On Commutativity of Completely Prime Gamma-Rings

Size: px

Start display at page:

Download "On Commutativity of Completely Prime Gamma-Rings"

Maud Brooks
5 years ago
Views:

1 alaysian Journal of athematical Sciences 7(2): (2013) ALAYSIAN JOURNAL OF ATHEATICAL SCIENCES Journal homepage: 1,2 I. S. Rakhimov, 3* Kalyan Kumar Dey and 3 Akhil Chandra Paul 1 Department of athematics, Faculty of Science, Universiti Putra alaysia, UP Serdang, Selangor, alaysia 2 Institute for athematical Research, Universiti Putra alaysia, UP Serdang, Selangor, alaysia 3 Department of athematics, Rajshahi University, Rajshahi-6205, Bangladesh risamiddin@gmail.com, kkdmath@yahoo.com and acpaulrubd_math@yahoo.com *Corresponding author ABSTRACT In this paper we prove that any completely prime Γ -ring satisfying the condition a bc c ( a, b, c and, Γ ) with nonzero derivation, is a commutative integral Γ -domain if its characteristic is not two. We also show that if the characteristic of is 2 the Γ -ring is either commutative or is an order in a simple 4-dimensional algebra over its center. We give necessary condition in terms of derivations for belongings of an element of the Γ -ring to the center of when the characteristic of is not two. If char = 2, and a Z( ), then we show that the derivation is the inner derivation. Keywords: Γ -ring, completely prime Γ -ring, centroid of Γ -rings, annihilator, extended centroid of Γ -rings. 1. INTRODUCTION The notion of a gamma ring was introduced as an extensive generalization of the concept of a classical ring. From its first appearance, the extensions and the generalizations of various important results in the theory of classical rings to the theory of gamma rings have attracted a wider attention as an emerging field of research to the modern algebraists to

2 Isamiddin S. Rakhimov, Kalyan Kumar Dey & Akhil Chandra Paul enrich the world of algebra. A good number of prominent mathematicians have worked out on this interesting area of research to develop many basic characterizations of gamma rings and have extended numerous significant results in this context in the last few decades. We start with the following necessary introductory definitions and examples. Definition 1.1. Let and Γ be additive abelian groups. If there exists a tri-additive mapping Γ : ( x,, y) x y satisfying the condition ( ab) c = a ( bc), for all a, b, c and Γ, then is said to be a Γ -ring. Examples. 1. Every associative ring is a Γ -ring. In this case Γ is a single element set. 2. Let R be an integral domain with the identity element 1. Take ( R) n 1 = 1 2 and Γ = n is an integer. Then is a Γ -ring. If 0 we assume that N = {( a, a) : a R}, then it is easy to verify that N also is a Γ -ring (in fact, N is a subring of ). 3. Let be a Γ -ring. We denote the sets of m n matrices with entries from and set of n m matrices with entries from Γ by m n and Γ n m, respectively, then m n is a Γ n m ring with the multiplication defined by ( a )( γ )( b ) = ( c ), where c a γ b. ij ij ij ij If m = n, then n is a Γn -ring. = ij ip pq qj p q Definition 1.2. Let be a Γ -ring. A subring I of is an additive subgroup which is also a Γ -ring. A right ideal of is a subgroup I such that IΓ I. The concept of a left ideal is defined similarly. If I is both the right and the left ideal then we say that I is an ideal. Definition 1.3. The ring is called a prime Γ -ring if aγ Γ b = 0 implies a = 0 or b = 0. The ring is called a completely prime Γ-ring if aγ b = 0 implies a = 0 or b = alaysian Journal of athematical Sciences

3 Let be a Γ -ring and N = {( a, a): a R}, then it is easy to verify that N is also a Γ -ring (in fact, N is a subring of ). It is clear that N is a completely prime Γ -ring satisfying the condition a bc c for all a, b, c N and, Γ. It is easy to see that every completely prime Γ -ring is a prime Γ - ring. Let be a Γ -ring and I be a subset of. The subset Ann ( I ) = { a aγ I = 0} of is called a left annihilator of I. A right l annihilator Ann ( I ) is defined similarly. If I is a nonzero ideal of then r Ann ( I) = Ann ( I) and we denote it by Ann( I ). l r Definition 1.4. Let be a Γ -ring. An additive mapping d : is called a derivation if d( a b) = d( a) b + a d( b) holds for every a, b and Γ. We write [ a, b] = a b b a (the commutator of a and b ). Let a. Define d : by d( x) = [ a, x] for every Γ. Then it is clear that d is a derivation. This derivation is called an inner derivation. We consider an assumption ( )ab c c and, Γ. By the definition of the commutator it is easy to see that the following basic commutator identities hold [ ab, c] = a[ b, c] + [ a, c] b + a[, ] c b [ a, bc] = b[ a, c] + [ a, b] c + b[, ] c. Taking into account the above assumption ( ) the basic commutator identities can be reduced to [ ab, c] = a[ b, c] + [ a, c] b alaysian Journal of athematical Sciences 285 a and [ a, bc] = b[ a, c] + [ a, b] c, respectively. In other words, the later shows that the left and right multiplication operators are inner derivations. This fact we extensively use in this paper. We refer to Ozturk and Jun (2000, 2001) for the definitions of centroid, extended centroid of Γ -rings. The Γ -rings were first introduced by Nobusawa (1964). These are generalizations of classical rings. Later, Barnes (1966) generalized the definition of Nobusawa (1964) which is given above. any researchers worked on Γ -rings and their derivations. The importance of study of derivations was noticed by Herstein (1978,

4 Isamiddin S. Rakhimov, Kalyan Kumar Dey & Akhil Chandra Paul 1979). The end 1990 th is considered the renaissance of the Γ -ring s theory. Ceven (2002) investigated the Jordan left derivations of completely prime Γ -rings and showed that every Jordan left derivation of a 2 torsion-free completely prime Γ -rings is a Jordan left derivation. The notion of k - derivation first is introduced by Kandamar (2000) and he proved that every prime Γ -ring of Nobusawa (1964) is commutative by the help of k - derivations. Chakraborty and Paul (2010) proved that every Jordan k -left derivations of a 2-torsion free completely prime Γ -rings is a k -left derivation. Sapanci and Nakajima (1997) showed that every Jordan derivation of a 2-forsion free prime Γ -rings is a derivation. The commutativity in prime Γ -rings was investigated with the help of two derivations by Soyturk (1994). He proved that is commutative when the two non-zero derivations are contained in the center of. We remind that commutativity conditions for Γ -near rings with one and two derivations have been considered in Rakhimov (2013a) and Rakhimov (2013b). In this paper, we prove that a completely prime Γ -ring with a nonzero derivation is either a commutative integral Γ -domain or is an order in a simple algebra of dimension four over its center, when char 2 or char = 2, respectively. We also prove the following (i) If a and [ a, d( x)] = 0 for all x and Γ, then Z( ), provided char 2, where Z( ) is the center of. (ii) If is of characteristic 2, then it is proven that a a is an element of the center of for all Γ. oreover, if a belongs to the center of, then there is an element c of the centroid of and such that d( x) = [ c a, x], that is d is an inner derivation of. 2. COUTATIVITY OF Γ-RINGS WITH DERIVATIONS. In this section first we prove a few subsidiary lemmas (Lemmas ) to use them in the proving of the main results (Theorems ). Lemma 2.1 Let be a prime Γ -ring and suppose that a centralizes a non-zero right ideal of. Then a Z( ). Proof. Suppose that a centralizes a non-zero right ideal A of. If x, r A, then r x A for every Γ, hence a ( r x) = r x a, for 286 alaysian Journal of athematical Sciences

5 , Γ. But a r = r a, for Γ, we thus get that r ( a x xa) = 0, which is to say that a [ a, x] ) = 0 for all x and, Γ. Since is prime and A 0, we conclude that [ a, x] = 0 for all x, Γ, hence a Z( ). Lemma 2.2 Let be any Γ -ring satisfying the condition a bc c and, Γ and let u. Then the set V = { a : a[ u, x] = 0, for all x and, Γ } is an ideal of. Proof. It is clear that V is a left ideal of. Now, we show that V also is the right ideal. Let a V and x, r. For all,, δ Γ, we have a[ u( rδ x) ( rδ x) u] = 0. The Jacobi identity for the commutators gives, a[ urδ x rδ xu] = ( ur ru) δ x + r ( uδ x xδ u), then using the condition, we get, 0 = a [ u ( rδ x) ( rδ x) u) = a[ u, r] x + a r[ u, x], that is, a r [ u, x] δ = 0, for any,, δ Γ. Hence a r V and V is a right ideal of. Lemma 2.3 Let be a prime Γ -ring satisfying the condition a bc c and, Γ and suppose that 0 u satisfies a [ u, x] = 0, for all x,, Γ. Then u Z( ). Proof. By Lemma 2.2, V = { a : a[ u, x] = 0, for all x and, Γ } is a nontrivial ideal of. Since, is prime and u x x u AnnrV we have u x x u = 0, for all x, Γ, hence u Z( ). Lemma 2.4 Let be a semiprime Γ -ring satisfying the condition a bc c and, Γ and suppose that a centralizes all the commutators [ u, x], for all x, y and Γ. Then u Z( ). δ δ alaysian Journal of athematical Sciences 287

6 Isamiddin S. Rakhimov, Kalyan Kumar Dey & Akhil Chandra Paul Proof. If x, y and, Γ, then since x ( ya) ( ya) x is a commutator, a must commute with x ( ya) ( ya) x. But x ( y a) ( y a) x = [ x, y] a + y[ x, y]. By the hypothesis, a commutes with the left sides and the first term of the right sides of the last relations. Thus the element a must commute with y [ x, a] for all x, y and, Γ. This gives [ y, a] δ [ x, a] = 0, for all x, y and,, δ Γ. This implies that [ y, a] V for all y and Γ. Now due to Lemma 2.2, V is an ideal of and [ y, a] Ann ( V ), hence in r [ y, a] V Annr ( V ). But is semiprime therefore V Annr ( V ) = 0, hence [ y, a] 0, for all y Γ. So we conclude that u Z( ). Theorem 2.5 Let be a Γ -ring, d be a nonzero derivation of such 3 that d 0. Then the subring A of generated by all d( a b), with Γ and a, b, contains a nonzero ideal of. Proof. Since 3 d 0 and d( ) Γ A we have d 2 ( Γ ) 0. Take y A such that d 2 ( y) 0. If x, then d( x y) = d( x) y + x d( y) A. Since y and d( x ) are in A, then x d( y) A. This implies that Γd( y) A. Similarly d( y) Γ A. If r, s, then d( r d( y) s) A. But 2 d( r d( y) s) = d( r) d( y) s) + r d ( y) s + r d( y) d( s). Since d( y) s A and r d( y) A this implies that r, s and, Γ. Hence we get Γd 2 ( y) A, r d 2 ( y) s A for all d 2 ( y) Γ A and Γd 2 ( y) Γ A. Therefore we conclude that the ideal of generated by d 2 ( y) 0 must be in A. This proves the theorem. 3 If d = 0, the result need not be true. Let be any completely prime Γ - ring having nilpotent elements, and let a 0 be such that a a = 0 for every Γ. Let d : be defined by d( x) = [ a, x] for all Γ. Then B = aγ Γ a is a subring of (since a a = 0 ) and contains 3 2 d( ). Also d = 0, d 0 (if char of 2 ). Yet B contains no nonzero ideal of, for 288 alaysian Journal of athematical Sciences

7 aγbγ a = aγ( aγ Γa) Γ a = aγaγ Γ a + aγγaγ a = 0. Theorem 2.6 Let be a completely prime Γ -ring satisfying the condition a bc c and, Γ. Let d be a nonzero derivation of such that [ d( x), d( y)] = 0 for all Γ and x, y. Then (i) If characteristic of 2 then is a commutative integral Γ - domain. (ii) If characteristic of = 2 then either is commutative or is an order of 4-dimentional simple algebra over its center. Proof. Let A be the subring of generated by all d( x ) for x. By the hypothesis the A is a commutative subring of. If a A and x then d( a) x + a d( x) = d( a x) A, hence centralizes A. Therefore, if b A then 0 = b d( a x) d( a x) b = d( a) ( b x xb) = d( a) [ b, x]. If A Z, we must have d( a ) = 0 for every a A and for the annihilator of all [ b, x] = 0, x, Γ, is an ideal of. Suppose that A Z. By the above d( A ) = 0, hence d( d( )) d( A) = 0, that is, d 2 ( x ) = 0 for all x. We obtain 2 d( x) d( y) = 0, so, if char 2 we get d( x) d( y) = 0. Using y = z x leads to d( x) Γ d( x) = 0 and hence d( x ) = 0 for all x. This contradicts d 0. Thus if A Z, we must have char = 2. Let T = { x d( x) = 0}. By the same argument above, we get T A. If t T and x, then d( t x) = d( t) x + t d( x) = t d( x), hence t d( x) centralizes A for all Γ. If a A then a commutes with d( x) A and t d( x), therefore, ( at t a) d( x) = 0 for all x and, Γ. Since is a completely prime Γ -ring, this gives that a t t a = [ a, t] = 0, hence A centralizes T. The T is clearly a Γ -subring of. oreover, it is a Lie ideal of. If t T and x, then d( tx xt) = td ( x) d( x) t = 0 (since d( t ) = 0 ) alaysian Journal of athematical Sciences 289

8 and d( x) Isamiddin S. Rakhimov, Kalyan Kumar Dey & Akhil Chandra Paul A centralizes T. Therefore, x xt T. Since is a completely prime Γ -ring, T is a subring and it is a Lie ideal of. Then either T contains a nonzero ideal of or T is commutative and tt Z( ) for t T. If T contains a nonzero ideal of then A must centralize this ideal, since A centralizes all of T in a completely prime Γ -ring, this forces A to be in Z( ), contrary to supposition. Hence, we have that T is commutative and tt Z( ) for all t T, Γ. Let a A but a Z( ), then T and a x x a T, thus a a Z( ) and ( a x a a) γ ( a x x a) Z( ) for all x and, γ Γ. It follows easily that must be an order of 4-dimentional simple algebra over its center. Hence if A Z( ). Now suppose that A Z( ). Thus d( x) Z( ) for all x. Hence if x, y then d( x y) = x d( y) + d( x) y Z( ). Commuting this with x and using the fact that d( x ), d( y ) are in Z( ), we obtain d( x) ( x y y x) = 0 for all, Γ. However if d( x) 0, since it is in Z( ) we get that it is not a zero divisor. Thus in this case we have x y = y x for all y and Γ. In short, if d( x) 0, since d( ) 0, pick x 0 such that d( x0 ) 0, then x0 Z( ) by the above. Also d( x + x0 ) = d( x) + d( x0 ) = d( x0 ) 0, hence x + x0 Z( ). This leaves us with x Z( ). In the other words, if A Z( ), then is commutative and since is a completely prime Γ -ring, it must be an integral Γ - domain. The theorem is now completely proved. Theorem 2.7 Let be a completely prime Γ -ring satisfying the condition a bc c and, Γ and let d be an nonzero derivation of. Suppose that a is such that [ a, d( x)] = 0 for all x, Γ. Then (i) If is not of char 2, then a is in the center of. (ii) If is of char 2, then a a Z( ). oreover, if a belongs to the center of, then there is an element c of the centroid of and, Γ such that d( x) = [ c a, x], that is d is an inner derivation of. 290 alaysian Journal of athematical Sciences

9 Proof. Suppose that a Z( ). According to the hypothesis one has [ a, d( x y)] = 0 for all x, y and, Γ. Since d( x y) = d( x) y + xd( y) substituting it into the former identity we get [ a, d( x) y + b d( y)] = 0 and this along with the condition ( ) gives d( x) [ a, y] + [ a, x] d( y) = 0. (1) If y commutes with a then [ a, y] = 0, hence the equation (1) reduces to [ a, x] d( y) = 0 for all x,, Γ. Because of a Z( ) we have d( y ) = 0. In the other words d vanishes on the centralizer C ( a) = { y [ a, y] = 0, Γ } of a in. But for any x, d( x) C ( a) by the hypothesis, thus we get d 2 ( x ) = 0. However, if is a completely prime Γ -ring of char is not 2 and d is a 2 derivation of such that d = 0, then d = 0. Since we have supposed that d 0 therefore if the char is not 2, by the results above, we are led to the conclusion that a Z( ). This settles the situation when the char is not 2. So, from this point on we assume that is of char 2 and that a Z( ). In this case the equation (1) becomes [ a, x] d( y) = d( x) [ a, y] (2) for all x, y and, Γ. Thus if d( y ) = 0 then from the equation (2) we obtain d( x) [ a, y] = 0 for all x. Since is a completely prime Γ -ring, we must have [ a, y] = 0, that is y C ( a). Combined with the previous result as we have said earlier d vanishes on C ( a ). Now one knows that C ( a ) coincides with T = { y d( y) = 0}. We return to the equation (2) and substitute in it xγ w for x, where x, w are arbitrary elements of and γ Γ. Hence we get [ a, x] γ w d( y) = d( x) γ w[ a, y] (3) If [ a, x] 0, using a result of [7] we have d( x) = c( x) [ a, x], where c( x ) is in the extended centroid of. oreover, since C ( ) { ( ) 0}, a = y d y = alaysian Journal of athematical Sciences 291

10 Isamiddin S. Rakhimov, Kalyan Kumar Dey & Akhil Chandra Paul we must have that c( x) 0. Also, if [ a, x] = 0, then d( x ) = 0, hence 0 = d( x) = 0 [ a, x]. Thus for all x, d( x) = c( x) γ[ a, x] where c( x ) is in the extended centroid of. We claim that a[ a, x] [ a, x] a = 0 i.e., [ a,[ a, x] ] = 0 for all x and, Γ. It is clear that if [ a, x] = 0 then [ a,[ a, x] ] = 0. On the other hand, if [ a, x] 0, then by the observation above d( x) = c( x) γ[ a, x] where c( x) 0, so since [ a, d( x)] = 0, one has c( x) γ [ a,[ a, x] ] = 0. Because c( x) 0 is invertible, we end up with [ a, [ a, x] ] = 0 for all x. Writing this out we conclude that a a Z( ). a ( x a + a x) = ( a x + x a) a, Now to the final part of the theorem, we just see that if a Z( ) then d( x) = c( x) γ[ a, x] with c( x ) is in the extended centroid for all x. We want to prove that c( x) is a constant. Let x, y, then d( x y) = c( x y) γ [ a, x y], that is d( x) y + x d( y) = c( x y) γ[ a, x] δ y + c( x y) γ xδ [ a, y], by the condition a bc c and, Γ. Because of d( x) = c( x) γ[ a, x] and d( y) = c( y) γ [ a, y]. We get c( x y) γ[ a, x] δ y + c( x y) γ xδ [ a, y] = c( x y) γ [ a, x] δ y + c( x y) δ[ a, y]. (4) Hence if p = c( x) + c( x y) and q = c( y) + c( x y) then we have ( c( x) + c( x y)) δ[ a, x] γ y = ( c( y) + c( x y)) δ xγ [ a, y] for all x, y and for all,, γ, δ Γ. 292 alaysian Journal of athematical Sciences

11 Since a a Z( ) commuting it with a we obtain [ a,[ a, x] ] = 0. Also due to (4) we have ( p + q) δ[ a, x] γ[ a, x] = 0, for all x, y and δ, γ Γ. If [ a, x] γ[ a, y] 0 then p + q = 0 i.e., c( x) + c( x y) + c( y) + c( x y) = 0 and so c( x) = c( y). This shows that c should be a constant on all elements failing to commute with a. Further C ( ) { ( ) 0}, a = y d y = would imply that d( x) = [ c a, x] for x and, Γ for some c in the extended centroid. This is the desired result. So to finish, we must show the existence of a such that w. In fact, we shall show a little more namely that there is an element w such that If this is not true then [ a, x] [ a, w] γ [ a, y] 0. [ a, x] [ a, z] γ [ a, y] = 0 for all z and,, γ Γ,,, γ Γ, that is, we get [ a, x] [ a, z] γ[ a, y] = [ a, x] a zγ [ a, y] [ a, x] z aγ [ a, y]. because of [ a, x] a = p[ a, x] where p is in the extended centroid of. Since a a = b Z( ) we get p p = b, the extended centroid is a field and is of char 2, the element p is uniquely determined by b, hence does not depend on x. But then [ a, x] ( a + q) = 0 for all x such that [ a, x] 0. If [ a, x] = 0, this relation is certainly true. So [ a, x] ( a + p) = 0, for all x,, Γ. But then this carries over to all x in central closure T of which is itself a completely prime Γ -ring. Since a Z( ) and [ a, x] ( a + p) = 0, for all x T and, Γ we alaysian Journal of athematical Sciences 293

12 Isamiddin S. Rakhimov, Kalyan Kumar Dey & Akhil Chandra Paul deduce that a + p = 0 and so a Z( ). We have a contradiction that completes the proof. ACKNOWLEDGENT The research was supported by grant RU Universiti Putra alaysia (UP), alaysia. 294 REFERENCES Barnes, W. E On the Γ-rings of Nabusawa. Pacific Journal of ath. 18: Chakraborty, S. and Paul, A. C On Jordan k-derivations of completely prime Γ-rings. Rocky ountain Journal of ath. 40(6): Ceven, Y Jordan mappings on completely prime Γ-rings. C.U. Fen-Edebiyat Fakultesi Fen Bilimeleri Dergisi. 23(2): Herstein, I. N A notes on derivations. Canad. ath. Bull. 21: Herstein, I. N A notes on derivations II. Canad. ath. Bull. 22(4): Herstein, I. N Rings with Involution. Chicago: Univ. of Chicago Press. Herstein, I. N Topics in Ring Theory. Chicago: Univ. of Chicago Press. Kandamar, H The k-derivation of a Gamma Ring. Turkish Journal of ath. 24: Nabusawa, N On a generalization of the Ring Theory. Osaka Journal of ath. 1: Ozturk,. A. and Jun, Y. B On the centroid of the prime Γ-ring. Comn. Korean ath. Soc. 15(3): alaysian Journal of athematical Sciences

13 Ozturk,. A. and Jun, Y. B On the centroid of the prime Γ-ring II. Turkish Journal of ath. 25: Rakhimov, I. S., Dey, K. K. and Paul, A. C. 2013a. Prime gamma-nearrings with (sigma,tau)-derivations. International Journal of Pure and Applied ath. 82(5): Rakhimov, I. S., Dey, K. K. and Paul, A. C. 2013b. Prime Gamma Rings with Derivations. International Journal of Pure and Applied ath. 83(2): Sapanaci,. and Nakajima, A Jordan derivations on completely prime Γ-rings. ath. Japonica. 46(1): Soyturk, The commutativity in prime Γ-rings with derivation. Turkish Journal of ath. 18: alaysian Journal of athematical Sciences 295

Jordan Γ * -Derivation on Semiprime Γ-Ring M with Involution

Jordan Γ * -Derivation on Semiprime Γ-Ring M with Involution Advances in Linear Algebra & Matrix Theory, 206, 6, 40-50 Published Online June 206 in SciRes. http://www.scirp.org/journal/alamt http://dx.doi.org/0.4236/alamt.206.62006 Jordan Γ -Derivation on Semiprime