1. Differential Rings
Let A be a commutative ring with identity. A derivation on A is a map d : A → A such that
(1) d(a + b) = da + db (2) d(ab) = da · b + a · db.
A differential ring is a ring A with a derivation d. Furthermore, if A is a field (integral domain e.t.c), A is called a differential field (differential integral domain e.t.c.). We write a0 for da, a00 for d2a and in general a(n) = dna, a ∈ A. In this section, A is always assumed to be a differential ring.
Lemma 1.1. Let a, b ∈ A and n be a natural number. Then (1) (ab)(n)=Pn
k=0 n
ka(k)b(n−k) for each n ≥ 0.
(2) (an)0 = nan−1a0.
(3) If a is a unit, (2) holds for all n ∈ Z.
(4) (1)0 = 0.
Proof. We leave it to the reader.
Definition 1.1. Let (A, dA) be a differential ring. If (B, dB) is a differential ring such that A is a subring of B and dB|A= dA, we say that (A, dA) is a differential subring of (B, dB) and (B, dB) is an extension of (A, dA).
Theorem 1.1. A derivation of an integral domain has a unique extension to its field of fraction.
Proof. Let A be a differential integral domain and Q(A) be its field of fraction. Define
∂ : Q(A) → Q(A) by
∂a b
= a0b − ab0 b2 , a
b ∈ Q(A).
We leave it to the reader to check that ∂ is a well-defined derivation. This proves the existence.
If δ : Q(A) → Q(A) is a derivation such that δ|Q(A) = d, by Lemma 1.1 (3), given a/b ∈ Q(A) with a, b ∈ A and b 6= 0,
δ(ab−1) = δa · b−1− ab−2δb = δa · b − aδb
b2 = a0b − ab0 b2 .
Here we use δa = a0 and δb = b0 since δ|A = d. Thus δ = ∂. This proves the uniqueness of
extension.
Example 1.1. Let C∞(a, b) be the ring of all real-valued differentiable functions on the open interval (a, b). Then
D : C∞(a, b) → C∞(a, b), f 7→ f0 is a derivation. Thus C∞(a, b) has a structure of a differential ring.
Example 1.2. Let k be a field and k[X] be the polynomial ring over k. Define D : k[X] → k[X],
n
X
i=0
aiXi 7→
n
X
i=0
iaiXi−1.
Then D is a derivation on k[X]. Since k[X] is an integral domain, D can be extended to the field k(X) of rational functions over k by Theorem 1.1.
1
2
Example 1.3. Let U be an open subset of C and O(U ) be the ring of holomorphic functions on U. The map ∂ : O(U ) → O(U ) sending f to f0(z) is a derivation. Furthermore, if U is connected, then O(U ) is an integral domain and thus ∂ can be extended to its field of fraction: Q(O(U )) = M(U ) the field of meromorphic functions on U.
Example 1.4. Let A[X] be the polynomial ring over a differential ring A. We define
∂ : A[X] → A[X],
n
X
i=0
aiXi 7→
n
X
i=0
a0iXi+
n
X
i=0
iaiXi−1. Then ∂ is an extension of d on A[X].
Example 1.5. Let A be a differential ring and A[{xi : i ≥ 0}] be the polynomial ring over A in the infinite sequence of variables {xi : i ≥ 0}. Define a D derivation on A[{xi : i ≥ 0}]
such that
dxi−1= xi, i ≥ 1.
We set x = x0. The resulting differential ring (A[{xi : i ≥ 0}], d) is denoted by A{x}.
Elements of A{x} are called differential polynomials in x.
Lemma 1.2. Let C be the subset of (A, d) consisting of a ∈ A such that a0 = 0. Then C is a subring of A containing the subring of A generated by unit elements of A. The ring C is called the ring of constants.
Proof. One check that if a, b ∈ C, so are a + b and ab.
Let I be an ideal in a differential ring A. We say that I is a differential ideal of A if a ∈ I implies a0 ∈ I.
Lemma 1.3. Let {Iλ} be a family of differential ideals of a differential ring A. ThenT
λIλ
is again a differential ideal.
Proof. Exercise.
Lemma 1.4. Let A be a differential ring and I be a differential ideal of A. On A/I, we define a derivation d by sending a + I to a0+ I. Then (A/I, d) is a differential ring.
Proof. Exercise.
Definition 1.2. A morphism from differential ring A to differential ring B is a ring homo- morphism ϕ : A → B such that
dB◦ ϕ = ϕ ◦ dA.
One can easily see that differential rings together with morphisms defined above forms a category.
Lemma 1.5. Let A be a differential ring and I be a differential ideal of A. Then the quotient map π : A → A/I is a morphism of differential rings.
Proof. Exericise.
Proposition 1.1. Let ϕ : A → B be a morphism of differential rings and I be its kernel.
Then I is a differential ideal of A and the morphism ϕ : A → B induces a monomorphism ϕ : A/I → B of differential rings such that ϕ = ϕ ◦ π, where π : A → A/I is the quotient map.
Proof. The proof is straightforward.
3
Let us recall that an ideal I of A is a radical ideal if an∈ I implies that a ∈ I.
Lemma 1.6. Let I be a differential radical ideal of a differential ring A. If ab ∈ I, then both ab0 and ab0 belong to I.
Proof. Since ab ∈ I and I is a differential ideal, then (ab)0 = a0b + ab0 ∈ I. Multiplying the equation by ab0, we see that a0b0(ab) + (ab0)2 ∈ I. Since ab ∈ I, a0b0(ab) ∈ I. Thus
(ab0)2= (ab)0(ab0) − a0b0(ab) ∈ I.
Thus ab0 ∈ I. This implies that a0b = (ab)0− ab0 ∈ I. Lemma 1.7. Let I be a radical differential ideal in a differential ring A and S be any subset of A. Let
T = {a ∈ A : aS ⊂ I}.
Then T is a radical differential ideal in A.
Proof. It is easy to see that T is an ideal of A. Suppose a ∈ T. We want to show that a0 ∈ I.
Since a ∈ T, ab ∈ I for all b ∈ S. By Lemma 1.6, a0b ∈ I for all b ∈ S. Thus a0S ⊂ I, i.e.
a0∈ T. This shows that T is a differential ideal.
Now, let us prove that T is radical. Let xn∈ T for some x ∈ A and n ∈ N. Hence xns ∈ I for all s ∈ S. Thus (xs)n ∈ I for all s ∈ S. Since I is radical, xs ∈ I for all s ∈ S. By
definition of T, x ∈ T.
In a ring, the intersection of any family of radical ideals is again a radical ideal. By Lemma 1.3, we obtain:
Lemma 1.8. The intersection of any family of differential radical ideals in a differential ring is again a differential radical ideal.
Definition 1.3. Let S be any subset of a differential ring A. The intersection of all differ- ential radical ideals of A containing S is denoted by {S}.
Lemma 1.9. Let S and T be subset of a differential ring A with S ⊂ T. Then {S} ⊂ {T }.
Lemma 1.10. Let a ∈ A and S ⊂ A. Here A is a differential ring. Then a{S} ⊂ {aS}.
Proof. Let I = {aS}. Then I is a differential radical ideal of A. Let T = {x ∈ A : xa ∈ I}.
Then T is a differential radical ideal containing I by Lemma 1.7. Since as ∈ I for all s ∈ S, T contains S. Since T is a differential radical ideal of A containing S, T contains {S}. Hence
a{S} ⊂ I.
Lemma 1.11. Let S, T be any subset of a differential ring A. Then {S}{T } ⊂ {ST }. Here ST = {st : s ∈ S, t ∈ T }.
Proof. Let I = {ST }. Consider J = {x ∈ A : a{T } ⊂ I}. By Lemma 1.10, s{T } ⊂ {sT } for all s ∈ S. By lemma 1.12, {sT } ⊂ {ST } for all s ∈ S. Thus J contains S. Since J is a differential radical ideal of A containing S, it contains {S}. Definition 1.4. A Ritt algebra is a differential algebra with identity over Q.
Lemma 1.12. Let I be a differential ideal in a Ritt algebra A. Let a ∈ A such that an∈ I for some n ∈ N. Then (a0)2n−1 ∈ I.
4
Proof. Since I is a differential ideal, (an)0 = nan−1a0 ∈ I. Since A is a Q-algebra, an−1a0 ∈ I.
Claim an−k(a0)2k−1 ∈ I for all k ≥ 1. We have prove that the statement is true for k = 1.
Assume that the statement is true for k = m, i.e. an−m(a0)2m−1 ∈ I. Differentiating an−m(a0)2m−1, we obtain
(n − m)an−m−1(a0)2m+ (2m − 1)an−m(a0)2m−2a00∈ I.
Multiplying the above equation by a0, we see that
(n − m)an−m−1(a0)2m+1+ (2m − 1)an−m(a0)2m−1a00∈ I.
By induction hypothesis, (2m−1)an−m(a0)2m−1a00∈ I. Thus (n−m)an−m−1(a0)2m+1∈ I. In other words, an−m−1(a0)2m+1∈ I. Thus the statement is true for k = m + 1. By induction, we prove that the statement is true for n ∈ N. Taking m = n, we obtain our result.
Corollary 1.1. In a Ritt algebra, the radical of a differential ideal is a differential ideal.
Proof. Let I be a differential ideal of a Ritt algebra A and √
I be its radical. Let a ∈√ I.
Then an∈ I for some n ≥ 1. By Lemma 1.12, (a0)2n−1 ∈ I. Hence a0 ∈√
I. By definition,
√
I is a differential ideal.