In this post we complete the proof of the Quillen–Suslin theorem using Vaserstein’s simplified proof. In Quillen–Suslin 1: Serre’s reduction, we showed that every finite projective module over a polynomial ring over a PID is stably free. Our remaining task is to show that every stably free module over the same ring is free.
The focus is the following definition.
Unimodular vectors
Definition 1.¶ Let be a ring.
- A unimodular vector is an ordered list of elements of which generate the unit ideal.
- A unimodular vector has the unimodular extension property (UEP) if there is a matrix in whose first column is .
- The ring has the unimodular extension property if every unimodular vector over has the UEP.
We regard vectors as columns from now on. This is a fundamentally geometric notion: unimodularity says precisely that the principal open sets cover . Equivalently, the section of the trivial rank- bundle generates a trivial line subbundle, and the UEP asks whether it extends to a frame of the whole bundle. The following proposition explains why the definition is useful, in light of the previous post.
Proposition 2.¶ A ring has the UEP if and only if every finite stably free -module is free.
Proof. Suppose first that has the UEP. By induction on the number of free summands, it suffices to show that if is finite free, then is free.
Choose an identification , and let be the image of . Projection to the second summand gives a linear map such that , so the coordinates of generate . We may therefore complete to a basis of .
For , replace by
These replacements preserve the basis and put in . If is written in the basis , applying shows that its -coefficient is zero. Hence is a basis of .
Conversely, let be unimodular. Choose such that and define
Since , we have . The second summand is stably free, hence free by assumption. A basis of it, together with , completes the required matrix. □
The first example of a ring with the UEP is a PID. There is a brute-force induction proof, but the following argument is a nice trick.
Proposition 3.¶ Every PID has the UEP.
Proof. Let be unimodular, choose with , and use the map from the preceding proof. Then
Since is a PID, the submodule is free. It has rank , as one sees after tensoring with . Thus extends to a basis. □
Definition 4.¶ For , we say that and are -equivalent, and write , if there is such that .
This is an equivalence relation. A unimodular vector has the UEP exactly when it is equivalent to the first standard basis vector . Permuting components, multiplying one component by a unit, and adding a multiple of one component to another all preserve equivalence.
Two small cases will be useful. A unimodular vector of length one is a unit. For a unimodular vector , choose ; then
has determinant and completes the vector. Thus vectors of length at most two have the UEP over every commutative ring.
Horrocks’ theorem
We next need Horrocks’ theorem. This is the machinery behind the inductive proof of the Quillen–Suslin theorem. We give the polynomial argument, including the degree reduction which makes the induction work.
Lemma 5.¶ Let be a ring, let be monic of degree , and let
If the coefficients generate , then the ideal contains a monic polynomial of degree .
Proof. Division by the monic polynomial gives a unique remainder of degree less than . For , let be the remainder of modulo , and let be its coefficient of . Each belongs to .
The term contributes to , terms with index less than have degree less than , and terms with index greater than contribute -multiples of their coefficients. Consequently,
Descending induction on shows that . Choose with . Then has degree and leading coefficient , as required. □
Theorem 6 (Horrocks).¶ Let be a local ring. If is a unimodular vector with a monic component, then has the UEP.
Proof. The cases were handled above, so assume . After permuting components, let be monic of degree . We induct on .
If , then and subtracting times the first component from the th component gives . Suppose . Divide each , for , by and subtract the quotient times . These operations let us assume that
At least one coefficient of one of these remaining components is a unit in . Otherwise all the for would vanish in , and unimodularity would say that the monic degree- polynomial generates the unit ideal of , which is impossible. Relabel so that has a unit coefficient.
By Lemma 5, there are such that
is monic of degree . Let be the coefficient of in . Replace by
This is an elementary change, and its coefficient of is .
We have obtained an equivalent unimodular vector with a monic component of degree . The induction hypothesis completes the proof. □
Corollary 7.¶ If is a local ring and is unimodular with a monic component, then over .
Proof. Evaluating a relation at zero shows that is unimodular over . At least one is a unit, since otherwise all of them would lie in . Hence over , and Horrocks’ theorem gives over . □
From local rings to a domain
The local theorem tells us when a polynomial vector is equivalent to its value at zero. The next step is to patch this equivalence over a domain. The useful intermediate statement concerns translations of the variable.
Lemma 8.¶ Let be a domain and a multiplicative subset. If over , then there is such that
over .
Proof. Choose with , and set
Then
Thus for a matrix over . Choose a common denominator for its finitely many coefficients. We then have
It remains to check invertibility over , rather than just over its localization. A unit in a polynomial ring over a domain is constant, so is a unit of . Therefore
The adjugate supplies an inverse over . The displayed identity now proves the desired equivalence. □
Proposition 9.¶ Let be a domain and a unimodular vector with a monic component. Then over .
Proof. Let
We first check that is an ideal. Clearly . If and , substituting for in an equivalence gives . For addition, suppose
Substitute for in the first identity and compose with the second. This gives
Both factors are invertible over , so .
Now let be a prime ideal. The ring is local, so Corollary 7 gives over . Applying Lemma 8 with produces . Thus is contained in no prime ideal, and .
In particular, there is with
Evaluating at zero gives . The specialized matrix is invertible because its inverse specializes as well. Renaming as proves the proposition. □
Making a component monic over a PID
We can now discard a variable whenever a component is monic in it. Over a field, a suitable change of variables makes any nonzero polynomial have constant nonzero leading coefficient, which we may then scale to . Over a PID, that coefficient need not be a unit. We therefore need a stronger version of the monic trick.
We will use unique factorization to prove exactly the version needed here. Recall that polynomial rings over a PID are noetherian UFDs. Saying that an ideal has no common irreducible divisor means that it is not contained in for any irreducible element .
Lemma 10.¶ Let be a noetherian UFD and a nonzero ideal with no common irreducible divisor. The set
is an ideal of with no common irreducible divisor.
Proof. To add two leading coefficients, multiply their polynomials by suitable powers of to give them the same degree and then add. If the coefficients cancel, their sum is zero, which is included in . Multiplication by elements of is equally immediate. Thus is an ideal.
Suppose, for a contradiction, that for an irreducible . Choose the finitely many minimal primes over . Noetherianity gives an integer such that
Indeed, the product is contained in , and a finitely generated ideal contained in has some power contained in . Leading coefficients multiply in the domain , so
Since is prime, for at least one of the , which we now call . Put . Then . In a UFD the only primes contained in are and : a nonzero prime contained in contains an irreducible factor of any of its nonzero elements, and that factor must be associated to .
If , every coefficient of every polynomial in lies in . To see this, its leading coefficient lies in by assumption; subtract that leading term, which belongs to , and repeat. Hence , contradicting the hypothesis on .
If , localize at . The nonzero prime becomes a nonzero proper prime of the PID , and is therefore generated there by an irreducible polynomial. Clear denominators and divide out its content to obtain a primitive irreducible polynomial . Gauss’ lemma says
A prime disjoint from the multiplicative set equals the contraction of its localization, so . Again has a common irreducible divisor, a contradiction. □
Lemma 11 (The monic trick over a PID).¶ Let be a PID and a nonzero ideal with no common irreducible divisor. If , there are polynomial coordinates over in which contains a polynomial monic in .
Proof. For , the ideal from Lemma 10 has no common irreducible divisor. Since is a PID and , it follows that . Thus contains a monic polynomial.
Suppose and put . Regard as an ideal in . By the preceding lemma, has no common irreducible divisor in . By induction we may choose coordinates for such that contains a polynomial
By the definition of , choose
Choose a positive integer so large that
for every nonzero with . Set , so . These are still polynomial coordinates over . Written in these coordinates, the leading -term of is with coefficient , and every term coming from has smaller -degree. Therefore is monic in . □
This is the UFD case of Suslin’s monic-polynomial argument; the more general formulation uses the height of an ideal. See Suslin, §6. The point of the preceding proof is that it applies to a PID without dividing by a nonunit coefficient.
Lemma 12.¶ Let be a PID and . Given a unimodular vector , with , elementary operations and a change of polynomial coordinates produce a vector with a monic component.
Proof. Permute the components so that . If is a unit, adding a suitable multiple of to makes , and we are done. Otherwise let be its distinct irreducible factors in the UFD .
For each such that divides , unimodularity supplies an index for which does not divide . Replace by
where the condition in the sum refers to the original . This is a sequence of elementary operations. Modulo any which originally divided , exactly its designated summand is nonzero. Modulo any other , all the added summands vanish and was already nonzero. Thus the new is divisible by none of the , and have no common irreducible divisor.
Apply Lemma 11 to the ideal . After changing coordinates, there is a polynomial
which is monic in . Choose with . Replacing by
makes the third component monic in . Again, these are elementary operations. □
Vaserstein’s theorem
We can finally complete the induction. The proof is an induction on the number of variables, not on the length of the vector.
Theorem 13 (Vaserstein).¶ Let be a PID and let be unimodular. Then has the UEP.
Proof. For , this is Proposition 3. Suppose the result holds for variables. Vectors of length one or two can be completed over any ring, so assume .
By Lemma 12, after elementary operations and a change of coordinates we may assume that a component of is monic in , where
Apply Proposition 9 with coefficient ring . It gives
over . The vector on the right is unimodular over , so the induction hypothesis completes it to an invertible matrix over that ring. The same matrix remains invertible after adjoining .
Hence our transformed vector is equivalent to . Undo the elementary operations to obtain a completion of the original vector. A change of coordinates does not affect the conclusion: an isomorphism of polynomial rings sends an invertible matrix, together with its inverse and its specified first column, to an invertible matrix with the corresponding first column. □
Corollary 14 (Quillen–Suslin).¶ Let be a PID. Then every finite projective module over is free. Equivalently, every finite-rank vector bundle on is trivial.
Proof. Vaserstein’s theorem says that has the UEP, so Proposition 2 shows that every finite stably free module over it is free. By Theorem 14, every finite projective module over the same ring is stably free. Combining the two statements proves the theorem. □
The two parts of the argument do rather different jobs. Serre’s reduction produces a stable basis; the unimodular-vector argument removes the additional free summands. Horrocks’ theorem and the translation ideal are what let us carry this cancellation from a local ring back to the polynomial ring itself.
For the paper version, including the geometric background and further applications, see Serre Schemes and the Quillen–Suslin Theorem.