TABLE]
Proof.
We apply the previous lemma with , which satisfies the conditions of the lemma by 2.1(ii); we then just need to check that the isomorphism respects the Hopf algebra structures, but this is easy to do. ∎
Recall that is dual to - this holds also in . The degree-doubling isomorphism between and the quotient of by the allows one to prove the following, cf. [MP72, Prop. 2.3].
Corollary 4.3**.**
In we have for .
Let be a profile function for ; note that since the classes are now exterior, is now only required to be a function from the set to the set . Let denote the quotient of by the ideal generated by the relations determined by and , and write for the corresponding Hopf subalgebra of . Similar to Proposition 3.4, we can define quotient Hopf algebras of by imposing conditions on the functions and .
Condition 4.1**.**
For all we have or .
Condition 4.2**.**
For all such that we have or .
Proposition 4.4** (Adams–Margolis [AM74]).**
If the profile function satisfies Conditions 4.1 and 4.2 then is a Hopf subalgebra of , and moreover every Hopf subalgebra is of this form.
Sketch of proof..
This is not quite proved in [AM74], however the proof is analogous. First, it is easy to check, as in Proposition 3.4, that the coproduct passes to the quotient, so that is a quotient Hopf algebra, and hence is a Hopf subalgebra. As in [AM74], to prove that every Hopf subalgebra is of this form, we define to be the polynomial subalgebra of generated by , and to be the subalgebra generated by and . For we interpret as and for we interpret as . One then proves the obvious analog of the current proposition for and , via an induction on , using the exact arguments given by Adams and Margolis. The only minor thing to keep in mind is that when Adams and Margolis talk about degree, this refers only to the topological degree - the motivic weight does not play a role. ∎
Since and for in , then given an -module , we can define Margolis homology groups with respect to these elements:
[TABLE]
For the following, we can then essentially quote the proof of the main theorem of [MW81]. Note that here -connected refers to the topological degree.
Theorem 4.5** (Miller–Wilkerson [MW81]).**
Let be a Hopf subalgebra of , and let be an -connected -module.
- (i)
If for all with and for all , then is -free. 2. (ii)
If for all such that and and if for all such that , then whenever with depending only on and . Moreover, if is of finite type, then for .
Sketch of proof..
As noted, we can simply quote Miller–Wilkerson, but for the benefit of the reader we outline the argument. Following their lead, we only prove (ii) and leave (i) for the reader. The first step is to reduce to being finite-dimensional using [MW81, Prop. 3.2]. The proof is then by induction on the -dimension of , using the observation that any finite-dimensional Hopf subalgebra of can be built out of extensions of the form
[TABLE]
where refers to the Hopf algebra .
Now given the dual of in the form , we let
[TABLE]
In the case where , there is an extension of the form
[TABLE]
If , then we have an extension of the form
[TABLE]
and we naturally need to consider the cases where and separately. All three cases are handled in the same way as in [MW81]; perhaps the only thing to check is that Prop. 4.1 of loc. cit. still holds, but it is easy to see that the same proof works in this case. ∎
5. Margolis homology, projective modules, and vanishing lines
In this section, we complete the proof of the main theorem and return to the motivating examples in the introduction. We begin with a definition, followed by a key lemma, essentially appearing in [GIR16].
Definition 5.1**.**
We say that a bigraded module is -connected provided when , and it is connective if it is -connected for some . is of finite type if it is connective and each is a finitely generated -module.
Lemma 5.2**.**
Let be a connected Hopf algebra over which is free and of finite type over , and let be an -module, also free and of finite type over . Then the following conditions are equivalent:
- (i)
* is projective as an -module.* 2. (ii)
* is projective as an -module.* 3. (iii)
* is free as an -module.* 4. (iv)
* is free as an -module.*
Proof.
(i) (ii): For an -module, and an -module, we have a change-of-rings isomorphism
[TABLE]
Since is free over , this derives to an isomorphism
[TABLE]
Now if is projective over , this proves that vanishes, so is projective over .
(ii) (iii): For a connected Hopf algebra over a field, projective implies free for connective modules. To see this, for a projective module form the minimal free resolution . This has the property that all the differentials have coefficients in the augmentation ideal of . So , and projective implies . But then is an isomorphism and is free.
(iii) (iv): Choose a basis of over . Lifting that to , we get a map from a corresponding free -module . By assumption, that map induces an isomorphism . Since is -free, , and analogously for . By induction via the five-lemma, we see that the induced map is an isomorphism for any . Since is finite type over , this implies that is an isomorphism.
(iv) (i) is clear. ∎
Finally, we note the following relation between Margolis homology groups in and in .
Lemma 5.3**.**
Let be an -module, which is free and of finite type over , and let be such that acts trivially on . Then if and only if .
Proof.
The long exact sequence in Margolis homology associated to the short exact sequence shows that if , then . Conversely, if , then is an isomorphism. By assumption on , vanishes in low motivic weight (depending on the degree), and since is an isomorphism, it must vanish in all weights, and hence .\myfnsymbolmult\myfnsymbolmult\myfnsymbolmultThis argument is similar to that given in [GIR16, Lem. 3.4]. ∎
We now give the proof of our main theorem.
Theorem 5.4**.**
Let be a free profile function, and the corresponding Hopf subalgebra of . Let be a -module that is free and of finite type over .
- (i)
If for all with and for all , then is -free. 2. (ii)
If for all such that and and if for all such that , then for all such that with depending only on and the connectivity of .
Proof.
- (i)
By 5.2 it suffices to show that is free as a -module. It is easy to see that if and only if , and similar for . The result then follows from 4.5. 2. (ii)
By 4.5 has a vanishing line of the claimed form. Associated to the filtration of by powers of there is a trigraded-graded algebraic Miller–Novikov type Bockstein spectral sequence
[TABLE]
This is convergent by the assumptions on (see [Mil81, Sec. 8]). Since has internal degree 0, the vanishing line of extends to the -page of the spectral sequence. Differentials in a spectral sequence cannot increase the vanishing line, and the result follows. ∎
Remark 5.5**.**
One can ask for generalizations of this result to other base fields . Let be a prime, and the characteristic of , so that we work with the algebra of bistable operations in the motivic cohomology of smooth -schemes with -coefficients. We start with the case . In this case, it is known from Voevodsky’s solution of the Milnor conjecture [Voe03a] that , where is the mod 2 Milnor -theory of . If is algebraically closed, then is known to be divisible for [Wei13, III. Ex. 7.2(b)] and , and so concentrated in degree 0. It follows that the motivic cohomology of a point . Moreover, the Steenrod algebra has the same form as presented in Section 2 (in positive characteristic see [HKØ, Thm 5.6]) and hence our main theorem applies also to these fields.
On the other hand, we do not yet see a way to extend the results to more general fields, even to . In this case is no longer a Hopf algebra, but rather a Hopf algebroid, and the dual Steenrod algebra becomes more complicated; there exist two distinguished elements and in the motivic cohomology of a point, and the dual Steenrod algebra takes the form
[TABLE]
Additionally the element is no longer invariant, since its right unit is given by . One can quotient by the ideal , but in this case we do not know how to prove an analog of 5.2. Moreover, has bidegree ( in homological grading conventions) and so a Bockstein spectral sequence from to can affect the vanishing line, and one would need to understand the differentials in the Bockstein spectral sequence in more detail.
If is odd, then we can make a similar conclusion as to above. In general, one has that the motivic Steenrod algebra is isomorphic to the odd-primary Steenrod algebra base-changed to the ring , see [Voe03b, Thm. 12.6] for characteristic 0, and [HKØ, Thm 5.6] for fields of positive characteristic. If is algebraically closed, then the motivic cohomology of a point is , and the motivic Steenrod algebra is simply the classic odd-primary Steenrod algebra base changed to . In this case it is easy to prove a version of our main theorem. When is not algebraically closed the situation may still be tractable at odd primes; for example, in [Sta16, Prop. 1.1(2)] Stahn shows that the motivic cohomology of is isomorphic to , with , and one can also prove a version of our main theorem in this case.
We now return to the examples given in the introduction.
Example 5.6**.**
[GIR16*, Prop. 4.7]** *Let be the Hopf subalgebra generated by and , with dual
[TABLE]
We have and in , and so we see that a finite type -module is free only if:
- (i)
is a free -module; 2. (ii)
; 3. (iii)
; 4. (iv)
,
which is in fact a slight strengthening of the result in loc. cit. (where is assumed to be finitely generated over ). The converse also follows by an easy calculation.
Our results also cover [GI15, Prop. 3.2], namely that a bounded below -module of finite type is free as an -module if and only if it is free as an -module and ; here we need to use 5.3 to lift from to = 0.
Example 5.7**.**
[GI15*]** *Although it is not proved explicitly, the method of Guillou and Isaksen can easily be used to show that if is a bounded-below -module of finite type that is free as a -module, then has a vanishing line of slope 1.
To recover this from our work, observe that since , 5.4 applies with , so that whenever (i.e. has a vanishing line of slope 1).
Guillou and Isaksen prove in [GI15, Prop. 7.2] that for such an the map
[TABLE]
is an isomorphism if and a surjection if .
Remark 5.8**.**
Note that we use a different grading convention to [GI15] - a class we write in corresponds to a class in under their conventions.
To recover this from our results (at least the slope - we are not precise about the exact value of the constant), note that corresponding to there is an exact sequence of modules
[TABLE]
It is easy to see that has vanishing Margolis homologies with respect to and , so that whenever (i.e. has a vanishing line of slope 1/2). The long exact sequence
[TABLE]
implies that (5.1) is an injection when and a surjection when . Hence, we see that -multiplication is an isomorphism above a line of slope 1/2 as claimed.
The reference list from the paper itself. Each links out to its DOI / PubMed record.
- 1[AD 73] Donald W. Anderson and Donald M. Davis, A vanishing theorem in homological algebra , Comment. Math. Helv. 48 (1973), 318–327. MR 0334207
- 2[Ada 66] J. F. Adams, A periodicity theorem in homological algebra , Proc. Cambridge Philos. Soc. 62 (1966), 365–377. MR 0194486
- 3[AM 71] J. F. Adams and H. R. Margolis, Modules over the Steenrod algebra , Topology 10 (1971), 271–282. MR 0294450
- 4[AM 74] by same author, Sub-Hopf-algebras of the Steenrod algebra , Proc. Cambridge Philos. Soc. 76 (1974), 45–52. MR 0341487
- 5[DHS 88] Ethan S. Devinatz, Michael J. Hopkins, and Jeffrey H. Smith, Nilpotence and stable homotopy theory. I , Ann. of Math. (2) 128 (1988), no. 2, 207–241. MR 960945
- 6[GI 15] Bertrand J. Guillou and Daniel C. Isaksen, The motivic Adams vanishing line of slope 1 2 1 2 \frac{1}{2} , New York J. Math. 21 (2015), 533–545. MR 3386536
- 7[GIR 16] Bogdan Gheorghe, Daniel C Isaksen, and Nicolas Ricka, The Picard group of motivic A ( 1 ) 𝐴 1 A(1) , ar Xiv preprint ar Xiv:1606.05191 (2016).
- 8[HKØ] Marc Hoyois, Shane Kelly, and Paul Arne Østvær, The motivic Steenrod algebra in positive characteristic , To appear in Journal of the European Mathematical Society .
