Commutative algebra
Commutative Algebra
Fall 2020
Notes by Patrick Lei
Lectures by Eric Urban
Columbia University
Disclaimer
These notes were taken during lecture using the vimtex package of the editor neovim. Any errors are mine and not the instructor’s. In addition, my notes are picture-free (but will include commutative diagrams) and are a mix of my mathematical style and that of the instructor. If you find any errors, please contact me at plei@math.columbia.edu.
Any similarties to Matsumura’s Commutative Algebra are due to the instructor.
Acknowledgements
I would like to thank Caleb Ji, Nicolás Vilches, and Alex Xu for pointing out typos in these notes.
1 Basic Notions
The references we will use in this course are Matsumura’s Commutative Algebra and Serre’s Algèbre Locale, Multiplicités. There is an English translation of Serre. We will begin with general results on rings and modules. We will assume all rings are commutative and unital. Recall that and ideal of a ring is prime if and only if is a domain, and is maximal if and only if is a field.
1.1 Basics of Ideals
Definition 1.1.1.
Let be an ideal. Then the radical of is the set
Definition 1.1.2.
An ideal is primary if and the zero divisors in are nilpotent. Thus if and , then for some .
Proposition 1.1.3.
If is primary, then is a prime ideal.
Proof.
If , then . If , then because . □
Remark 1.1.4.
The converse to Proposition 1.1.3 is false in general.
Definition 1.1.5.
Let be a ring. Then the spectrum of as a set is the set of prime ideals of . We may place the Zariski topology on this set, where the basis of open sets is given by , where is the set of prime ideals containing .
If is a morphism of rings, the morphism is continuous in the Zariski topology.
Exercise 1.1.6.
In particular, if , then is an embedding.
Exercise 1.1.7.
Let be an ideal. Then let be ideals of that are all prime except possibly two of them. Show that if for all , then .
Exercise 1.1.8.
Let be ideals of such that . Then
- 1.
- 2.
There is an isomorphism of rings .
1.2 Localization
Let be a multiplicative subset. The main examples are and for a prime ideal . Then if , there is at least one ideal such that . Denote the set of such by . Then any maximal element of is a prime ideal in . Existence of a maximal element is seen using Zorn’s lemma.
To see that maximal elements of are prime ideals, note that is not in , so if , there exist and such that and . Therefore , so is not in
Lemma 1.2.1.
Let be the set of all nilpotent elements. Then
Proof.
One direction is easy, so let be contained in all prime ideals. Then consider the set . If , then is nonempty, so it has a maximal element. This is a prime ideal, which implies is not contained in some prime. □
Corollary 1.2.2.
Let be an ideal of . Then is the intersection of all prime ideals containing .
Now fix a multiplicative subset . Then we will define an equivalence relation on . We write
if there exists such that . If is a domain, then this says that . Now we will define the localization to be the set of equivalence classes for this relation. Note there is a natural morphism that sends .
Note that the localization has a universal property: If is a morphism such that , then factors uniquely through .
Localization gives a map , and in particular, if , we recover the set .
1.3 Modules
Let be a ring. Then an -module is an abelian group with an action of . If is an -module and is a multiplicative set, then is the set of equivalence classes for if there exists such that . This is an -module.
Lemma 1.3.1.
Let be an -module. Then the map
is injective.
Proof.
Let be nonzero. Then the annihilator of is a proper ideal of , so it is contained in a maximal ideal. This implies that is nonzero. □
Corollary 1.3.2.
Let be a domain. Then , where this intersection makes sense inside the fraction field of .
Proof.
Apply the previous lemma to . □
Definition 1.3.3.
Let be an -module and . Then is -regular if the morphism is injective. Additionally, if is -regular, then it is called regular.
The set of all regular elements in is multiplicative, and the ring is called the total ring of fractions. If is a domain, then , and is the field of fractions.
Definition 1.3.4.
A ring is a local ring if has only one maximal ideal. In this case, all elements not in the maximal ideal are units.
Remark 1.3.5.
If is an ideal such that , then is a local ring and its maximal ideal.
Example 1.3.6.
Now let be a general ring and . Then is a local ring with maximal ideal .
Definition 1.3.7.
Now suppose are local rings. Then a morphism of rings is local if . This means we have a commutative diagram
| (1.1) |
where is the residue field of .
Recall that the nilradical is the set of all nilpotent elements, or equivalently the intersection of all prime ideals. Then the Jacobson radical is defined to be the intersection of all maximal ideals.
Proposition 1.3.8.
Let . Then if and only if is a unit for any .
Proof.
If , then is contained in some maximal ideal , which implies . In the other direction, suppose there exists some maximal ideal such that . Then is nonzero in . Thus there exists such that , which contradicts the assumption that is a unit for any . □
Lemma 1.3.9 (Nakayama’s Lemma).
Let be a finitely generated -module. Then let be an ideal such that . Then there exists such that . In particular, if , then .
Proof.
We will induct on the number of generators. If , then for some , and thus . Now suppose . Let . By the inductive hypothesis, for some . Therefore , so . Therefore for some , and thus . Thus . □
Corollary 1.3.10.
Let and such that . Then if either
- 1.
is nilpotent;
- 2.
and is finitely generated,
then .
Proof.
- 1.
Suppose is nilpotent. Then
because is nilpotent.
- 2.
Let and be finitely generated. Then set , where is the image of inside . Because is finitely generated, so is . Therefore , so .
□
Remark 1.3.11.
Most of the time, we apply this result when is local and is the maximal ideal of . In this case, is a finite-dimensional vector space over .
1.4 Artinian and Noetherian Rings
Definition 1.4.1.
We say that an -module satisfies the ascending chain condition if any ascending chain of submodules of becomes stationary. Similarly, satisfies the descending chain condition if any descending chain of submodules becomes stationary. If satisfies the ascending chain condition, it is called Noetherian, and if satisfies the descending chain condition, it is Artinian.
Proposition 1.4.2.
Assume we have a short exact sequence of modules
| (1.2) |
Then is Noetherian (resp. Artinian) if and only if and are.
Proof.
Proving that if is Noetherian, then and are is left to the reader. Now consider a chain
Then let be the image of the in and . Then we have an exact sequence
Because and stabilize, so must from the exact sequence. □
Corollary 1.4.3.
If is Noetherian (resp. Artinian), then any finitely generated -module is Noetherian (resp. Artinian).
Corollary 1.4.4.
Assume is Noetherian. Then any finitely generated -module has a projective resolution by finite free -modules. In other worse, there exists an exact sequence
such that each .
Proof.
Suppose is finitely generated. Then , so we have a sequence
Then and . Then we repeat this process with taking the role of . □
Proposition 1.4.5.
An -module is noetherian if and only if any submodule of is finitely generated.
Proof.
Let . Then choose . Then if , choose . This process will stop because is Noetherian, so is finitely generated.
Now suppose any submodule is finitely generated. Given a chain
set . This is finitely generated and is also equal to the first that contains all of the generators. □
This means that a ring is noetherian if and only if all ideals of are finitely generated. In particular, fields and principal ideal domains are Noetherian.
Proposition 1.4.6.
Let be Noetherian and suppose is a multiplicative subset of . Then is Noetherian.
Proof.
Consider the morphism . Then let be a chain of -modules in . Their preimages form a chain, and they are stationary, so is also stationary. □
Theorem 1.4.7.
Let be a Noetherian ring. Then is Noetherian.
Proof.
Let . Then be generated by the dominant coefficients of polynomials in of degree at most . Then we can write as where and a dominant coefficient of a polynomial of degree at most in . Thus the form a chain of ideals of that stabilizer for . Then . Set . If , then there exists such that such that and .
Therefore , so and is thus finitely generated. □
Corollary 1.4.8.
Let be a finitely-generated -algebra. Then if is Noetherian, is also Noetherian.
Corollary 1.4.9.
Any finitely generated algebra over a field is Noetherian.
Remark 1.4.10.
Suppose is Noetherian and an -module. If is finitely generated, then is Noetherian, but submodules are not necessarily Noetherian. However, they are finitely generated.
Suppose is an inclusion of rings. Then we say that is integral over if there exists a monic polynomial such that .
Proposition 1.4.11.
The following are equivalent:
- 1.
is integral over ;
- 2.
is a finitely-generated -module;
- 3.
There exists such that is a finitely-generated -module.
- 4.
There exists a faithful -module which is finitely generated over .
Proof.
- 1 implies 2
-
Note that is generated by , where has degree .
- 2 implies 3
-
Set .
- 3 implies 4
-
Choose .
- 4 implies 1
-
Write . is an -module, so we can consider . Then for all , we have , so if we write consider the matrix , then this matrix represents the map given by multiplication by . Therefore we have
so set . This is monic and for all , and therefore acts by on . Because is a faithful -module, we have .
□
Exercise 1.4.12.
Let . Show that if and are integral over then so are .
Proposition 1.4.13.
Let . Assume that is Noetherian and that is a finitely-generated -algebra. If is a finitely-generated -module, then is a finitely-generated -algebra.
Proof.
Write . Also, we can write for some . Then we can write and for . Then is a finitely-generated -algebra. Any element of is a polynomial in the with coefficients in , so is a finitely-generated -module. In particular, is Noetherian. Because , this implies that is a finitely generated -module, so it is a finitely-generated -algebra. □
Corollary 1.4.14.
Let be a field and a finitely-generated -algebra. If is a field, then is a finite extension of .
Proof.
Let be a finitely-generated -algebra. Then there exist that are algebraically independent over . Then is algebraic over , which is the field of fractions of . However, this gives an inclusion , where is a finitely-generated -algebra and is algebraic over .
By the proposition, is a finitely-generated -algebra. Therefore, we can write , where . Because is a UFD, then we can write
is repatively prime to all of the , so . This gives a contradiction, so must be algebraic over . □
1.4.1 Primary Decomposition in Noetherian Rings
Definition 1.4.15.
An ideal is irreducible if for any decomposition , then either or .
Example 1.4.16.
If is a prime ideal, then is irreducible. To see this, if , then contains one of , and so either or .
Remark 1.4.17.
Suppose is a maximal ideal. Then any power of is primary.
Proof.
We want to prove that the zero divisors of are nilpotent. Because is maximal, then is a local ring with maximal ideal . But then are all units, so everything in is nilpotent. □
Lemma 1.4.18.
If is Noetherian, then every irreducible ideal is primary.
Proof.
Let be irreducible. Then we can pass to the quotient, so we may asusme . Let be nonzero with . We want to show that is nilpotent.
Because is Noetherian, then there exists such that . We want to show that , so choose . Then , so . However, this means . Because is irreducible, then , so . □
Corollary 1.4.19.
If is Noetherian, then every ideal of has a primary decomposition. In other words, we can write , where each is primary.
Proof.
Let be the set of ideals with no primary decomposition. If is nonempty, then has a maximal element . To see this, we can use the fact that is Noetherian, so any chain of ideals in eventually stabilizes. We know that is not irreducible, we can write such that . In addition, , so they have a primary decomposition. This implies that has a primary decomposition. □
Remark 1.4.20.
This decomposition is not unique. For example, consider . Then for all .
1.4.2 Artinian Rings
Proposition 1.4.21.
Assume that is Artinian.
- 1.
Every prime ideal of is maximal.
- 2.
has finitely many maximal ideals.
- 3.
The Jacobson radical of is nilpotent.
Proof.
- 1.
Fix a prime ideal and consider the domain . Choose and consider the decreasing chain of ideals. This stabilizes, so there exists , so we can write for some , and therefore because is a domain. Therefore has an inverse, so is a field. Thus is maximal.
- 2.
Suppose we have infinitely many maximal ideals that are pairwise distinct. Then we form a chain
which becomes stationary. Therefore , so contains some . Because these ideals are maximal, this is a contradiction.
- 3.
Consider . Then the chain stabilizes, so for some . Let .1 We will show that . If not, let such that is minimal for this property. Such a exists because is Artinian.
Let and consider the ideal By minimality of , we see that (otherwise by Nakayama’s lemma). Therefore , so and thus . Therefore, . This implies , so and thus . □
Definition 1.4.22.
An -module is called irreducible if and are the only submodules of .
Definition 1.4.23.
An -module is said to be of finite length if there exists a (finite) decreasing sequence of submodules
such that is irreducible for . In this case, is actually unique and depends only on . We will call the length of .
Proposition 1.4.24.
Let be a ring. Then is Artinian if and only if is of finite length as an -module.
Proof.
If is of finite length, then we have a sequence where is irreducible. If is a decreasing chain of ideals, so is a decreasing chain of ideals. However, each is either or , so this chain stabilizes. Similarly, the chain also stabilizes for all . Therefore, there exists such that for all , is constant for all , so is constant for all .
Now suppose that is Artinian. Choose , where the are the maximal ideals of . Then is nilpotent, so there exists such that
Then by the Chinese remainder theorem, so is clearly a local ring and is of finite length as an -module. Note that the -vector space is finite-dimensional because is Artinian. Therefore is of finite length. □
Exercise 1.4.25.
If there is an exact sequence of -modules, then is of finite length if and only if and are of finite length. Moreover, .
Theorem 1.4.26.
is Artinian if and only if is Noetherian and .
Proof.
If is Artinian, we have already proved that . By the previous proposition, because is of finite length, is Noetherian. To see this, for a chain , note that stabilizes. We can do this for each , so any increasing chain stabilizes.
Now assume is Noetherian and has dimension . We know that has a primary decomposition, so we can write , where each is primary. Then is a prime ideal, so it is maximal because . Because is Noetherian and for all , for , so there exists such that for each . Therefore
so . Therefore, . Each is of finite length (because each is a finite-dimensional vector space), so is of finite length. □
Proposition 1.4.27.
Let be a Noetherian local ring with maximal ideal . Then one of the following holds:
- (a)
Either for all , or;
- (b)
for and in this case, is Artininian.
Proof.
If , then by Nakayama’s lemma. This implies that is of finite length. Then if is prime, then , so . Because is maximal, , so . □
Theorem 1.4.28 (Structure Theorem for Artinian Rings).
An Artinian ring is uniquely up to isomorphism a finite product of Artinian local rings.
Proof.
Previously, we proved that . Each of these is a local Artinian ring. □
2 Linear Algebra of Modules
Proposition 2.0.1.
Assume are -modules.
- 1.
The sequence is exact if and only if for all -modules , the sequence
is exact.
- 2.
The sequence is exact if and only if for all -modules , the sequence
is exact.
Proof.
This is left as an exercise. □
Remark 2.0.2.
In general, if is exact, then
are exact but the last morphism is not necessarily surjective.
Definition 2.0.3.
A module is projective if the functor is exact. Here, exact means that short exact sequences are preserved. Similarly, a module is injective if the functor is exact.
Proposition 2.0.4.
A module is projective if and only if is a direct factor of a free module. In other words, there exists a free module and -module such that .
Proof.
Suppose is projective. Then there is a surjection . Because is projective, there exists a map such that . Therefore , where is the kernel of .
On the other hand, if , then for any diagram of the form
| (2.1) |
we can embed in and then use projectivity of free modules (because ). □
Remark 2.0.5.
If is projective and finitely generated, then it is a direct factor of a finite free module.
Definition 2.0.6.
A projective resolution of an -module is a right bounded complex
of projective modules such that there exists such that
is exact.
Exercise 2.0.7.
Show that any module has a projective resolution (Hint: construct a free resolution). In addition, any two projective resolutions are homotopic.
Definition 2.0.8.
such that .
2.1 Tor and Ext Functors
Note that for a complex, we can compute the homology . This measures the defect of the complex from being exact. For functors that are not exact, we can construct derived functors that measure the defect of exactness. Let be right exact. Then for any , we can consider a projective resolution . Applying to , then the left derived functor is defined by .
Proposition 2.1.1.
If is exact, then we have a long exact sequence
Recall that the tensor product of two modules is an -module with a bilinear map such that all bilinear maps factor through .
Proposition 2.1.2.
The functors are an adjoint pair.
Corollary 2.1.3.
If is exact, then
is exact.
Definition 2.1.4.
A module is flat if is exact.
We can defined the left derived functors of the tensor product.
Proposition 2.1.5.
Any projective module is flat.
Proof.
Clearly free modules are flat, so write and then note that the tensor product distributes over the sum. □
Proposition 2.1.6.
The following are equivalent:
- 1.
is flat over .
- 2.
If , then .
- 3.
For all finitely generated ideals , .
- 4.
For any finitely generated ideals , .
- 5.
For any finitely generated module , we have .
- 6.
For all and such that there exist and such that .
Proof.
It is clear that 1 is equivalent to 2 implies 3 implies 4 implies 5. The directions 3 implies 2 and 4 implies 3 are left to the reader.
- 1 implies 6
-
Choose such that . Then define a map by
and define . Because is -flat, we have an exact sequence
Then . Therefore there exists and such that
Writing , we obtain the identity
and thus .
- 6 implies 3
-
Choose an ideal . Consider the map . Then for any element in the kernel, we can write
Then we can write and so
and thus is injective. □
Let be a map of rings and let be a -module. Define to be flat if is flat as an -module.
Proposition 2.1.7.
If is flat and is a flat -module, then is also flat as an -module.
Proof.
Let be an -module. Then . If is an exact sequence of -module, by flatness of as an -module, then
is exact. Because is flat over , we see that
is exact, as desired. □
Now let be an -module. Then for any map , we can consider the -module .
Proposition 2.1.8.
If is -flat, then is -flat.
Proof.
For a -module , write
Thus if is exact, then is exact because is -flat, as desired. □
Proposition 2.1.9.
Let be a multiplicative subset of . Then the morphism of rings is flat.
The proof is left to the reader. This can be reformulated as .
Now we will give some remarks about the functors. For any left exact functor, we may define the right derived functors by
where is an injective resolution. Then we will define the right derived functors of by .
Proposition 2.1.10.
If is injective, then for all . Similarly, if is projective, then for all .
Remark 2.1.11.
We can compute using a projective resolution of .
Proposition 2.1.12.
Let be a morphism of rings and let be -modules. Then let be their base changes to . Then we have
and
if is -flat.
Proof.
This follows from the definition of using projective resolutions using the following facts.
- 1.
If is -projective, then is -projective.
- 2.
Since is -flat, for any complex of -modules, then .
□
2.2 Flatness
Proposition 2.2.1.
Let be a local ring. Then any finitely generated flat -module is free. In particular, free, projective, and flat are equivalent for -modules.
Proof.
We know that free implies projective implies flat. Therefore we will show that if is flat, then it is free. Assume that is finitely generated and -flat. Let be the residue field of . Define , which is a vector space of finite dimension over . Then there exists that descend to a basis of .
Then the map is surjective by Nakayama’s lemma. We will prove that this map is injective by induction. If , then suppose . Then there exist such that
where for all . Because , there exists such that so is invertible in . Thus .
Now suppose . Then there exist and such that
and
Because , we see that for some and thus is a unit. Then , so we can write
We know that are linearly independent over , so from the induction and thus . □
Remark 2.2.2.
If is not finitely generated, the proposition is false. An example is given by taking the field of fractions of a local domain.
When proving the proposition, we in fact proved that
Lemma 2.2.3.
If with a flat -module for a local ring and are linearly independent in , then are linearly independent in .
Proposition 2.2.4.
Suppose that is flat and are ideals of . Then
- 1.
;
- 2.
If is finitely generated, then .
Proof.
The proof is a formal consequence of flatness.
- 1.
Consider the exact sequence . Tensoring with , we obtain an exact sequence
But then , but the kernel of the last map is clearly .
- 2.
Set . Then because
it suffices to prove the result for a principal ideal. We have an exact sequence
Tensoring by , we obtain
and by analysing the kernel, we see that . By repeated application of the previous part, the desired result follows. □
Example 2.2.5.
We will give an example where the previous proposition is not true in general. Let and . Then choose , so . But then we have , but .
Another example is . Here we can check that , that , but . Viewing this geometrically as , we can check the fiber over and see that the map is not flat.
Proposition 2.2.6.
Let be a ring homomorphism. The following are equivalent:
- 1.
is flat over ;
- 2.
is flat over for all and .
- 3.
is flat over for any maximal.
Proof.
- 1 implies 2:
-
We know that is flat over . But then is flat over because it is a localization. By transitivity of flatness, is flat over .
- 2 implies 3:
-
This is obvious.
- 3 implies 1:
-
Note that for all maximal, for . This implies that , and thus is flat over . To get that the first is zero we need to use the lemma belos. □
Lemma 2.2.7.
Let be a morphism of rings and choose . Then let and an -module. Then is a -module and .
Proof.
Let be a projective resolution. Then is computed by the homology of the complex . When we localize, we localize the homology at the term. However, , so because is -projective, then is -projective, and thus is a projective resolution of . Thus the complex computes the as desired. □
Definition 2.2.8.
An -module is said to be faithfully flat if
- 1.
is -flat;
- 2.
For any sequence of -modules, if is exact, then is exact.
Theorem 2.2.9.
Let be an -module. Then the following are equivalent:
- 1.
is faithfully flat over ;
- 2.
is flat and for any nonzero , ;
- 3.
is flat and for all maximal ideals , .
Proof.
- 1 implies 2:
-
Choose the sequence . Then tensor with . If , the sequence is now exact, then the original sequence is exact, and thus .
- 2 implies 3:
-
Consider . Then , so .
- 3 implies 2:
-
Choose and set for some maximal ideal . Then , so . Because injects into , injects into , which must be nonzero.
- 2 implies 1:
-
Consider the sequence . Then because is flat, and . If , then , which happens iff . Then .
If is exact, then is a complex. Finally, we need that . By flatness of , we can tensor to find that and then we see that . □
Corollary 2.2.10.
Let be a local homomorphism and let be a finitely-generated =module. Then if flat over if and only if is faithfully flat over .
Proof.
Clearly faithfully flat implies flat. Then we need to show that . By Nakayama’s lemma, we know that , so . In particular, . In particular, this implies that item 3 of the previous theorem holds. Thus is faithfully flat over . □
Remark 2.2.11.
This also shows that flat and faithfully flat are equivalent over local rings. Alternatively, we can use the equivalence of flat and free.
Remark 2.2.12.
Faithful flatness is transitive. In addition, if is a morphism of rings, and is faithfully flat over , then is faithfully flat over .
Proposition 2.2.13.
Let be a faithfully flat -module which is faithfully flat over . Then is faithfully flat over .
Proof.
Let be an -module. Then if . This implies that is nonzero. Now it suffices to show that is flat over .
Let be an exact sequence of -modules. Then if we consider , this is exact by flatness of over . By faithful flatness of over , this implies that is exact. □
Proposition 2.2.14.
Let be faithfully flat. Then
- 1.
For any -module , the map is injective;
- 2.
If is an ideal, then ;
- 3.
The map is surjective.
Proof.
- 1.
Let . Then . Because is faithfully flat, .
- 2.
Recall that . Then the map is injective. Therefore we have a map which has kernel .
- 3.
Choose . Then and is faithfully flat by base change. This means that . Thus if we choose to be a maximal ideal of containing , we see that and thus they are equal by maximality of . Then if we choose , we see that
Thus the image of is . □
Theorem 2.2.15.
Let be a map of rings. The following are equivalent:
- 1.
The map is faithfully flat.
- 2.
The map is flat and is surjective.
- 3.
The map is flat and for all maximal ideals of , there exists some maximal ideal of such that .
Proof.
- 1 implies 2:
-
This is the previous proposition.
- 2 implies 3:
-
Choose a maximal ideal . Then there exists such that . But then if is any maximal ideal containing , we see that by maximality of .
- 3 implies 1:
-
We want to prove that for any maximal ideal of . Then there exists such that . But then . □
Proposition 2.2.16 (Descent).
Let be faithfully flat and be an -module. Then
- 1.
is flat (resp. faithfully flat) if and only if is -flat (resp. -faithfully flat).
- 2.
Assume is a local ring and is finitely-generated. Then is free if and only if is -free.
Proof.
- 1.
Let be an exact sequence. Then is exact, so is exact. By faithful flatness of , is exact. Now if is another -module, we know that , but this is the same as , so is nonzero.
- 2.
Assume that is local. Then suppose is free. Therefore is faithfully flat. But then, is faithfully flat over , which means that is free because is finitely generated. □
Exercise 2.2.17.
Let be integral domains. Assume that and have the same field of fractions. Prove that is faithfully flat if and only if .
2.3 More on Integral Dependence
Recall Proposition 1.4.11.
Corollary 2.3.1.
Let . If each is integral over , then is a finitely-generated -module.
Corollary 2.3.2.
Let be the set of integral elements over . Then is a subring of .
Proof.
Note that , which is a finitely-generated module. Therefore they are integral over . □
Remark 2.3.3.
The ring is not necessarily finitely-generated over . For an example, choose .
Definition 2.3.4.
Let . Then we say that is integral over if all elements of are integral over .
Corollary 2.3.5.
Let be extensions of rings. Then if is integral over and is integral over , then is integral over .
Proof of this is left to the reader.
Definition 2.3.6.
Let be an integral domain. We say that is integrally closed if for all , then is integral over if and only if .
Definition 2.3.7.
Assume that is an inclusion of rings. Then the integral closure of inside is the set of all elements of that are integral over .
Example 2.3.8.
A typical example of this situation is when is a domain, is its fraction field, and is a field extension. Then we can consider the integral closure of inside . In number theory, if is a number field, we define its ring of integers to be the integral closure of in .
Exercise 2.3.9.
If , prove that .
Exercise 2.3.10.
Let be an integral domain and . Prove that the integral closure of inside is integrally closed.
Lemma 2.3.11.
Let be a domain that is integral over . Then is a field if and only if is a field.
Proof.
Assume that is a field. Now choose . But then we know that
for some and . But then we have
and so .
Now assume that is a field. Choose . Then . This means that is integral over , which means that
for some . Then if we multiply by , we obtain
which means . □
Corollary 2.3.12.
Let and be integral over . Let and . Then is maximal if and only if is maximal.
Proof.
Note that is integral over . Then we apply the lemma to . □
We can refine this into going-up and going-down. Let be a morphism of rings and let be the induced map of spaces.
Definition 2.3.13 (Going-up).
A ring homomorphism satisfies the Going-up property if the following holds:
Let be prime ideals of and suppose that . Then there exists such that .
Definition 2.3.14 (Going-down).
A ring homomorphism satisfies the Going-down property if the following holds:
Let be prime ideals of . Then let satisfy . Then there exists with .
Lemma 2.3.15.
The going-down property is equivalent to the following:
For all and a minimal prime ideal of containing , we have .
Proof.
First suppose that going-down holds. Then choose be a minimal prime containing . Then . If , then there exists such that , which contradicts minimality.
Now suppose the other condition holds. Suppose goes to . Then we know that . If we fix to be the minimal prime containing , then we see that . □
Theorem 2.3.16.
If is flat, then going-down holds.
Proof.
Fix and let lie over . Then we know that is flat over . Because is local and the map is local, it is faithfully flat. This implies that the map is surjective, so there exists such that .
Now set . Then we see that
We will see consequences of this result in algebraic geometry.
We will now consider integral ring extensions .
Theorem 2.3.17 (Cohen-Seidenberg).
Suppose is an integral extension. Then the following hold:
- 1.
The map is surjective.
- 2.
There are no inclusion relations between the prime ideals of which are above a fixed prime ideal of .
- 3.
Going-up holds for .
- 4.
If is local with maximal ideal , then the prime ideals of lying over are precisely the maximal ideals of .
- 5.
Assume further that and are integral domains and that is integrally closed. Then going-down holds for .
- 6.
If is the integral closure of in a normal extension of field of , then any two prime ideals of lying over the same prime ideal of are conjugate by an element of .
Proof.
We prove 4, then 1, 2, and 3, then 6, and finally 5.
- 1.
Let . Then is integral over . Applying 4, we obtain the desired result.
- 2.
Consider again. By 4, because ideals lying over are maximal, there cannot be inclusion relations between them.
- 3.
Let and lying over . Then is integral over . By 1, we know that is surjective. Thus there exists lying over . Then we know that for some prime ideal of , and this is the ideal we are looking for.
- 4.
This is a consequence of Lemma 2.3.15.
- 5.
Write . Then let be the normal closure of . Then let in and in lie over . Then let in the integral closure of in . These exist thanks to 1 and 3.
Let in such that . Then there exists such that because both ideals are above . Then we can choose
We need to show that . But this is simply
- 6.
We know that is integrally closed in . Then let be a finite Galois (we can always reduce to this case) extension and the integral closure of in . Then let lie above . We will show there exists such that .
Suppose that no such exists. Then for all , . In particular, . Then there exists which is not in any then we see that
is integral over , so . Also, because , so and thus , so . This gives a contradiction. □
Corollary 2.3.18.
Assume that is integral over .
- 1.
If is a chain of prime ideals of , the the for a chain of prime ideals of .
- 2.
If is a chain of prime ideals of , then there exists a chain of prime ideals of above it.
- 3.
If is integrally closed and is a domain, then for any chain and above , there exists a chain above the chain in .
Proof.
The proof is clear and left to the reader. □
Definition 2.3.19.
Let . Then define the height of by
Then define the dimension of by
Corollary 2.3.20.
Let be an integral extension. Then
- 1.
Suppose lies above . Then .
- 2.
.
- 3.
If is integrally closed and is a domain, then we have .
Proof.
This is an immediate consequence of the previous corollary. □
2.4 Associated Primes
Let be an -module and .
Definition 2.4.1.
We say that is an associated prime of if one of the two following equivalent conditions hold.
- 1.
There exists such that ;
- 2.
There is an injection .
We will denote the set of associated primes using the unfortunate notation . Then the set of primes such that will be denoted .
Proposition 2.4.2.
Let be a maximal element of . Then .
Proof.
We will show that such a maximal element is actually a prime ideal. Suppose . Then for some nonzero , so . Then . By maximality, . Because , then . □
Corollary 2.4.3.
Let be Noetherian.
- 1.
is nonzero if and only if is nonempty.
- 2.
The set of zero divisors for is the union of the associated primes of .
Proof.
- 1.
If there is some associated prime, then clearly . In the other direction, the set of annihilators has a maximal element because is Noetherian, so there must be an associated prime.
- 2.
Let for some nonzero . Then is contained in some associated prime (because it is contained in some maximal element), and thus every zero divisor is contained in an associated prime. The other direction is obvious. □
Lemma 2.4.4.
Let be a multiplicative set and an -module. Then
Proof.
Let . Then for some , so must be empty. Next, we see that the set contains some maximal element because is Noetherian. But then .
On the other hand, if , then , which means for some . Then . Thus . Thus we have shown that
The other inclusion is clear. □
Theorem 2.4.5.
Let be Noetherian and and -module. Then and any minimal element of is inside .
Proof.
Let . Then injects in , so we have an injection . Thus .
Now choose a minimal . Thus is nontrivial, so there exists a prime ideal such that . Thus is nonzero, so . By minimality, and thus . Therefore . □
Definition 2.4.6.
If , then is not necessarily minimal in the support of . Then such a prime is called an embedded prime.
Proposition 2.4.7.
Let be Noetherian and a finitely-generated -module. Then
- 1.
There exists a chain
such that for and .
- 2.
Given such a sequence, we have . In particular this set is finite.
Proof.
- 1.
Suppose Then choose and let be the image of in . Then if is nonzero, choose and defined analogously to . This gives a sequence of submodules of such that . Because is Noetherian, this sequence becomes stationary. Thus there exists such that .
- 2.
This is a consequence of the next lemma.
□
Remark 2.4.8.
In general the support of a module is not finite.
Lemma 2.4.9.
Assume we have an exact sequence of modules . Then .
Proof.
If , there exists such that . Then if , and . If the intersection is nonzero, then there exists some nonzero such that because is a domain. Thus . □
Definition 2.4.10.
We say that is coprimary if .
Definition 2.4.11.
Let . Then we say that is -primary if . Alternatively, we say that belongs to .
Lemma 2.4.12.
A module is coprimary if and only if is nonzero and any zero divisor for is locally nilpotent (for all , there exists such that ).
Proof.
Suppose that is coprimary. Now suppose that and . Then , so is minimal in the support of , which is . Therefore, . Thus for , .
In the other direction, let be the set of locally nilpotent elements with respect to . This is clearly an ideal of . Then let . Then , so . Therefore because is a prime ideal. However, is contained in the set of zero divisors, which is precisely . □
Remark 2.4.13.
Let be an ideal. Then if and only if the zero divisors of are locally nilpotent. This is equivalent to being primary.
Lemma 2.4.14.
- 1.
Let be -primary submodules. Then is -primary.
- 2.
Let be an irredundant decomposition (i.e. is -primary) for distinct . Then .
Proof.
- 1.
Note that injects in . The desired result follows from the previous lemma.
- 2.
First, note that . Then suppose . Then we have an injection
and thus is contained in . By the exact sequence
we see that . □
Theorem 2.4.15.
Let be a module over a Noetherian ring . Then for all , there exists a -primary submodule such that
Proof.
Fix . Consider the set
This set is nonempty because . Next, if is a chain, then the module is a submodule of . In addition, . This implies that contains a maximal element by Zorn’s lemma. Choose such a maximal element .
We will show that is coprimary. By the exact sequence
if , then because otherwise would inject in as . Then , so , contradicting minimality. Thus .
The second part of the claim follows immediately from the fact that . □
Corollary 2.4.16.
Let be an -module of finite type. Then any has a primary decomposition
such that
- 1.
The are -primary;
- 2.
No can be omitted;
- 3.
This decomposition is irredundant: .
Proof.
Apply the previous theorem to . Because is of finite type, is finite. Then use the previous lemma. □
Exercise 2.4.17.
Let be a morphism of rings and let be a -module. Then prove that
where is the induced map of spaces.
3 Dimension Theory
3.1 Graded Rings and Modules
Let be a graded ring. This means that . Then an -module is a graded module if
such that . We will call the homogeneous elements of degree on .
Now let be a submodule. We say that is a graded submodule if . is also called homogeneous. A homogeneous element of is an element of some . Being a graded submodule is the same as every element being a sum of homogeneous elements.
Lemma 3.1.1.
The following are equivalent:
- 1.
is a homogeneous submodule.
- 2.
is generated by homogeneous elements.
- 3.
If for , then for all , .
Moreover, if is homogeneous, then so is , and
Proof.
The proof is left as an exercise to the reader. □
Example 3.1.2.
Let be any ring. Then the ring is a graded ring where the grading is by the degree of each monomial. In particular, . Then an ideal is graded if where . In addition, is a graded ring.
Proposition 3.1.3.
Let be a Noetherian graded ring and a graded -module. Then
- 1.
If , then is a graded ideal of and there exists a homogeneous such that .
- 2.
One can choose a -primary graded submodule such that .
Proof.
Let and set . Then write . Then for , write . If , then we can write
and deduce that . Then , so . By induction, all , so is graded.
The proof of the second part is simply the following lemma. □
Lemma 3.1.4.
Let be a graded prime ideal and such that is -primary. Let be the submodule of generated by the homogeneous elements of . Then is -primary.
Proof.
This will be proved later. □
We will now discuss filtrations of rings. A filtration is a sequence of subgroups
such that . If we set
then is a graded ring.
The basic example is for some fixed ideal . in this case, the filtration is called the -adic filtration.
Lemma 3.1.5.
Let be a Noetherian ring and set . Then
is a Noetherian graded ring.
Proof.
Because is finitely-generated, then is a finitely-generated -module. Thus is a finitely-generated -algebra. If is a set of generators of , then
is surjective, so because is Noetherian, so is and thus so is . □
Let be an Artinian ring and be a graded ring. Then let be a finitely-generated graded -module. Each graded piece is an -module, so write . Because is finitely generated, we have a map
Here, as a -module with the gradation . Thus is generated by homogeneous elements of degree . This gives us the map
and thus . But then is a free -module, and thus
Theorem 3.1.6.
Let be as above. Then there is a polynomial such that
for . This is called the Hilbert-Samuel polynomial for . The degree of this polynomial will give the first definition for the dimension of .
Proof.
Say that satisfies the property if there exists such that for .
- 1.
First, we will show that if and hold, then holds.
- 2.
Second, if is irreducible, then holds.
If we prove these two things, then we simply use a primary decomposition of . Then the second statement implies that holds and then we simply repeatedly apply the first to see that holds.
- 1.
Suppose with graded. We then have an exact sequence
and we know that is graded. Then we see that , so we only need to prove that exists. But then and because holds, so does .
- 2.
Let be irreducible. We know that is coprimary, so is -primary for some prime ideal . Write . If , then for . Indeed, if is the maximal degree of a system of generators of , then . On the other hand, because is -primary, then elements in are locally nilpotent. Thus there exists such that and thus for . Thus exists and is identically zero.
In the second case, . Then suppose that . Thus is not a zero divisor for . Thus, we have an exact sequence
which then gives
when restricting to a single graded piece. Thus . This implies that exists because above, we proved that if holds for any implies that holds. Then for , we have . This implies that
Then for some polynomial of degree and then . □
Now let and be an -module with filtration
We say that the filtration is
- 1.
-admissible if for all ;
- 2.
-acic if for all ;
- 3.
essentially -adic if for .
Remark 3.1.7.
A filtration on defines a topology on so that is a topological group. Here, a system of neighborhoods of is . If , then the topology is Hausdorff. If the filtration is essentially -adic, then the topology is called the -adic topology.
Lemma 3.1.8.
Let and be as before. Let be an -module with an admissible filtration. Let and
- 1.
is a -module.
- 2.
The filtration is essentially -adic if and only if is a finitely-generate -module.
Proof.
- 1.
This is trivial.
- 2.
Note that is a graded -module. If is finitely generated, then write . Then we see that for . Thus is essentially -adic.
Conversely, if for , then, then it is clear that is generated by and is thus finitely generated. □
Theorem 3.1.9 (Artin-Rees).
Let be a Noetherian ring and . Then let be a finitely-generated -module and be a submodule. Then there exists such that
for all .
Proof.
Let be the -adic filtration. Then is -admissible. Then both are both -modules. We know that is Noetherian, so because , is finitely generated. Thus is also Noetherian, so it is also finitely-generated. This implies that is essentially -adic, as desired. □
Remark 3.1.10.
This theorem is saying that the filtration is essentially -adic.
Corollary 3.1.11 (Krull Intersection Theorem).
Let as above.
- 1.
If , then .
- 2.
If , then .
Proof.
- 1.
Note that . Then apply the Artin-Rees theorem to .
- 2.
Apply Nakayama’s lemma. □
Corollary 3.1.12.
Let be a Noetherian domain and let be a proper ideal. Then .
Proof.
Let . By the previous corollary, . Then is finitely generated because is Noetherian. Thus there exists such that , which implies that because is a domain. □
Exercise 3.1.13.
Let be Noetherian and a finitely-generated -module. Then let be generated by -regular elements. Then there exists such that . Here, .
3.2 Other Notions of Dimension
Let be a ring. Then we define the Krull dimension
Then for any ideal , define the height of to be
Proposition 3.2.1.
For any ideal , we have .
Proof.
Consider a chain
where . Then we see that , as desired. □
Then if is an -module, define .
Proposition 3.2.2.
Assume that is Noetherian and is a finite -module. Then the following are equivalent:
- 1.
is of finite length.
- 2.
is Artinian.
- 3.
.
Proof.
Clearly conditions 2 and 3 are equivalent. Then is a quotient of , so 2 implies 1. Thus we need to prove that 1 implies 3.
Assume that . If we write , then is a finite -module. If , then there exists that is minimal but not maximal. Then because , we have . But then and thus we have an embedding . But then , so and thus . □
Now let be a semilocal ring. Let . Then an ideal is called an ideal of definition of if there exists such that .
Remark 3.2.3.
is an ideal of definition if and only if is Artinian.
Let be the graded ring with respect to the -adic filtration and let be the corresponding graded -module. Then if , define
Then we have a map , so is a -module. Now define . If is a finite -module, then is of finite length (because is Artinian) and thus
Then if is a polynomial in of degree at most for , then is a polynomial of degree at most for .
Now if is another ideal of definition, then there exists such that and thus . Therefore
and so the degree of is independent of . Denote this degree by . We know that , which is the number of generators of .
Lemma 3.2.4.
Assume we have an exact sequence of finite -modules
Then and is a polynomial of degree strictly less that .
Proof.
For each we have an exact sequence
Then . By Artin-Rees, there exists such that . But then
because has degree strictly less than , and the desired result follows. □
Lemma 3.2.5.
Let be a Noetherian semilocal ring. Then . In particular, .
Proof.
We will induct on . If , then for . By Nakayama, we see that , so and thus is Artinian.
Assume that and . Let
be a chain of prime ideals of length . Choose . Then . Because we have the exact sequence
we have and that has degree less than . Therefore
By induction, and thus . This holds for any chain of ideals, so . □
Corollary 3.2.6.
If is Noetherian and , then .
Proof.
by the previous lemma. □
Lemma 3.2.7.
Let be a Noetherian semilocal ring and a finite -module. Choose . Then .
Proof.
Let be an ideal of definition such that . Then
Then because , we see that , so
This implies that , so . □
Lemma 3.2.8.
Let and be as before. Let . Then there exists such that .
Proof.
Let be the minimal prime ideals containing such that . Because , then the are not maximal and therefore . In particular, it is not contained in . Choose .
If is prime and minimal, then because . This implies . By induction, we can then find such that
where . □
Theorem 3.2.9.
Let be semilocal and a finite -module. Then is the smallest integer such that there exists such that .
Proof.
Choose . If , then we know that . Then because has finite length, its dimsnion is zero and thus . Then let be the smallest such integer. By the previous lemma, we deduce that .
We will show that . Consider a sequence
such that for some prime ideals . Then are the minimal primes containing , so we see that
Remark 3.2.10.
If , then is the smallest integer such that there exists such that is an ideal of definition.
Corollary 3.2.11.
If is Noetherian and , then any minimal prime ideal containing has height at most . In particular, .
Proof.
First, note that is Artinian because the image of is both maximal and minimal. Then and thus . □
Now let be a finitely generated -module. Define for an ideal of definition . We call the -adic completion of .
Corollary 3.2.12.
.
Proof.
We know that . Thus the two modules have the same Hilbert-Samuel polynomial. □
Corollary 3.2.13.
Let be Noetherian with . Let be an integer. The following are equivalent:
- 1.
.
- 2.
There exits generated by elements such that is minimal in .
Proof.
- 1 implies 2
-
Suppose that . Then there exists an ideal of definition of generated by elements. If , then and is minimal containing .
- 2 implies 1
-
Let such that is minimal. Therefore is Artinian, so it has finite length. Thus , as desired. □
Definition 3.2.14.
A system of parameters for is a set of elements such that
- •
;
- •
.
Proposition 3.2.15.
Let . Then and we have equality if and only if belong to a system of parameters for .
Proof.
By induction for any . Then we know that
Then assume that we have equality. Let be a system of parameters for . Then . However, if , then
and thus is a system of parameters for .
Conversely suppose that is a system of parameters for . Then
but we have equality because
and so . □
Now we turn to the case of local Noetherian rings with maximal ideal and residue field . Then if , any ideal of definition has at least generators. Then let such that . Thus is an ideal of definition and is a system of parameters of .
Definition 3.2.16.
A local ring is a regular local ring if there is a system of parameters generating the maximal ideal of . Such a system is called a regular system of parameters.
Note that and that is regular if and only if .
Proposition 3.2.17.
Let be a local Noetherian ring. Let be a system of parameters of . Then and the image of in is a system of parameters of this quotient.
3.3 Dimension in the Relative Setting
Consider a morphism . We have the pullback . For , we have a bijection between and . This latter ring is isomorphic to , where is the residue field of .
Theorem 3.3.1.
Let lie over . Then
- 1.
. Equivalently, .
- 2.
Equality holds is equivalent to the going-down property for and in particular if is flat.
- 3.
If is surjective and the going-down property holds, then and for .
Proof.
- 1.
Set . We need to prove that , where is the maximal ideal of . Write and let be a system of parameters for . Then is an ideal of definition, so for some . Thus and all of these ideals have the same nilradical. Therefore
for some integer . If is a system of paramaters for , then generate an ideal of definition for , so .
- 2.
Let be a chain of ideals of of length . Then for we know and thus for all . Now by the Going Down property we can find
such that . Thus we have
where . This gives us the chain
and thus .
- 3.
The first inequality follows from 2. Note that .
To prove the equality, let be minimal such that . Let . Then and is minimal, so . This tells us that and thus . Thus .
Conversely, let be minimal with . Let such that . Then and so we may suppose it is minimal for this property. Then we see that
as desired. □
Corollary 3.3.2.
Let be Noetherian rings such that is integral over .
- 1.
;
- 2.
For all , .
- 3.
If the going-down property holds, then for any ideal , have .
Proof.
The proof of this is left as an exercise to the reader. □
Exercise 3.3.3.
Let be a morphism of rings and assume that going-down holds for . Let be prime ideals of . Prove that .
Now we will consider finitely generated extensions of rings. Here will be a finitely-generated -algebra.
Theorem 3.3.4.
Let be Noetherian. Then .
Proof.
Let and let such that . Choose to be maximal for this property. We will show that . After localization at , we may assume that is maximal and is local. Then , and is a field. Thus is a PID, so is a nonzero principal ideal, so it must have height exactly equal to .
Previous we have seen that because is flat, , and thus , and we obtain the desired result. □
Corollary 3.3.5.
- 1.
.
- 2.
If is a field, then . Moreover, .
Proof.
We only need to prove the part about the height of . Then we have
Then clearly and the inequality cannot be strict because otherwise has height strictly larger than . □
Exercise 3.3.6.
Let be Noetherian, , and . Suppose . Show that .
Theorem 3.3.7 (Noether Normalization).
Let be a finitely generated -algebra over a field . Let be a chain of prime ideals of . Then there exist elements algebraically independent such that:
- 1.
is integral over ;
- 2.
For all , there exists an integer such that . In particular, .
Remark 3.3.8.
Recall that are algebraically independent means that the map sending is injective.
Proof of Theorem.
We will treat the case where because any finitely generated algebra over which is an integral domain is integral over such a ring. We will prove the result by induction on and . The case when is clear because is a PID.
We will assume the result is true for . Now we will form an induction on . In the case , first assume that , where . Then . For we will introduce for some integer . We want to choose the such that is integral over . Writing
we see that is the maximal degree of in this expression. Then it is possible to choose such that are all distinct, for example for , where the max is taken over all that occur in the polynomial.
Then choosing the for which is maximal, we can write
Thus is integral over , so is integral over . Therefore is integral over . Finally, are algebraically independent because otherwise the transcendance degree of is smaller than . We now show that . If , then write for some . But then , so because is integrally closed, and thus .
For the general case, suppose is generated by more than one element. Choose and choose such that is integral over and . By the induction hypothesis on , there exist such that is integral over and . To see this, choose . Then there exist such that
bceause . Thus and thus . This finishes the case .
Now we complete the induction on . Suppose we have a chain of prime ideals in . Then we choose such that
- •
is integral over
- •
for .
Now we apply the case to the ideal where . Thus there exist such that is integral over and .
First, it is clear that is integral over . If we set for , then we will show that
One direction is obvious from the inductive hypothesis. In the other direction, if we write , then because , we see that
Thus , as desired. □
Corollary 3.3.9.
Let be an integral domain of finite type over a field . Then equals the transcendence degree of the fraction field of .
Proof.
There exist such that is integral over . Then is algebraic over . On the other hand, we know that , which is the transcendance degree of . □
Corollary 3.3.10 (Nullstellensatz).
Let be an algebra of finite type over a field . Then for any maximal ideal , is algebraic over .
Proof.
Note that , but this is also the transcendence degree over by the previous corollary. □
Proposition 3.3.11.
Let be an integral domain of finite type over a field . Then for any prime ideal , we have .
Proof.
Let and be its height. Then by Noether normalization, there exist , where , such that is integral over and . Then we know that . Because is integral over , we see that . But then because is integrally closed and is integral over . Therefore,
Remark 3.3.12.
Let be a finitely generated -algebra. Then for any maximal ideal , we know that is an algebraic extension of . Therefore we have a correspondence
In addition, any prime ideal
Proposition 3.3.13.
Let
Proof.
Choose
Remark 3.3.14.
We can think of
Proposition 3.3.15 (Hilbert’s Nullstellensatz).
Let
Proof.
One direction is obvious because
Then writing
3.4 Rings of Dimension 1
Definition 3.4.1.
A local ring
This prime ideal is naturally maximal becuase if
Definition 3.4.2.
A discrete valuation on
- •
;v ( x y ) = v ( x ) + v ( y ) - •
.v ( x + y ) ≥ min { v ( x ) , v ( y ) }
We define
If
Proposition 3.4.3.
If
Proof of this is left as an exercise.
Example 3.4.4.
Consider the ring
Now let
Proposition 3.4.5.
Let
- 1.
is a DVR.A - 2.
is a local noetherian ring andA is generated by an element𝔪 A which is not nilpotent.π
Proof.
One direction is clear. If
In the other direction, suppose
Proposition 3.4.6.
Let
- 1.
is integrally closed;A - 2.
has a unique nonzero prime ideal.A
Proof.
One direction is clear. In the other direction, assume
this is an
Let
where
Proposition 3.4.7.
Let
- 1.
For all
, the localization0 ≠ 𝔭 ∈ Spec A is a discrete valuation ring.A 𝔭 - 2.
is integrally closed and of dimensionA .1
Proof.
- 1 implies 2:
-
Let
be integral overx ∈ K . But thenA for allx ∈ A 𝔭 and thus𝔭 ≠ 0 . Being of dimensionx ∈ ⋂ 𝔭 ≠ 0 A 𝔭 = A is easy. If1 , we localize at0 ≠ 𝔭 ⊂ 𝔪 ⊂ A and see that𝔪 , so𝔭 A 𝔪 = 𝔪 A 𝔪 .𝔭 = 𝔪 - 2 implies 1:
-
For all
, we know that𝔭 is integrally closed. Because it has dimensionA 𝔭 , it must be a DVR.1
□
Definition 3.4.8.
A ring
Example 3.4.9.
Example 3.4.10.
Let
Here is another example of this. Let
Remark 3.4.11.
All smooth curves can be obtained in this way (of taking the integral closure of some ring). Also, normalization resolves all singluarities of curves.
Definition 3.4.12.
A fractional ideal
Note that for
Proposition 3.4.13.
If
Proof.
It is easy to see that the multiplication is associative and commutative. Now we need to show that inverses exist. For
Clearly we know that for
and thus
Now if
Lemma 3.4.14.
If
Proof.
Let
is a strictly decreasing infinite chain of ideals containing
but this is impossible because
Remark 3.4.15.
For any nonzero ideal
Definition 3.4.16.
Let
Corollary 3.4.17.
Let
Proof.
If we denote the product by
Remark 3.4.18.
This corollary gives the primary decomposition of an ideal in a Dedekind domain. In number theory, this replaces the prime factorization of an integer.
3.5 Depth
Let
Definition 3.5.1.
The sequence
- 1.
is not a zero divisor ofa i forM / ( a 1 , … , a i − 1 ) M .i = 1 , … , r - 2.
. Therefore we havea ¯ ⋅ M ⊊ M a 1 M ⊊ ( a 1 , a 2 ) M ⊊ ⋯ ⊊ ( a 1 , … , a r ) M .
Lemma 3.5.2.
Assume that
Proof.
We will induct on
Thus
Theorem 3.5.3.
Assume that
Proof.
It is sufficient to prove that
Now if
By induction on
Multiplying this by
By the previous lemma, we see that
Definition 3.5.4.
The sequence
- •
For all
homogeneous of degreeF ( x 1 , … , x r ) ∈ M [ x 1 , … , x r ] = A [ x 1 , … , x r ] ⊗ A M such thatn , this implies thatF ( a 1 , … , a r ) ∈ I n + 1 M , whereF ( x 1 , … , x r ) ∈ I M [ x 1 , … , x r ] .I = ( a 1 , … , a r ) - •
If
is homogeneous and such thatF ( x 1 , … , x r ) ∈ M [ x 1 , … , x r ] , thenF ( a 1 , … , a r ) = 0 F ∈ I M [ x 1 , … , x r ] - •
The map
M / I M [ x 1 , … , x r ] → gr I M = ⨁ n = 0 ∞ I n M / I n + 1 M , F ↦ F ( a 1 , … , a r ) is an isomorphism.
Lemma 3.5.5.
Assume that
Proof.
We will induct on
and then quasi-regularity gives us that
as desired. □
Proposition 3.5.6.
Using the same notation, if
Proof.
First we prove that regular implies quasi-regular by induction on
For the inductive step, suppose that
where
Because
Now in the other direction, we will induct on
Now for the induction, we know that
vanishes at
We can repeat this until
is homogeneous of degree
Then
Remark 3.5.7.
For
Definition 3.5.8.
Let
Before we continue, recall that
and similarly,
Theorem 3.5.9.
Assume that
- 1.
for allExt n ( N , M ) = 0 and any finitelyi < n -moduleA such thatN .supp ( N ) ⊂ V ( I ) - 2.
for allExt i ( A / I , M ) = 0 .i < n - 3.
for allExt i ( N , M ) = 0 for some finitei < n -moduleA such thatN .supp ( N ) = V ( I ) - 4.
There exists a
-regular sequenceM of length( a 1 , … , a n ) insiden .I
Proof.
Clearly 1 implies 2 implies 3. Now we prove 3 implies 4 implies 1.
- 3 implies 4:
-
Assume that
. IfHom ( N , M ) = 0 does not contain anyI -regular element, thenM and thusI ⊂ ⋃ 𝔭 ∈ Ass M 𝔭 for someI ⊂ 𝔭 . Then𝔭 ∈ Ass ( M ) , which is equivalent toA / 𝔭 ↪ M . On the other hand, we knowA 𝔭 / p A 𝔭 ↪ M 𝔭 , so𝔭 ∈ V ( I ) = supp ( N ) . By Nakayama, we see thatN 𝔭 ≠ 0 , and thusN 𝔭 ⊗ k ( 𝔭 ) ≠ 0 , but this is simplyHom A 𝔭 ( N 𝔭 , M 𝔭 ) ≠ 0 . ThusHom A ( N , M ) 𝔭 and thus there existsHom ( N , M ) ≠ 0 that isa 1 ∈ I -regular. Then we have an exact sequenceM 0 → M → M → M / a 1 M → 0 . Writing
, we have proved the caseM 1 = M / a 1 M and now proceed by induction onn = 1 . Applyingn to the above exact sequence, we have the exact sequenceExt n ( N , − ) ⋯ → Ext i ( N , M ) → Ext i ( N , M 1 ) → Ext i + 1 ( N , M ) and deduce that
forExt i ( N , M 1 ) = 0 . Applying the casei < n − 1 ton − 1 , we obtain anM 1 -regular sequenceM 1 and thusa 2 , … , a n ∈ I is( a 1 , … , a n ) -regular.M - 4 implies 1:
-
We will induct on
. We have a sequencen that isa 1 , … , a n ∈ I -regular. Then setM , so we have an exact sequenceM 1 = M / a 1 M 0 → M → M → M 1 → 0 . Then choose
such thatN . Then we have an exact sequencesupp ( N ) ⊂ V ( I ) Ext i − 1 ( N , M 1 ) → Ext i ( N , M ) → Ext i ( N , M ) . If
, theni < n . By the inductive hypothesis, this implies thati − 1 < n − 1 . Thus the mapExt i − 1 ( N , M 1 ) = 0 is injective, so becauseExt i ( N , M ) → × a 1 Ext i ( N , M ) , we havesupp ( N ) ⊂ V ( I ) . Thus there existsI ⊂ Ann ( N ) such thatm . From the exact sequence, we knowa 1 m ∈ Ann ( N ) is injective, which means thatExt i ( N , M ) → × a 1 m Ext i ( N , M ) . □Ext i ( N , M ) = 0
Definition 3.5.10.
If
This is the same as the maximal length of an
Corollary 3.5.11.
Let
Proof.
We know that the equivalence of 2 and 4 from the theorem implies the corollary except for
is exact because this implies that
Lemma 3.5.12.
Let
Proof.
We use induction on
where
and thus
Remark 3.5.13.
If
Theorem 3.5.14.
Let
Proof.
We know
Corollary 3.5.15.
Let
In general, this inequality is strict, so we will later study the rings for which this is an equality.
Lemma 3.5.16.
Let
Proof.
We prove this by induction on
Lemma 3.5.17.
Let
Proof.
Let
and therefore
4 Geometric Aspects of Commutative Algebra
4.1 Cohen-Macaulay Rings and Modules
Let
Definition 4.1.1.
Recall that
Theorem 4.1.2.
Let
- 1.
If
is Cohen-Macaulay, then for anyM ,𝔭 ∈ Ass ( M ) .depth M = dim A / 𝔭 - 2.
If
isf ∈ A -regular andM , thenM ′ = M / f M is Cohen-Macaulay if and only ifM is.M ′ - 3.
If
is Cohen-Macaulay, then for allM ,𝔭 ∈ Spec A is Cohen-Macaulay andM 𝔭 .depth 𝔭 M = depth A 𝔭 M 𝔭
Proof.
- 1.
Let
andM ≠ 0 . Then letdim M = depth M . This implies that𝔭 ∈ Ass ( M ) ⊂ supp ( M ) , but alsodim A / 𝔭 ≤ dim M .dim A / 𝔭 ≥ depth ( M ) = dim M - 2.
Let
bef -regular. Then we knowM (this follows from the Theorem 3.5.9, applyingdepth M / f M = depth M − 1 to the exact sequenceExt ( k , − ) ). But then we proved earlier that0 → M → M → M / f M → 0 , and thusdim M / f M = dim M − 1 is C-M if and only ifM is C-M.M / f M - 3.
Let
. Then𝔭 ∈ supp ( M ) and𝔭 ⊃ Ann ( M ) . But then ifM 𝔭 ≠ 0 isx 1 , … , x r ∈ 𝔭 -regular whereM , thenr = depth 𝔭 M isx 1 , … , x r ∈ 𝔭 A 𝔭 -regular by exactness of localization. This implies thatM 𝔭 . We know thatdepth M 𝔭 ≥ r = depth 𝔭 M , so we need to prove thatdim M 𝔭 ≥ depth M 𝔭 . We do this by induction onr = dim M 𝔭 . Ifdepth 𝔭 M , then we knowr = 0 . Thus there existsHom ( A / 𝔭 , M ) ≠ 0 such that𝔭 ′ ⊃ 𝔭 , soA / 𝔭 ′ ↪ M . By minimality of associated primes, we have𝔭 ′ ∈ Ass M . Now𝔭 ′ = 𝔭 becausedim M 𝔭 ′ = 0 is maximal in𝔭 and minimal inA 𝔭 .supp ( M ) Now in general, assume
. Letdepth 𝔭 ( M ) > 0 bea ∈ 𝔭 -regular. ThusM isa -regular, so setM 𝔭 . We know thatM 1 = M / a M by the inductive hypothesis. This implies thatdim ( M 1 ) 𝔭 = depth 𝔭 M 1 becausedim M 𝔭 = depth 𝔭 M anddim ( M 1 ) 𝔭 = dim M 𝔭 − 1 . □depth 𝔭 M 1 = depth 𝔭 M − 1
Theorem 4.1.3.
Let
- 1.
For all proper ideals
, we haveI ⊊ A .ht ( I ) + dim A / I = dim A - 2.
For all
prime ideals, thenP ⊃ Q , where the last height is taken inht ( P ) − ht ( Q ) = ht ( P / Q ) .A / Q - 3.
Let
. Then the following are equivalent:x 1 , … , x r ∈ 𝔪 - (a)
arex 1 , … , x r -regular.A - (b)
for allht ( x 1 , … , x i ) = i .i ≤ r - (c)
.ht ( x 1 , … , x r ) = r - (d)
There exists
withx r + 1 , … , x n ∈ 𝔪 such thatn = dim A is a system of parameters forx 1 , … , x n .A
- (a)
Proof.
We will first prove 3 and then use this to prove 1 and 2.
- a implies b:
-
We already know that
. First, we know thatht ( x 1 , … , x i ) ≤ i does not belong to any minimal prime ideal ofx 1 becauseA isx 1 -regular. ThusA . Then inht ( x 1 ) = 1 , the sequenceA / x 1 isx ¯ 1 , … , x ¯ r -regular, and thusA / x 1 . Thus we can findht ( x ¯ 2 , … , x ¯ i ) = i − 1 𝔭 i ⊋ 𝔭 i − 1 ⊊ ⋯ ⊋ 𝔭 1 ⊃ ( x 1 ) such that
. This implies that𝔭 ¯ i ⊋ ⋯ ⊋ 𝔭 ¯ 1 ⊋ 0 and thusht ( 𝔭 i ) ≥ 1 .ht ( x 1 , … , x r ) ≥ i - b implies c:
-
This is trivial.
- c implies d:
-
Assume
. Then there existsr < dim A that does not belong to any minimal prime ideal containingx r + 1 ∈ 𝔪 . Therefore( x 1 , … , x r ) and then we can findht ( x 1 , … , x r + 1 ) = r + 1 such thatx r + 2 , … , x n . But this implies thatht ( x 1 , … , x n ) = n is an ideal of definition forI = ( x 1 , … , x n ) and soA is a system of parameters.( x 1 , … , x n ) - d implies a:
-
Note that so far we have not used the fact that
is Cohen-Macaulay. We show that a system of parameters isA -regular. LetA . Then𝔭 ∈ Ass ( A ) . This implies thatdim A / 𝔭 = n is minimal and therefore𝔭 inx ¯ 1 ≠ 0 because otherwiseA / 𝔭 would be a system of parameters inx ¯ 1 , … , x ¯ n . But then this meansA / 𝔭 for allx 1 ∉ 𝔭 . This implies that𝔭 ∈ Ass ( A ) isx 1 -regular, and thusA is Cohen-Macaulay of dimensionA / x 1 A . On the other hand,n − 1 is a system of parameters forx ¯ 2 , … , x ¯ n and the desired result follows by induction.A / x 1 A
Now we will prove 1 and 2.
- 1.
Let
be an ideal with heightI ⊊ A . Then we can chooseht ( I ) = r such thata 1 … , a r ∈ I . By 3a, this means thatht ( a 1 , … , a r ) = r isa 1 , … , a r -regular and thusA . On the other hand, if we haver ≤ depth I ( A ) anb 1 , … , b s ∈ I -regular sequence, we see thatA . This tells us thats ≤ ht ( I ) . But then recall thatht ( I ) = depth I A ht ( I ) = inf { ht ( p ) ∣ p ⊃ I } and that
dim A / I = sup { dim A / 𝔭 , 𝔭 ⊃ I } . This means we can assume that
is a prime ideal withI = p . In this case, we know that𝔭 = r is Cohen-Macaulay andA 𝔭 ht ( p ) = dim A 𝔭 = depth 𝔭 ( A ) so there exists
anx 1 , … , x r ∈ I -regular sequence such thatA is Cohen-Macaulay of dimensionA / ( x 1 , … , x r ) . Thenn − r is minimal in𝔭 ¯ because otherwise we would be able to fineA / ( x 1 , … , x r ) which is not a zero divisor inx r + 1 ∈ 𝔭 and thenA / ( x 1 , … , x r ) . This impliesdepth ( 𝔭 ) ≥ r + 1 dim A / 𝔭 = dim A / ( x 1 , … , x r ) = n − r and thus
.dim A / 𝔭 + ht ( 𝔭 ) = n − r + r = n = dim A - 2.
Suppose
are prime ideals. We knowP ⊃ Q is Cohen-Macaulay. Then we know thatA P by 1 fordim A P = ht ( Q A P ) + dim ( A P / Q A P ) . But then we have the desired result becauseA P ,dim A P = ht ( P ) , andht ( Q A P ) = ht ( Q ) . □dim ( A P / Q A P ) = ht ( P / Q )
4.2 Normal and Regular Rings
Let
Remark 4.2.1.
Clearly if
To prove the first part, write
Then choose
To prove the second part, note that if
Definition 4.2.2.
We say that
Remark 4.2.3.
Thering
We also note that if
where
Now we consider the normal rings of dimension
Proposition 4.2.4.
Let
- 1.
is a discrete valuation ring.A - 2.
is normal andA .dim A = 1 - 3.
is normal and there existsA such that0 ≠ a ∈ 𝔪 is( a ) -primary.𝔪 - 4.
The maximal ideal
of𝔪 is principal.A
Proof.
We will proceed quickly.
- 1 implies 2:
-
This was done previously.
- 2 implies 3:
-
Note that
is the only prime ideal containing𝔪 because( a ) has dimensionA and is a local domain. This means that1 is( a ) -primary.𝔪 - 3 implies 4:
-
We know that
and thus𝔪 ∈ Ass ( A / ( a ) ) for some𝔪 = Ann A ( x ¯ ) . If0 ≠ x ¯ ∈ A / a A is a lift ofx ∈ A , thenx ¯ . However, we know thatx a − 1 ∉ A , so𝔪 ⋅ x ⊆ ( a ) . This tells us that𝔪 x a − 1 ⊂ A is a strict inclusion because otherwise𝔪 ⊊ 𝔪 x a − 1 is integral overx a − 1 and thusA . Thereforex a − 1 ∈ A , so there exists𝔪 x a − 1 = A such thatt ∈ 𝔪 and thust x a − 1 = 1 and thust A ⊂ 𝔪 ∋ y , wherey = t x a − 1 y = t ( x a − 1 y ) , and thusx a − 1 y ∈ A , soy ∈ t A is principal.𝔪 = t A - 4 implies 1:
-
If
, then we know that𝔪 = ( t ) ⋂ n ≥ 0 𝔪 n = 0 and so for all
there existsx ≠ 0 such thatm wherex = u t m , and thusu ∈ A × is a DVR. □A
Now let
Theorem 4.2.5.
Let
- 1.
If
is a domain, then so isA ∗ .A - 2.
If
is Noetherian andA , then ifI ⊂ rad ( A ) is a normal domain, so isA ∗ .A
Proof.
- 1.
Choose
nonzero. Thena , b ∈ A and soa ∗ , b ∗ ≠ 0 becausea ∗ b ∗ ≠ 0 is a domain. This means that ifA ∗ , we know thata ∈ I n ∖ I n + 1 , b ∈ I q ∖ I q + 1 and thusa b ∈ I n + q ∖ I n + q + 1 .a b ≠ 0 - 2.
Let
be integral overx ∈ K . WriteA wherex = a b . Thena , b ∈ A is Noetherian and therefore by the Krull intersection theorem this is the same thing asA / b A ⋂ n ≥ 0 I n + b A = b A . To show that
, we will prove thatx ∈ A or thata ∈ b A for alla ∈ I n + b A . Ifn ≥ 0 , then we know thata ∈ I n − 1 + b A , wherea = a ′ + b s anda ′ ∈ I n − 1 . Then we know that becauses ∈ A is integral, so isa b . Then we may assume thata ′ b anda ∈ I n − 1 is almost integral. Then there existsa b such thatc ∈ A for allc a m ∈ b m A and thus becausem ≥ 1 , we know thata ∗ ≠ 0 and thus( a ∗ ) m = ( a m ) ∗ ≠ 0 . Becausec ∗ ( a ∗ ) m ∈ ( b ∗ ) m A is Noetherian and normal, it is completely normal, and thusA ∗ and thusa ∗ b ∗ ∈ A ∗ for somea ∗ = b ∗ d ∗ . This means that0 ≠ d ∈ A fora − b d ∈ I N + 1 such thatN . But we know this fora ∈ I N ∖ I N + 1 , and thusN = n − 1 , soa − b d ∈ I n . □a ∈ I n + b A
Recall that if
Theorem 4.2.6.
Let
where
Proof.
Assume that
This map is clearly surjective, so we need to show that it is injective. If
and thus
Now assume that
Corollary 4.2.7.
Let
- 1.
is a discrete valuation ring.A - 2.
is a normal ring of dimensionA .1 - 3.
is regular of dimensionA .1
Proof.
We have seen before that 1 is equivalent to 2 which implies 3. To prove that 3 implies 1, we need to show that
Corollary 4.2.8.
If
Theorem 4.2.9.
Let
- 1.
is a normal domain.A - 2.
is anx 1 , … , x d -regular sequence andA is Cohen-Macaulay.A - 3.
The ideal
is a prime ideal of height𝔭 i = ( x 1 , … , x i ) andi is regular of dimensionA / 𝔭 i .d − i - 4.
If
and𝔭 ∈ Spec ( A ) is regular, then there exists a system of parametersA / 𝔭 such thaty 1 , … , y d and𝔭 = ( y 1 , … , y i ) .dim A / 𝔭 = d − i
Proof.
We prove each part.
- 1.
We already proved that the graded ring
is a normal domain.gr 𝔪 ( A ) - 2.
Use the previous theorem. Recall that
. This means thatk [ x 1 , … , x d ] ≃ gr 𝔪 ( A ) is anx 1 , … , x d -quasi-regular sequence. BecauseA is Noetherian, thenA is anx 1 , … , x d -regular sequence. This implies thatA d ≤ depth ( A ) ≤ d and thus
is Cohen-Macaulay.A - 3.
The previous implies that
. Then the idealdim A / ( x 1 , … , x i ) = d − i is generated by the𝔪 ¯ = 𝔪 / 𝔭 i elementsd − i and these form a system of parameters. By definition,x ¯ i + 1 , … , x ¯ d is regular.A / 𝔭 i - 4.
Write
. Then let𝔪 ¯ = 𝔪 / 𝔭 such thati ≤ d . This means thatdim 𝔪 ¯ / 𝔪 ¯ 2 = d − i , so we can choosei = dim ( 𝔭 + 𝔪 2 ) / 𝔪 2 such thaty 1 , … , y i ∈ 𝔭 is a basis ofy ¯ 1 , … , y ¯ i . We then complete this to a basis of( 𝔭 + 𝔪 2 / 𝔪 2 ) . Thus𝔪 / 𝔪 2 is a regular system of parameters fory 1 , … , y d , and is thus is regular sequence. ThenA is a prime ideal and𝔭 ′ = ( y 1 , … , y i ) ⊂ 𝔭 . But thendim A / 𝔭 ′ = d − i becausedim A / 𝔭 = p − i is regular by assumption, so we must haveA / 𝔭 . □𝔭 = 𝔭 ′
We now return to some definitions.
Definition 4.2.10.
Let
Recall that if
Definition 4.2.11.
Let
Definition 4.2.12.
We say that the unmixed theorem holds for
Remark 4.2.13.
This is equivalent to saying that for all
To see this, note that if
Theorem 4.2.14.
Let
Proof.
Suppose that the unmixed theorem holds for
Now assume that
Returning to normal domains, we have the following result:
Theorem 4.2.15.
Let
In particular,
Proof.
Let
This means that
and thus we can replace
Then one of these two inclusions is an equality, so if the first inclusion is an equality, we know that elements of
Now we will prove that
4.3 Homological Theory
Lemma 4.3.1.
Let
- 1.
is projective if and only ifM for allExt A 1 ( M , N ) = 0 -modulesA .N - 2.
is injective if and only ifM for all idealsExt A 1 ( A / I , M ) = 0 .I ⊂ A
Proof.
We only need to prove the second part. One direction follows from the direction, so now suppose we have
We will show that
Now we apply the functor
Definition 4.3.2.
Let
Lemma 4.3.3.
Let
- 1.
All
-modulesA have projective dimension at mostM ;n - 2.
All
-modulesA of finite type have projective dimension at mostM ;n - 3.
All
-modulesA have injective dimension at mostM ;n - 4.
for allExt A n + 1 ( M , N ) = 0 -modulesA .M , N
Proof.
Clearly 1 implies 2.
- 2 implies 3:
-
Consider
0 → M → U 0 → U 1 → ⋯ → U n − 1 → C → 0 such that
are injective. Now we have an exact sequenceU 0 , … , U n − 1 0 → M → U 0 → V 0 → 0 where
is the kernel ofV 0 . For any idealU 1 → U 2 , we know thatI becauseExt n + 1 ( A / I , M ) = Ext n ( A / I , V 0 ) is injective. Now we obtain thatU 0 whereExt n ( A / I , V 0 ) = Ext n − 1 ( A / I , V 1 ) . This implies thatV 1 = ker ( U 2 → U 3 ) Ext A n + 1 ( A / I , M ) = Ext A 1 ( A / I , C ) . By 2, we know that
and thereforeExt n + 1 ( A / I , M ) = 0 is injective. Thus there is an injective resolution ofC of lengthM .n - 3 implies 4:
-
This is trivial.
- 4 implies 1:
-
The proof is similar to the proof that 2 implies 3. □
Corollary 4.3.4.
The two numbers
are equal and are called the global homological dimension of
Lemma 4.3.5.
Suppose
Proof.
Because
where
Lemma 4.3.6.
Let
Proof.
Let
In the other direction, first assume that
we obtain the exact sequence
and this implies
where each
Corollary 4.3.7.
Let
- 1.
The projective dimension of
is the supremum of the projective dimensions ofM for all maximal idealsM 𝔪 of𝔪 .A - 2.
if and only ifproj . dim M ≤ n for all maximal idealsTor n + 1 ( M , A / 𝔪 ) = 0 of𝔪 .A
Proof.
- 1.
This follows from the fact that
and the fact that for anExt A i ( M , N ) 𝔪 = Ext A 𝔪 i ( M 𝔪 , N 𝔪 ) -moduleA we haveM if and only ifM = 0 for allM 𝔪 = 0 maximal.𝔪 - 2.
This follows from 1, the previous lemma, and the fact that
commutes with localization. □Tor
Lemma 4.3.8.
Let
- 1.
.gl . dim A ≤ n - 2.
for all finiteproj . dim M ≤ n -modulesA .M - 3.
for all finiteinj . dim M ≤ n -modulesA .M - 4.
for all finiteExt A n + 1 ( M , N ) = 0 -modulesA .M , N - 5.
for all finiteTor n + 1 A ( M , N ) = 0 -modulesA .M , N
Proof.
We know that 1 is equivalent to 2, which implies 3. We know that 3 implies 4 and 2 implies 5 are trivial. The previous corollary gives us 5 implies 2, so we need to prove that 4 implies 2. Let
such that
This implies that
Corollary 4.3.9.
We have the local-global identity
Theorem 4.3.10.
Let
Proof.
Suppose that
Lemma 4.3.11.
Let
Proof.
Let
so after applying the functor
but
Theorem 4.3.12.
Let
Proof.
Let
and thus
Corollary 4.3.13.
Let
Proof.
Let
4.4 Koszul Complex
Our goal is to prove the following result:
Theorem 4.4.1 (Serre).
A local Noetherian ring is regular if and only if its global dimension is finite.
To do this, we will introduce the Koszul complex. First, recall that if
For
Then for an integer
This satisfies
Now consider the exact sequence
Now if
This gives us a long exact sequence of modules
By definition of
Thus if
Theorem 4.4.2.
Let
for all
Corollary 4.4.3.
If
Now we will discuss minimal resolutions. Suppose
satisfies
Proposition 4.4.4.
Now assume
Proof.
To prove existence, begin with
To prove uniqueness, if
that makes this diagram commute. In degree zero, we have
Lemma 4.4.5.
Let
- 1.
is a finite freeF p -module for allA ;p - 2.
is surjective.F ¯ 0 → M ¯ - 3.
for alld ( F i ) ⊂ 𝔪 F i − 1 and induces a surjectioni > 0 .F ¯ i ↪ 𝔪 / 𝔪 2 ⊗ F ¯ i − 1 = 𝔪 ⊗ F ¯ i − 1
Then
Proof.
Consider a minimal free resolution
Then the map
Now we have a diagram
where we have
and therefore
Theorem 4.4.6.
Let
In particular,
Proof.
Write
Reducing modulo
We are now ready to prove the theorem of Serre.
Proof of Theorem 4.4.1.
We proved one direction last time. Now set
and because
Corollary 4.4.7.
Let
Proof.
This requires showing that
4.5 Unique Factorization
Recall that a unique factorization domain, or a UFD, is a domain
Exercise 4.5.1.
If
We say that two irreducibles
then
Lemma 4.5.2.
A Noetherian domain is a UFD if and only if
Proof.
Suppose
In the other direction, we will prove that Gauss’s lemma is satisfied. This says that if
Lemma 4.5.3.
Let
Proof.
Suppose
In the other direction, we will show that for any irreducible
Lemma 4.5.4.
Let
Proof.
Suppose
we must have
Theorem 4.5.5.
Let
Proof.
We induct on
For the inductive step, choose
where
and so because
where
Because
Writing the first direct sum as
Definition 4.5.6.
Let
4.6 Criterion for Normality
Let
:( R k ) -
For all
such that𝔭 ∈ Spec A , thenht ( 𝔭 ) ≤ k is regular.A 𝔭 :( S k ) -
For all
,𝔭 ∈ Spec A .depth ( A 𝔭 ) ≥ inf ( k , ht ( 𝔭 ) )
Theorem 4.6.1.
Let
Before we prove this, we will prove several lemmas.
Lemma 4.6.2.
The ideal
Proof.
Assume that
Conversely, suppose
Corollary 4.6.3.
The property
Proof.
Use the previous lemma and an induction argument. Also use the fact that
Recall that a domain
Definition 4.6.4.
A ring
It is not hard to see that it suffices to check this condition for maximal ideals.
Corollary 4.6.5.
Assume that
Here, the intersection is well-defined because the intersection of all minimal prime ideals is
Proof.
Assume that
Lemma 4.6.6.
Assumes that
Proof.
Choose
Now assume that
Proof of Theorem 4.6.1.
First assume that
Now assume that
Let