Lie and Jordan structures of differentially semiprime rings

Properties of Lie and Jordan rings (denoted respectively by RL and RJ) associated with an associative ring R are discussed. Results on connections between the differentially simplicity (respectively primeness, semiprimeness) of R, RL and RJ are obtained.

Збережено в:
Бібліографічні деталі
Опубліковано в: :Algebra and Discrete Mathematics
Дата:2015
Автори: Artemovych, O.D., Lukashenko, M.P.
Формат: Стаття
Мова:English
Опубліковано: Інститут прикладної математики і механіки НАН України 2015
Онлайн доступ:https://nasplib.isofts.kiev.ua/handle/123456789/154758
Теги: Додати тег
Немає тегів, Будьте першим, хто поставить тег для цього запису!
Назва журналу:Digital Library of Periodicals of National Academy of Sciences of Ukraine
Цитувати:Lie and Jordan structures of differentially semiprime rings / O.D. Artemovych, M.P. Lukashenko // Algebra and Discrete Mathematics. — 2015. — Vol. 20, № 1. — С. 13-31 . — Бібліогр.: 23 назв. — англ.

Репозитарії

Digital Library of Periodicals of National Academy of Sciences of Ukraine
id nasplib_isofts_kiev_ua-123456789-154758
record_format dspace
spelling Artemovych, O.D.
Lukashenko, M.P.
2019-06-15T20:01:23Z
2019-06-15T20:01:23Z
2015
Lie and Jordan structures of differentially semiprime rings / O.D. Artemovych, M.P. Lukashenko // Algebra and Discrete Mathematics. — 2015. — Vol. 20, № 1. — С. 13-31 . — Бібліогр.: 23 назв. — англ.
1726-3255
2010 MSC:Primary 16W25, 16N60; Secondary 17B60, 17C50.
https://nasplib.isofts.kiev.ua/handle/123456789/154758
Properties of Lie and Jordan rings (denoted respectively by RL and RJ) associated with an associative ring R are discussed. Results on connections between the differentially simplicity (respectively primeness, semiprimeness) of R, RL and RJ are obtained.
en
Інститут прикладної математики і механіки НАН України
Algebra and Discrete Mathematics
Lie and Jordan structures of differentially semiprime rings
Article
published earlier
institution Digital Library of Periodicals of National Academy of Sciences of Ukraine
collection DSpace DC
title Lie and Jordan structures of differentially semiprime rings
spellingShingle Lie and Jordan structures of differentially semiprime rings
Artemovych, O.D.
Lukashenko, M.P.
title_short Lie and Jordan structures of differentially semiprime rings
title_full Lie and Jordan structures of differentially semiprime rings
title_fullStr Lie and Jordan structures of differentially semiprime rings
title_full_unstemmed Lie and Jordan structures of differentially semiprime rings
title_sort lie and jordan structures of differentially semiprime rings
author Artemovych, O.D.
Lukashenko, M.P.
author_facet Artemovych, O.D.
Lukashenko, M.P.
publishDate 2015
language English
container_title Algebra and Discrete Mathematics
publisher Інститут прикладної математики і механіки НАН України
format Article
description Properties of Lie and Jordan rings (denoted respectively by RL and RJ) associated with an associative ring R are discussed. Results on connections between the differentially simplicity (respectively primeness, semiprimeness) of R, RL and RJ are obtained.
issn 1726-3255
url https://nasplib.isofts.kiev.ua/handle/123456789/154758
citation_txt Lie and Jordan structures of differentially semiprime rings / O.D. Artemovych, M.P. Lukashenko // Algebra and Discrete Mathematics. — 2015. — Vol. 20, № 1. — С. 13-31 . — Бібліогр.: 23 назв. — англ.
work_keys_str_mv AT artemovychod lieandjordanstructuresofdifferentiallysemiprimerings
AT lukashenkomp lieandjordanstructuresofdifferentiallysemiprimerings
first_indexed 2025-11-25T22:29:15Z
last_indexed 2025-11-25T22:29:15Z
_version_ 1850563564855623680
fulltext Algebra and Discrete Mathematics RESEARCH ARTICLE Volume 20 (2015). Number 1, pp. 13–31 © Journal “Algebra and Discrete Mathematics” Lie and Jordan structures of differentially semiprime rings Orest D. Artemovych and Maria P. Lukashenko Communicated by A. P. Petravchuk Abstract. Properties of Lie and Jordan rings (denoted respectively by RL and RJ) associated with an associative ring R are discussed. Results on connections between the differentially simplicity (respectively primeness, semiprimeness) of R, RL and RJ are obtained. 1. Introduction Throughout here, R is an associative ring (with respect to the addition “+” and the multiplication “ · ”) with an identity, Der R is the set of all derivations in R. On the set R we consider two operations: the Lie multiplication “[−, −]” and the Jordan multiplication “(−, −)” defined by the rules [a, b] = a · b − b · a and (a, b) = a · b + b · a for any a, b ∈ R. Then RL = (R, +, [−, −]) is a Lie ring and RJ = (R, +, (−, −)) 2010 MSC: Primary 16W25, 16N60; Secondary 17B60, 17C50. Key words and phrases: Derivation, semiprime ring, Lie ring. 14 Lie and Jordan structures is a Jordan ring (see [13] and [14]) associated with the associative ring R. Recall that an additive subgroup A of R is called: • a Lie ideal of R if [a, r] ∈ A, • a Jordan ideal of R if (a, r) ∈ A for all a ∈ A and r ∈ R. Obviously, A is a Lie (respectively Jordan) ideal of R if and only if AL (respectively AJ) is an ideal of RL (respectively RJ). In all that follows ∆ will be any subset of Der R (in particular, ∆ = {0}) and δ ∈ Der R. A subset K of R is called ∆-stable if d(a) ∈ K for all d ∈ ∆ and a ∈ K. An ideal I of a (Lie, Jordan or associative) ring A is said to be a ∆-ideal if I is ∆-stable. A (Lie, Jordan or associative) ring A is said to be: • simple (respectively ∆-simple) if there no two-sided ideals (respec- tively ∆-ideals) other 0 or A, • prime (respectively ∆-prime) if, for all two-sided ideals (respectively ∆-ideals) K, S of A, the condition KS = 0 implies that K = 0 or S = 0 (if ∆ = {δ} and A is ∆-prime, then we say that A is δ-prime), • semiprime (respectively ∆-semiprime) if, for any two-sided ideal (respectively ∆-ideal) K of A, the condition K2 = 0 implies that K = 0, • primary if, for any two-sided ideals K, S of A, the condition KS = 0 implies that K = 0 or S is nilpotent. Every non-commutative ∆-simple ring is ∆-prime and every ∆-prime ring is ∆-semiprime. We say that R is Z-torsion-free if, for any r ∈ R and integers n, the condition nr = 0 holds if and only if r = 0. If the implication 2r = 0 ⇒ r = 0 is true for any r ∈ R, then R is said to be 2-torsion-free. Let Fp(R) = {a ∈ R | a has an additive order pk for some non-negativek = k(a)} be the p-part of R, where p is a prime. Then Fp(R) is a ∆-ideal of R. If R is ∆-semiprime, then pFp(R) = 0. O. D. Artemovych, M. P. Lukashenko 15 In particular, in a ∆-prime ring R it holds Fp(R) = 0 (and so the characteristic char R = 0) or Fp(R) = R (and therefore char R = p). Obviously that the additive group R+ of a ∆-prime ring R is torsion-free if and only if char R = 0. Recall that a ring R is said to be of bounded index m, if m is the least positive integer such that xm = 0 for all nilpotent elements x ∈ R. We say that a ring R satisfies the condition (X) if one of the following holds: (1) R or R/P(R) is Z-torsion-free, where P(R) is the prime radical of R, (2) R is of bounded index m such that an additive order of every nonzero torsion element of R, if any, is strictly larger than m. As noted in [16, p.283], a Z-torsion-free δ-prime ring is semiprime. In this way we prove the following Proposition 1. For a ring R the following hold: (1) if R is a ∆-semiprime ring with the condition (X), then it is semiprime, (2) if R is both semiprime (respectively satisfies the condition (X)) and ∆-prime, then R is prime. Relations between properties of an associative ring R, a Lie ring RL and a Jordan ring RJ was studied by I.N. Herstein and his students (see [7, 8, 11] and bibliography in [9] and [5]); he has obtained, for a ring R of characteristic different from 2, that the simplicity of R implies the simplicity of a Jordan ring RJ [7, Theorem 1], and also that every Lie ideal of a simple Lie ring R is contained in the center Z(R) [7, Theorem 3]. K. McCrimmon [20, Theorem 4] has proved that R is a simple algebra if and only if RJ is a simple Jordan algebra. Our result is the following Theorem 1. For a 2-torsion-free ring R the following statements are true: (1) R is a ∆-simple ring if and only if RJ is a ∆-simple Jordan ring, (2) R is a ∆-prime ring if and only if RJ is a ∆-prime Jordan ring, (3) R is a ∆-semiprime ring if and only if RJ is a ∆-semiprime Jordan ring. 16 Lie and Jordan structures Let us d ∈ ∆. Since C(R) and ann C(R) are ∆-ideals, the rule d : R/ ann C(R) ∋ r + ann C(R) 7→ d(r) + ann C(R) ∈ R/ ann C(R) determines a derivation d of the quotient ring R/ ann C(R). Then ∆ = {d | d ∈ ∆} ⊆ Der(R/ ann C(R)). Inasmuch d(Z(R)) ⊆ Z(R), the rule d̂ : RL/Z(R) ∋ r + Z(R) 7→ d(r) + Z(R) ∈ RL/Z(R) determines a derivation d̂ of the Lie ring RL/Z(R). Then ∆̂ = {d̂ | d ∈ ∆} ⊆ Der(RL/Z(R)). Since the center Z(R) is a nonzero Lie ideal of an associative ring R with an identity, a Lie ring RL is not ∆-simple. Our next result is the following Theorem 2. Let R be a 2-torsion-free ring. Then the following are true: (1) if the quotient ring RL/Z(R) is a ∆̂-simple Lie ring, then R is non-commutative and R/ ann C(R) is a ∆-simple ring, (2) if R is a ∆-simple ring, then RL/Z(R) is a ∆̂-simple Lie ring or R is commutative, (3) if RL/Z(R) is a ∆̂-semiprime Lie ring, then R is non-commutative and the quotient ring R/ ann C(R) is a ∆-semiprime ring, (4) if R is a ∆-semiprime ring, then RL/Z(R) is a ∆̂-semiprime Lie ring or R is commutative, (5) if RL/Z(R) is a ∆̂-prime Lie ring, then R is non-commutative and R/ ann C(R) is a ∆-prime ring, (6) if R is a ∆-prime ring, then RL/Z(R) is a ∆̂-prime Lie ring or R is commutative. Throughout, let Z(R) denote the center of R, [A, B] (respectively (A, B)) an additive subgroup of R generated by all commutators [a, b] (respectively (a, b)), where a ∈ A and b ∈ B, C(R) the commutator ideal of R, N(R) the set of nilpotent elements in R, char R the characteristic of R, annl I = {a ∈ R | aI = 0} the left annihilator of I in R, annr I = {a ∈ R | Ia = 0} the right annihilator of I in R, ann I = (annr I) ∩ (annl I), CR(I) = {a ∈ R | ai = ia for all i ∈ I} the centralizer of I in R and ∂a(x) = [a, x] for a, x ∈ R. All other definitions and facts are standard and it can be found in [10], [17] and [19]. O. D. Artemovych, M. P. Lukashenko 17 2. Differentially prime right Goldie rings Let agree that d0 = idR is the identity endomorphism for d ∈ ∆. Lemma 1. The following conditions are equivalent: (1) R is a ∆-semiprime ring, (2) for any ∆-ideals A, B of R the implication AB = 0 ⇒ A ∩ B = 0 is true, (3) if a ∈ R is such that aRδm1 1 . . . δmk k (a) = 0 for any integers k > 1, mi > 0 and derivations δi ∈ ∆ (i = 1, . . . , k), then a = 0. Proof. A simple modification of Proposition 2 from [17, §3.2]. Lemma 2. The following conditions are equivalent: (1) R is a ∆-prime ring, (2) a left annihilator annl I of a left ∆-ideal I of R is zero, (3) a right annihilator annr I of a right ∆-ideal I of R is zero, (4) if a, b ∈ R are such that aRδm1 1 . . . δmk k (b) = 0 for any integers k > 1, mj > 0 and derivations δj ∈ ∆ (j = 1, . . . , k), then a = 0 or b = 0. Proof. A simple consequence of Lemma 2.1.1 from [10]. If I is an ideal of a ring R, then CR(I) = {x ∈ R | x + I is regular in the quotient ring R/I} (see [19, Chapter 2, §1]). The next lemma extends Proposition 1 of [15]. 18 Lie and Jordan structures Lemma 3. Let R be a right Goldie ring and δ ∈ Der R. If R is δ-prime, then: (a) the set N = N(R) of nilpotent elements of R is its prime radical, (b) ⋂k i=1 δ−1(N) = 0 for some integer k, (c) CR(0) = CR(N). Proof. From Theorem 2.2 of [16] (see the part (ii) ⇒ (iii) of its proof), we obtain (a) and (b). By Proposition 4.1.3 of [19], CR(0) ⊆ CR(N). By the same argument as in [16, p.284], we can obtain that CR(0) = CR(N). Corollary 1. If R is a commutative δ-prime Goldie ring and δ ∈ Der R, then N(R) contains all zero-divisors of R. By Corollary 1.4 of [6], if I is a δ-prime ideal of a right Noetherian ring R and R/I has characteristic 0, then I is prime. The following lemma is an extension of Lemma 2.5 from [6]. Lemma 4. Let R be a 2-torsion-free commutative Goldie ring and δ ∈ Der R. If R is δ-prime, then it is an integral domain. Proof. Assume that a ∈ ann N(R), b ∈ N(R) and r ∈ R. Then 0 = δ2(arb) = δ(δ(a)rb + aδ(r)b + arδ(b)) = δ2(a)rb + 2δ(a)δ(r)b + 2δ(a)rδ(b) + aδ2(r)b + 2aδ(r)δ(b) + arδ2(b) and so 2δ(a)Rδ(b) ⊆ N(R). This means that δ(a) ∈ N(R) or δ(b) ∈ N(R). Hence N(R) is δ-stable. By Lemma 3, N(R) is a ideal and therefore N(R) = 0. By Lemma 1.2 of [4], R is prime and consequently it is an integral domain. Proof of Proposition 1. (1) By Proposition 1.3 of [6] and Theorem 1 of [1], the prime radical P(R) is a ∆-ideal and so P(R) = 0 is zero. (2) Since P(R) = 0, R is prime by Lemma 1.2 from [4]. By Theorem 4 of [22], a ∆-simple ring R of characteristic 0 is prime. Since every non-commutative ∆-simple ring is ∆-prime, in view of Propo- sition 1 we obtain the following Corollary 2. Let R be a semiprime ring (respectively a ring R satisfy the condition (X)). If R is ∆-simple, then it is prime. O. D. Artemovych, M. P. Lukashenko 19 3. Differential analogues of Herstein’s results For the proof of Theorem 2 we need the next results. In the proofs below we use the same consideration, as in [12, Chapter 1, §1], and present them here in order to have the paper more self-contained. Let agree that everywhere in this section k > 1 and mi > 0 are integers (i = 1, . . . , k). Lemma 5. Let R be a ∆-semiprime ring, A and B its ∆-ideals. Then the following statements hold: (i) if AB = 0, then BA = 0. (ii) annl A = annr A. (iii) A ∩ annr A = 0. Proof. (i) Indeed, BA is a ∆-ideal and (BA)2 = 0 and so BA = 0. (ii) We denote (annr A)A by X. Since X is a ∆-ideal and X2 = 0, we deduce that X = 0. This means that annr A ⊆ annl A. The inverse inclusion we can prove similarly. (iii) Since A ∩ annr A is a nilpotent ∆-ideal, the assertion holds. Henceforth Xa = {[δm1 1 . . . δmk k (a), x] | x ∈ R, δi ∈ ∆, mi > 0 and k > 1 are integers (i = 1, . . . , k)}. It is clear that [a, x] ∈ Xa. Lemma 6. Let R be a ∆-semiprime ring and a ∈ R. Then the following statements hold: (i) if a[δm1 1 . . . δmk k (a), R] = 0 for any integers k > 1, mi > 0 and derivations δi ∈ ∆ (i = 1, . . . , k), then a ∈ Z(R), (ii) if I is a right ∆-ideal of R, then Z(I) ⊆ Z(R), (iii) if I is a commutative right ∆-ideal of R and I is nonzero, then I ⊆ Z(R). If, moreover, R is ∆-prime, then it is commutative. 20 Lie and Jordan structures Proof. (i) Let x, y ∈ R and d, δ ∈ ∆. Since [b, xy] = [b, x]y + x[b, y] (3.1) for any b ∈ Xa and a[b, xy] = 0, we conclude that ax[b, y] = 0. This gives that ayx[b, y] = 0 and yax[b, y] = 0 and consequently (R[a, y]R)2 = 0. (3.2) In addition, 0 = d(a[b, x]) = d(a)[b, x]. Multiplying (3.1) by d(a) on left we get d(a)x[b, y] = 0. Moreover, 0 = δ(ax[d(b), y]) = δ(a)x[d(b), y] and, by the similar argument, we obtain that δm1 1 . . . δmk k (a)x[δm1 1 . . . δmk k (a), y] = 0 for any integers k > 1, mi > 0 and derivations δi ∈ ∆ (i = 1, . . . , k). As in the proof of the condition (3.2), we deduce that (R[δm1 1 . . . δmk k (a), y]R)2 = 0. Then I = ∞∑ k=1 ∑ δ1...δk∈∆ y∈R R[δm1 1 . . . δmk k (a), y]R is a sum of nilpotent ideals and therefore it is a nil ideal. Since I is a ∆-ideal, we conclude that I = 0 and, as a consequence, a ∈ Z(R). (ii) Let a ∈ Z(I) and y ∈ R. Then, for δ1, . . . , δk ∈ ∆, we have δm1 1 . . . δmk k (a) ∈ Z(I) and ay ∈ I. This gives that a(δm1 1 . . . δmk k (a)y) = δm1 1 . . . δmk k (a)(ay) = a(yδm1 1 . . . δmk k (a)), and thus a[δm1 1 . . . δmk k (a), y] = 0. By (i), a ∈ Z(R) is central. O. D. Artemovych, M. P. Lukashenko 21 (iii) By (ii), I ⊆ Z(R). Assume that R is ∆-prime, u, v ∈ R and a ∈ I. Then au ∈ I and so au ∈ Z(R). Since a(uv) = (au)v = v(au) = (va)u = a(vu), we see that [u, v] ∈ annr I. By Lemma 2(3), [u, v] = 0 and hence R is commutative. Lemma 7. Let R be a ∆-prime ring and a ∈ R. If a ∈ CR(I) for some nonzero right ∆-ideal I of R, then a ∈ Z(R). Proof. Let us y ∈ R and b ∈ I. Then by ∈ I and so bay = a(by) = bya. This yields that I[a, y] = 0 = [a, y]I. By Lemma 2(3), [a, y] = 0. Hence a ∈ Z(R). Lemma 8. The left annihilator annl(Xa) is a left ∆-ideal of R. Proof. Immediate from the definition. Lemma 9. If R is a ∆-semiprime ring, then CR([R, R]) ⊆ Z(R). Proof. Let us a ∈ CR([R, R]), d, δ ∈ ∆ and x, y ∈ R. Putting x for a and xd(a) for xy in (3.1) we obtain [x, xd(a)] = [x, x]d(a) + x[x, d(a)] and, as a consequence, [a, x[x, d(a)]] = 0 and [a, x][x, d(a)] = 0. Then, by the same reasons as in the proof of Lemma 6(i), we obtain that [a, x] ∈ annl(Xa) and A = annl(Xa) is a ∆-ideal. Then [δ(a), x][d(a), x] = δ([a, x][d(a), x]) = 0. Since A ∩ annl A = 0, we deduce that is a nilpotent ∆-ideal and so a ∈ Z(R). Lemma 10. Let R be a 2-torsion-free ∆-semiprime ring. If a ∈ R commutes with all elements of Xa, then a ∈ Z(R). 22 Lie and Jordan structures Proof. Let r, x, y ∈ R and d ∈ ∆. It is clear that ∂2 a(x) = 0. From ∂2 a(xy) = 0 it follows that 2∂a(x)∂a(y) = 0 and so ∂a(x)∂a(y) = 0. Since 0 = ∂a(x)∂a(rx) = ∂a(x)∂a(r)x + ∂a(x)r∂a(x) = ∂a(x)r∂a(x), we deduce that ∂a(x)R∂a(x) = 0 and (∂a(x)R)2 = 0. Moreover, a[b, x] = [b, x]a for any [b, x] ∈ Xa and therefore d(a)[b, x] + a[d(b), x] + a[b, d(x)] = [b, x]d(a) + [d(b), x]a + [b, d(x)]a. From this it holds that d(a)[b, x] = [b, x]d(a). This means that CR(Xa) is ∆-stable and (∂d(a)(x)R)2 = 0. As a conse- quence, I = ∞∑ k=1 ∑ x∈R mk>0 δ1,...,δk∈∆ ∂δ m1 1 ...δ mk k (a)(x)R is a sum of nilpotent ideals and so I is a nil ideal. Since I is a ∆-ideal, we deduce that I = 0. Hence a ∈ Z(R). The next lemma is an extension of Lemma 1 from [11] in the differential case. Lemma 11. Let R be a 2-torsion-free ∆-semiprime ring, T its Lie ∆- ideal. If [T, T ] ⊆ Z(R), then T ⊆ Z(R). Proof. Let x ∈ R and t ∈ T . 1) If [T, T ] = 0, then [t, x] ∈ T and so [t, [t, x]] = 0. By Lemma 10, T ⊆ Z(R). 2) Now assume that 0 6= [a, b] ∈ [T, T ] for some a, b ∈ T . Then ∂a(b) ∈ Z(R) and ∂2 a(R) ⊆ Z(R). Moreover, we have that Z(R) ∋ ∂2 a(bx) = ∂a(∂a(b)x + b∂a(x)) = ∂2 a(b)x + 2∂a(b)∂a(x) + b∂2 a(x) = 2∂a(b)∂a(x) + b∂2 a(x) O. D. Artemovych, M. P. Lukashenko 23 and hence [2∂a(b)∂a(x) + b∂2 a(x), b] = 0. Then 0 = 2∂b(∂a(b))∂a(x) + 2∂a(b)∂b(∂a(x)) + ∂b(b)∂2 a(x) + b∂b(∂ 2 a(x)) = 2∂a(b)∂b(∂a(x)) (3.3) and ∂a(ba) = ∂a(b)a + b∂a(a) = ∂a(b)a. Replacing ba for x in (3.3) we have 0 = 2∂a(b)∂b(∂a(b)a) = 2∂a(b)(∂b(∂a(b)) + ∂a(b)∂b(a)) = −2∂a(b)3 and thus ∂a(b)3 = 0. Then R∂a(b) is a nilpotent ideal in R and, as a consequence, ∑ a,b∈T R∂a(b) is a nonzero nil ∆-ideal, a contradiction. Lemma 12. If U is a Lie ∆-ideal of a ring R and I(U) = {u ∈ R | uR ⊆ U}, then I(U) is the largest ∆-ideal of R such that I(U) ⊆ U . Proof. Let u, v ∈ I(U), x, y ∈ R and δ ∈ ∆. Clearly that I(U) is an additive subgroup of R, I(U) ⊆ U and (ux)y = u(xy) ∈ (ux)R = u(xR) ⊆ uR ⊆ U that is ux ∈ I(U). From u(xy) − (yu)x = (ux)y − y(ux) = [ux, y] ∈ U (and so (yu)x ∈ U) it holds that yu ∈ I(U). Hence U is a two-sided ideal of R. Moreover, δ(u)x + uδ(x) = δ(ux) ∈ δ(U) ⊆ U and uδ(x) ∈ uR ⊆ U . Therefore δ(u)x ∈ U . This means that I(U) is a ∆- ideal of R. If A is a ∆-ideal of R that is contained in U , then AR ⊆ A ⊆ U and hence A ⊆ I(U). Lemma 13. Let U be a Lie ∆-ideal of R. If U is an associative subring of R, then [U, U ] = 0 or U contains a nonzero ∆-ideal of R. 24 Lie and Jordan structures Proof. Assume that x ∈ R and [U, U ] 6= 0. Then [u, v] 6= 0 for some u, v ∈ U and [u, vx] = u(vx) − (vx)u = (uv − vu)x + v(ux − xu). Since [u, x], [u, vx] ∈ U and v[u, x] ∈ U , we deduce that [u, v]x ∈ U . This means that [u, v] ∈ I(U). In view of Lemma 12, I(U) is a nonzero ∆-ideal of R that is contained in U . Proposition 2. If U is a Lie ∆-ideal of R, then [U, U ] = 0 or there exists a nonzero ∆-ideal IU of R such that [IU , R] ⊆ U . Proof. By Lemma 3 of [7], T (U) = {t ∈ R | [t, R] ⊆ U} is both a Lie ideal and an associative subring of R and U ⊆ T (U). Moreover, for δ ∈ ∆, we have [δ(t), R] + [t, δ(R)] = δ([t, R]) ⊆ δ(U) ⊆ U and so [δ(t), R] ⊆ U . Hence T (U) is ∆-stable. If [U, U ] 6= 0, then, by Lemmas 12 and 13, IU = I(T (U)) ⊆ T (U) is a nonzero ∆-ideal of R such that [IU , R] ⊆ U . Lemma 14. Let U be a Lie ∆-ideal of a ring R. If [U, U ] = 0, then the centralizer CR(U) is a Lie ∆-ideal and an associative subring of R. Proof. Is immediately. We extend Theorem 1.3 of [9] in the following Proposition 3. Let R be a ∆-simple ring of characteristic 2. If U is a Lie ∆-ideal of R, then one of the following holds: (1) [R, R] ⊆ U , (2) U ⊆ Z(R), (3) R contains a subfield P such that U ⊆ P and [P, R] ⊆ P . O. D. Artemovych, M. P. Lukashenko 25 Proof. If [U, U ] 6= 0, then [R, R] ⊆ U by Proposition 2. Therefore we assume that [U, U ] = 0. By Lemma 14, CR(U) is a Lie ∆-ideal and an associative subring of R such that U ⊆ CR(U). a) If CR(U) is non-commutative, then CR(U) = R by Lemma 13. Hence U ⊆ Z(R). b) Now assume that the centralizer CR(U) is commutative. If c ∈ CR(U) and x ∈ R, then c2 ∈ CR(U) and [c2, x] = [[c, x], x] = 2c[c, x] = 0. This gives that c2 ∈ Z(R). By Theorem 2 of [22], Z(R) is a field. As a consequence, c2 (and so c) is invertible in CR(U). Hence CR(U) is a field. Corollary 3. Let R be a ∆-simple ring. If U is a Lie ∆-ideal of R, then one of the following holds: (1) [R, R] ⊆ U , (2) U ⊆ Z(R), (3) char R = 2 and R contains a subfield P such that U ⊆ P and [P, R] ⊆ P . 4. Jordan properties Lemma 15. Let R be a ∆-simple ring of characteristic 6= 2, U its proper Jordan ∆-ideal and a ∈ U . If [a, R] ⊆ U , then a = 0. Proof. Let us x, y ∈ R. Since [a, x] ∈ U and (a, x) ∈ U , we obtain that 2ax ∈ U and, as a consequence, ax ∈ U and (ax, y) ∈ U . Moreover, from axy ∈ U it follows that yax ∈ U . This means that RaR ⊆ U . Since d(a) ∈ U for any d ∈ ∆, in view of [21, Lemma 1.1] we obtain that ∞∑ k=1 ∑ δ1,...,δk∈∆ (m1,...,mk)∈N k Rδm1 1 . . . δmk k (a)R is a proper ∆-ideal of R that is contained in U . Hence a = 0. Remark 1. Let R be a 2-torsion-free ring, U its Jordan ∆-ideal. If ∆ contains all inner derivations of R, then U is an ideal of R. 26 Lie and Jordan structures In fact, we have 2xa = [a, x] + (a, x) ∈ U for any a, b, x ∈ U and so xa ∈ U . By the same argument, we can conclude that ax ∈ U . Proof of Theorem 1. (1) (⇐) If A is a nonzero proper ∆-ideal of a ring R, then AJ is a nonzero proper ∆-ideal of RJ , a contradiction. (⇒) Let U be a proper Jordan ∆-ideal of R, a, b ∈ U and x ∈ R. By Lemma 1 of [7], [(a, b), x] ∈ U , and, by Lemma 15, we see that (a, b) = 0. (4.4) In particular, 2a2 = 0 and, as a consequence, a2 = 0 and 2axa = (a, (a, x)) = 0. It follows that axa = 0. Since 0 = (a + b)x(a + b) = axb + bxa and 0 = (b, (a, x)) = b(ax + xa) + (ax + xa)b = bax + bxa + axb + xba, we deduce that bax + xab = 0. But ab = −ba and so bax − xba = 0. This means that ba ∈ Z(R). Then (RabR)2 = 0. Since I = ∞∑ k=1 ∑ a,b∈U, δ1,...,δk∈∆ (m1,...,mk)∈N k Raδm1 1 . . . δmk k (b)R is a ∆-ideal of R that is a sum of nilpotent ideals, we obtain that I = 0. Therefore 0 = (b, x)a = (bx + xb)a = bxa + xba = 2bxa. We conclude that URU = 0. From (RUR)2 = 0 and δ(RUR) ⊆ RUR for any δ ∈ ∆ it holds that U = 0. (2) (⇐) If A, B are ∆-ideals of R such that AB = 0, then (BA)2 = 0 and so BA is a Jordan ideal of R satisfying the condition (BA, BA) = 0. O. D. Artemovych, M. P. Lukashenko 27 Thus the condition (4.4) is true for U = BA. As in the proof of the part (1), we obtain that BA = 0. Then AJ , BJ are ∆-ideals of a Jordan ring RJ such that (AJ , BJ) = 0. Hence A = 0 or B = 0. (⇒) Let a1, a2 ∈ A and x, y ∈ R. Suppose that RJ is not ∆-prime and therefore there exist nonzero Jordan ∆-ideals A, B of R such that (A, B) = 0. By the same reasons as above, we conclude that A ∩ B = 0. Then, by Lemma 1 of [7], we have [(a1, a2), x] ∈ A and hence [(a1, a2), x] ± ((a1, a2), x) ∈ A. Therefore x(a1, a2)y ∈ A. Thus R contains ∆-ideals R(A, A)R ⊆ A and R(B, B)R ⊆ B such that R(A, A)R(B, B)R ⊆ A ∩ B = 0. Hence (A, A) = 0 or (B, B) = 0 and this leads to a contradiction. (3) (⇐) If A is a nonzero ∆-ideal of R such that A2 = 0, then AJ is a nonzero ∆-ideal of the Jordan ring RJ such that (AJ , AJ) = 0, a contradiction. (⇒) Suppose that R has a nonzero Jordan ∆-ideal U such that (U, U) = 0. Then the condition (4.4) is true for any a, b ∈ U . As in the proof of the part (1), we obtain that U = 0. � If R is a ring, then on the set R we can to define a left Jordan multiplication “〈−, −〉” by the rule 〈a, b〉 = 2ab for any a, b ∈ R. Then the equalities 〈〈〈a, a〉, b〉, a〉 = 〈〈a, a〉, 〈b, a〉〉 and 〈〈a, b〉, a〉 = 〈a, 〈b, a〉〉 28 Lie and Jordan structures are true and hence RlJ = (R, +, 〈−, −〉) is a non-commutative Jordan ring (which is called a left Jordan ring associated with an associative ring R). It is clear that, for commutative ring R, we have RJ = RlJ . If A is an additive subgroup of R that 〈a, r〉, 〈r, a〉 ∈ A for any a ∈ A and r ∈ R, then A is called an ideal of RlJ . If δ ∈ ∆ and a, b ∈ R, then δ(〈a, b〉) = δ(2ab) = 2δ(a)b + 2aδ(b) = 〈δ(a), b〉 + 〈a, δ(b)〉 and therefore δ ∈ Der(RlJ). By the other hand, if δ ∈ Der(RlJ), then 2δ(ab) = δ(〈a, b〉) = 〈δ(a), b〉 + 〈a, δ(b)〉 = 2(δ(a)b + aδ(b)). If R is a 2-torsion-free ring, then δ ∈ Der R. Similarly, as in Theorem 1, we can prove the following Proposition 4. For a 2-torsion-free ring R the following conditions are true: (1) R is a ∆-simple ring if and only if RlJ is a ∆-simple Jordan ring, (2) R is a ∆-prime ring if and only if RlJ is a ∆-prime Jordan ring, (3) R is a ∆-semiprime ring if and only if RlJ is a ∆-semiprime Jordan ring. 5. Proofs The next lemma in the prime case is contained in [18, Lemma 7]. Lemma 16 ([2, Lemma 1.7]). Let R be a ring. If [[R, R], [R, R]] = 0, then the commutator ideal C(R) is nil. Corollary 4. If R is a non-commutative ∆-semiprime ring, then [R, R] is non-commutative. Proof of Theorem 2. (1) It is clear that a ring R is non-commutative. If A is a nonzero proper ∆-ideal of R, then AL is a nonzero proper ∆-ideal of RL. Therefore A ⊆ Z(R) and, as a consequence, A · C(R) = 0. O. D. Artemovych, M. P. Lukashenko 29 (2) Suppose that a ∆-simple ring R is non-commutative and U is its nonzero proper Lie ∆-ideal. By Proposition 2, [U, U ] = 0. Then, by Lemma 11, U ⊆ Z(R). Hence the quotient ring RL/Z(R) is ∆̂-simple. (3) Let A be a nonzero ∆-ideal of R such that A2 = 0. Then AL is a nonzero ∆-ideal of a Lie ring RL and, moreover, [AL, AL] = 0. By Lemma 11, A ⊆ Z(R) and hence A · C(R) = 0. (4) Suppose that R is non-commutative. Let A be a nonzero Lie ∆-ideal of R such that [A, A] = 0. Then, by Lemma 11, A ⊆ Z(R) and, as a consequence, the Lie ring RL/Z(R) is ∆̂-semiprime. (5) Let A, B be nonzero ∆-ideals of R such that AB = 0. Obviously, [A, B] ⊆ Z(R). Then A ⊆ Z(R) or B ⊆ Z(R). (6) Assume that R is non-commutative and A, B are nonzero Lie ∆-ideals of R such that [A, B] = 0. Then A ∩ B ⊆ Z(R). Since A ∩ B ⊆ ann C(R) in a ∆-prime ring R, we have that the intersection A ∩ B = 0 is zero. If T (A) = R (see proof of Proposition 2), then [R, R] ⊆ A and B ⊆ CR([R, R]). By Lemma 9, B ⊆ Z(R). So we assume that T (A) 6= R. If [T (A), T (A)] = 0, then [A, A] = 0 and, by Lemma 11, A ⊆ Z(R). Suppose that [T (A), T (A)] 6= 0. By Lemma 13, T (A) contains a nonzero ∆-ideal I of R. Since [I, B] ⊆ A ∩ B = 0, we conclude that B ⊆ Z(R) by Lemma 7. The map ∂a : R ∋ x 7→ [a, x] ∈ R is called an inner derivation of a ring R induced by a ∈ R. The set IDer R of all inner derivations of R is a Lie ring. Every prime Lie ring is primary Lie. Lemma 17. There is the Lie ring isomorphism IDer R ∋ ∂a 7→ a + Z(R) ∈ RL/Z(R). Proof. Evident. Corollary 5. Let R be a ring. Then the following statements hold: 30 Lie and Jordan structures (1) IDer R is a simple Lie ring if and only if RL/Z(R) is a simple Lie ring, (2) IDer R is a prime Lie ring if and only if RL/Z(R) is a prime Lie ring, (3) IDer R is a semiprime Lie ring if and only if RL/Z(R) is a semipri- me Lie ring, (4) IDer R is a primary Lie ring if and only if RL/Z(R) is a primary Lie ring. References [1] K. I. Beidar, A. V. Mikhaĺ’ev, Ortogonal completeness and minimal prime ideals (in Russian), Trudy Sem. Petrovski, 10, 1984, pp.227-234. [2] H. E. Bell, A. A. Klein, Combinatorial commutativity and finiteness conditions for rings, Comm. Algebra, 29, 2001, pp.2935-2943. [3] J. Bergen, I. N. Herstein, E. W. Kerr, Lie ideals and derivations of prime rings, J. Algebra, 71, 1981, pp.259-267. [4] J. Bergen, S. Montgomery, D. S. Passmann, Radicals of crossed products of en- veloping algebras, Israel J. Math., 59, 1987, pp.167-184. [5] M. Brešar, M. A. Chebotar and W.S. Martindale, 3rd, Functional identities, Birkh́’ auser Verlag, Basel Boston Berlin, 2000. [6] K. R. Goodearl, R. B. Warfield, Jr., Primitivity in differential operator rings, Math. Z., 180, 1982, pp.503-523. [7] I. N. Herstein, On the Lie and Jordan rings of a simple associative ring, Amer. J. Math., 77, 1955, pp.279-285. [8] I. N. Herstein, The Lie ring of a simple associative ring, Duke Math. J., 22, 1955, pp.471-476. [9] I. N. Herstein, Topics in Ring Theory, The University of Chicago Press, Chicago London, 1965. [10] I. N. Herstein, Noncommutative rings, The Mathematical Association of America, J. Wiley and Sons, 1968. [11] I. N. Herstein, On the Lie structure of an associative ring, J. Algebra, 14, 1970, pp.561-571. [12] I. N. Herstein, Rings with involution, The University of Chicago Press, Chicago London 1976. [13] N. Jacobson, Lie algebras, Interscience, New York, 1962. [14] N. Jacobson, Structure and representations of Jordan algebras, Amer. Math. Soc. Colloq. Publ., V. 39, Providence, R.I., 1968. [15] C. R. Jordan, D. A. Jordan, The Lie structure of a commutative ring with derivation, J. London Math. Soc.(2), 18, 1978, pp.39-49. [16] D. A. Jordan, Noetherian Ore extensions and Jacobson rings, J. London Math. Soc. (2), 10, 1975, pp.281-291. O. D. Artemovych, M. P. Lukashenko 31 [17] J. Lambek, Lectures on rings and modules, Blaisdell Publ. Co. A division of Ginn and Co. Waltham Mass. Toronto London, 1966. [18] W. S. Martindale, 3rd, Lie isomorphisms of prime rings, Trans. Amer. Math. Soc., 142, 1969, pp.437-455. [19] J. C. McConnell, J. C. Robson, Noncommutative Noetherian rings. With the cooperation of L. W. Small. Revised edition. Grad. Stud. in Math., 30. Amer. Math. Soc., Providence, RI, 2001. [20] K. McCrimmon, On Herstein’s theorem relating Jordan and associative algebras, J. Algebra, 13, 1969, pp.382-392. [21] A. Nowicki, The Lie structure of a commutative ring with a derivation, Arch. Math., 45, 1985, pp.328-335. [22] E. C. Posner, Differentiably simple rings, Proc. Amer. Math. Soc., 11, 1960, pp.337-343. [23] I. I. Zuev, Lie ideals of associative rings (in Russian), Uspehi Mat. Nauk, 18, 1963, pp.155–158. Contact information O. D. Artemovych Institute of Mathematics Cracow University of Technology ul. Warszawska 24 Cracow 31-155 POLAND E-Mail(s): artemo@usk.pk.edu.pl M. P. Lukashenko Faculty of Mathematics and Informatics PreCarpathian National University of Vasyl Ste- fanyk Shevchenko St 57 Ivano-Frankivsk 76025 UKRAINE E-Mail(s): bilochka.90@mail.ru Received by the editors: 22.01.2015 and in final form 22.03.2015.