DAMTP/94-70 to app. Proc. 3rd Colloq.

Quantum Groups, Prague, June, 1994

SOME REMARKS ON THE QUANTUM DOUBLE

arXiv:hep-th/9409056v1 10 Sep 1994

Abstract We recall the abstract theory of Hopf algebra bicrossproducts and double cross products due to the author. We use it to develop some less-well known results about the quantum double as a twisting, as an extension and as q-Lorentz group. Keywords: quantum double quantum group gauge theory twisting q-Lorentz group q-Minkowski space

Introduction

The quantum double Hopf algebra D(H) was introduced by Drinfeld[1]. Since then it has been extensively studied by the author in [2][3][4][5][6][7] as well as by Reshetikhin and Semenov-Tsian Shanskii in [8]. It was also proposed as q-Lorentz group in [9] and connected with an R-matrix approach[10] in our paper [6, Sec. 4]. Here we collect some of the main results and develop two less well known ones. They demonstrate the power of some algebraic techniques for Hopf algebras relating to cross products and cross coproducts. We write formulae over C but in fact these abstract constructions work over a general eld too. We use the summation convention and also the standard abstract notation h = h(1) h(2) for the quantum group coproduct. This stands for a sum of terms in the tensor product.

Quantum double as a double cross product

The denition of the quantum double that we use is the algebraic form due to the author[2] as an example of a double cross product. If H is a nite dimensional Hopf algebra or quantum group, its quantum double D(H) is built on the vector space H H with product (a h)(b g) = b(2) a h(2) g Sh(1) , b(1) h(3) , b(3)
1

(1)

Royal Society University Research Fellow and Fellow of Pembroke College, Cambridge

and tensor product unit, counit and coproduct. It is a quasitriangular Hopf algebra with universal R-matrix R = (f a 1) (1 ea ). This is equivalent to Drinfelds denition[1] with generators and relations. The rst thing to notice is that the Hopf algebra construction here works perfectly well if H is not nite-dimensional: in this case it is really a function of two Hopf algebras A, H which are dually paired by a map , : H A C. We call this D(A, H, , ) the generalised

quantum double. Moreover, it is not necessary to have Hopf algebras either. All this works at the bialgebra level if we replace S( ), by the convolution inverse map ,
1 .

The linear

maps from H A C form an algebra using the coproduct of H A, and we need our pairing invertible in this algebra[11]. Next, we introduce a general construction of which this generalised quantum double is an example. This is the concept of a double cross product. We consider two bialgebras or Hopf algebras which act on each other by maps : H A H, : H A A

which are coalgebra homomorphisms (one says that H is a right A-module coalgebra and A is a left H-module coalgebra). We suppose further that these actions are a matched pair in that they obey the compatibility conditions[2] (hg)a = (h(g(1) a(1) ))(g(2) a(2) ), h(ab) = (h(1) a(1) )((h(2) a(2) )b), 1a = (a) h1 = (h) (2)

h(1) a(1) h(2) a(2) = h(2) a(2) h(1) a(1) In this case, one has the theorem[2] that there is a double cross product bialgebra AH built on the vector space A H with product (a h)(b g) = a(h(1) b(1) ) (h(2) b(2) )g (3)

and tensor product unit, counit and coproduct. It is clear that A, H are subalgebras and that AH factorises into them. The theorem is: Theorem 2.1 cf.[2]. If a bialgebra X factorises in the sense that there are sub-bialgebras AX H such that (i j) : A H X is an isomorphism of vector spaces, then A, H are a matched pair as above and X AH. =
i j

Proof

This is based on [2, Sec. 3.2] but in a cleaner form. For this reason we give the full

details. Since (i j) is a linear isomorphism we have a well-dened linear map : H A A H dened by j(h)i(a) = (i j) (h a). Associativity in X and that i, j are algebra maps tells us that (i j) (hg a) = j(hg)i(a) = j(h)j(g)i(a) = j(h) (i j) (g a) = (i j) (id ) 12 23 (h g a) (i j) (h ab) = j(h)i(ab) = j(h)i(a)i(b) = (i j) (h a) i(b) = (i j) ( id) 23 12 (h a b) so we conclude from this and by a similar consideration of the identity element in X that ( id) = (id ) 12 23 , (id ) = ( id) 23 12 , (1 a) = a 1 (h 1) = 1 h

(4)

where the suces refer to the tensor factor on which acts. We see that the products in H and A respectively commute with in a certain sense. This much works just at the algebra level (factorisations of algebras are of this form). Note that the braided tensor product of algebras[12] is an example of such an algebra factorisation. Its introduction was motivated by this part of the double cross product theorem. Next, we use the counits to dene linear maps : H A H and : H A A by = (id ) , = ( id) .

Applying id to the rst line of (4) and id to the second tells us that is a left action and is a right action as required. Applying id to the rst and id to the second tells us that (hg)a = (h(g a)), h(ab) = ((h a)b), 1a = (a) h1 = (h).

This much works for algebras equipped with homomorphisms . Next we use that i, j are coalgebra maps to deduce that (i j) : A H X and (j i) : H A X are coalgebra maps too. The inverse of (i j) is also a coalgebra map hence the ratio is also a coalgebra map. I.e., we conclude that A H = ( ) H A , 3 ( ) (h a) = (a)(h).

Now we apply id id both sides of the rst of these to conclude that h(1) a(1) h(2) a(2) = (h a) which means that our results above prove the rst two of (2). Applying instead the map id id gives by contrast h(1) a(1) h(2) a(2) = (h a) where is usual transposition. This proves the third of (2). Likewise, applying instead

id id gives that is a coalgebra map, while applying id id gives that is a coalgebra map too. Hence we have all the conditions needed for a matched pair in the sense of (2). Finally, looking again at the equation for j(h)i(a) above tells us now that (i j) becomes an isomorphism between the corresponding double cross product AH and our original bialgebra X. The quantum double ts as an example of this theory as follows: given two paired bialgebra or Hopf algebras we dene ha = h(2) h(1) , a(1)
1

h(3) , a(2) ,

ha = a(2) h(1) , a(1)

h(2) , a(3)

and check that they obey the conditions (2) for a matched pair (Aop , H, , ). We recover the generalised quantum double as D(A, H, , ) = Aop H which is how the generalised quantum double was rst introduced in [11]. In the converse direction, one knows that the double factorises[1] and can use the above theorem to deduce , . Next, we make a trivial and completely cosmetic change: we denote Aop in the above generalised quantum double by A. Then of course this new A is no longer dually paired with H: instead we denote the same linear map axioms (hg a) = (h a(1) )(g a(2) ), (h ab) = (h(1) b)(h(2) a). which are just the axioms for , (5) , as a skew pairing : H A C. It obeys the

in terms of the new A which has the opposite product to the

old one which was dually paired. Then D(A, H, ) AH (a h)(b g) = ab(2) h(2) g 1 (h(1) b(1) )(h(3) b(3) ) 4 (6)

with tensor product unit and counit is a double cross product by ha = h(2) 1 (h(1) a(1) )(h(3) a(2) ) ha = a(2) 1 (h(1) a(1) )(h(2) a(3) ). Finally, it is well-known in quantum groups that every construction has a dual one. It is an easy exercise to cast the above into their dual form (and not a new result to do so!). We list the result here for completeness. Firstly, the general construction is a double cross coproduct. We consider two bialgebras or Hopf algebras coacting on each other by maps : A A H, : H AH

which are algebra homomorphisms (comodule algebras). We require further that the coactions are matched in the sense ( id) (a) = ((id ) (a(1) )) (1 (a(2) )) (id ) (h) = ((h(1) ) 1) (( id) (h(2) )) (a)(h) = (h)(a). In this case, one has the theorem that there is a bialgebra H A with tensor product algebra structure and counit, and (h a) = h(1) (a(1) )(h(2) ) a(2) . In the Hopf algebra setting this is a Hopf algebra. The double cross coproduct comes equipped with bialgebra or Hopf algebra surjections H H AA
p q

(7)

(8)

into which factors the double cross coproduct decomposes by (p q) . Conversely, and bialgebra decomposing like this is a double cross coproduct. Our generalised quantum double example (6) becomes a generalised quantum codouble associated to a skew copairing between bialgebras or Hopf algebras A, H. This is an element A H such that ( id) = 13 23 , (id ) = 13 12 , (9)

which is just the dual of (5). We suppose is invertible. Then D (H, A, ) = H A

(10)

with tensor product counit and algebra structure is a double cross coproduct by coactions (a) = 1 (a 1), (h) = 1 (1 h).

This general construction includes the dual of Drinfelds quantum double as D(H) = H cop H . As well as being the dual of our generalised quantum double, the resulting ex amples (10) of double cross coproducts associated to a skew copairing were found independently in [8] by another route as a twisted product. The double cross coproduct construction itself is however, much more general than this.

Quantum double as a twisting, I

If A is a dual quasitriangular bialgebra (such as A(R) according to results in [13, Sec. 3]) then the universal-R-matrix functional R : A A C as in [14][15] obeys (5). Hence we have a double cross product[11] D(A, A, R) = AA which we identied[6] as a natural construction for the q-Lorentz group. It is not in general isomorphic to Drinfelds double, i.e. we use the more general double (6) coming out of the theory of double cross products. There is however, a Hopf algebra homomorphism [6, Sec. 4] AA D (H), a b (id a(1) b(1) )(R1 R31 ) a(2) b(2) 12

if A is dual to H. It is an isomorphism in some (factorisable) cases such as SUq (2)SUq (2). In matrix generators s, t for the latter, the above map is[6] s 1 Sl+ t, 1 t Sl t

where l are the matrix FRT generators[16] of Uq (su2 ). Its codouble as an algebra is the tensor product Uq (su2 ) SUq (2). The same formula works for general Uq (g). As explained in [6] this is just a dual version of a corresponding result in enveloping algebra form which is essentially due to [8]. Thus, if H is a strict quantum group with universal Rmatrix (such as Uq (g)) then = R is a skew copairing and we have an example of a generalised codouble or double cross coproduct D (H, H, R) = H H recovering the twisted square introduced in another framework in [8]. These authors showed that D(H) H H, a h (id a)(R1 R23 ) (h) 31 6

is a homomorphism of Hopf algebras, and argued that it should be an isomorphism if Q = R1 R1 is invertible when regarded as a linear map H H by evaluation on its rst tensor 21 factor (a version of the factorisable case). The required inverse map was not given in [8] and appears here I think for the rst time (being obtained in fact from the bosonisation theory in Section 6). It is h g (Q1 (hSg(1) ))(1) R(2) , (Q1 (hSg(1) ))(2) R(1) g(2) . According to [6, Sec. 4] the example Uq (su2 ) Uq (su2 ) is the correct construction for the enveloping algebra of the q-Lorentz group dual to[10] and is isomorphic to D(Uq (su2 )) by the above maps. In matrix l generators the isomorphism from the quantum double to the twisted square is t 1 l l+ , 1 l l l .

We also introduced in [6] a -structure in AA given by (a b) = b a in the case when R is of real-type in the sense in [7]. It is not a new result to repeat this in the dual form for H H where it immediately comes out as (h g) = R21 (g h )R1 ; 21 R = R21 .

It is also clear from the form of the product of AA and H H that they are examples of dual-twisting and twisting respectively in the more modern sense by a non-Abelian 2-cocycle[17], a theory due to Drinfeld in [18] and restricted to Hopf algebras in [19]. Briey, a 2-cocycle on a quantum group is : A A C such that (b(1) c(1) )(a b(2) c(2) ) = (a(1) b(1) )(a(2) b(2) c) (1 a) = (a). The other side (a 1) = (a) follows. Given such a cocycle, the new product a = (a(1) b(1) )a(2) b(2) 1 (a(3) b(3) ) b (12) (11)

denes a new bialgebra A. Similarly for the antipode and any dual quasitriangular structures of A. It is clear from the form (6) of AA that it is the twisting by ((a b) (c d)) = (a)R1 (b c)(d) as a 2-cocycle on A A. In the braided geometrical approach to space time SUq (2) SUq (2) acts covariantly on a q-Euclidean space and this cocycle twists it to SUq (2)SUq (2) acting on q-Minkowski space, i.e. it is physically a quantum Wick rotation[20]. 7

Likewise, a 2-cocycle for a quantum group H is H H such that[18][19] 23 (id ) = 12 ( id), ( id) = 1 (13)

and means that H with = ( )1 etc., is also a quantum group. This twisting is even more clear for H H where the coproduct (10) is obviously a twisting by a 2-cocycle = R1 for 23 H H. So this picture of the q-Lorentz enveloping algebra is likewise an immediate corollary of its description introduced by the author in [6] as the twisted square Uq (su2 ) Uq (su2 ). Some of this has subsequently been reiterated by other authors following [6].

Quantum double as a twisting, II

Now we look at a more well-known topic which is the quantum double of the Borel subalgebra Uq (b ) say of Uq (sl2 ). It is known that D(Uq (b+ )) projects onto Uq (sl2 ) and this is indeed how the quasitriangular structure of the latter was deduced[1]. Likewise for general Uq (g). We begin by recalling this calculation with a modern proof based on q-geometry. For Uq (b ) we take the Hopf algebra H generated by X, g, g1 say and gX = q 2 Xg, X = 0, X = X 1 + g X, Sg = g1 , g = g g

g = 1,

SX = g1 X.

We show that when q 2 = 0, 1 it is a self-dual Hopf algebra in the sense that it has a Hopf algebra pairing with itself given by g, g = q 2 , g, X = 0, X, g = 0, X, X = 1 . 1 q 2

This is essentially due to Drinfeld in [1], but here is a new proof. We dene a right action of H on itself by (X, g)g = (q 2 X, q 2 g), where (X, g) = (X, g)X = q (X, g); = 1 1 q 2

ab X a gb is assumed to have powers of X to the left and q X m gn = [m; q 2 ]X m1 gn ; [m; q 2 ] = 1 q 2m 1 q2

is a q-derivative. Since q lowers the degree by 1, it is clear that (g)X = q 2 (X)g for all H, so indeed denes a right action of our Hopf algebra on itself. Moreover, one has ()g = (g)(g), ()X = (X) + (g)(X) 8 (14)

for all normal-ordered polynomials (X, g), (X, g). Note that the product must be normalordered using the q-commutation relations between X, g before we can apply the denition of . The action of g on products is clear since the normal ordering process commutes with scaling of the generators by q 2 . The action of X is easily veried on monomials as (X m gn X k gl )X = (q 2nk X m+k gn+l )X = q 2nk [m + k; q 2 ]X m+k1 gn+l = q 2nk [m; q 2 ] + q 2m [k; q 2 ] X m+k1 gn+l = [m; q 2 ]X m1 gn X k gl + q 2(m+n) X m gn [k; q 2 ]X k gl = ((X m gn )X)X k gl + ((X m gn )g)((X k gl )X) from the denition of and the relations of H. This result (14) implies that the product of H is covariant (a module algebra) under the action . Given this action , we now dene h, a = (ha) and deduce that it obeys half the axioms of a pairing from our module algebra property (14). But since the resulting pairing X m gn , X k gl = ((X k gl )X m )gn = m,k m [m; q 2 ]!q 2nl is symmetric in the two arguments of , , we also deduce the other half by symmetry. (15)

This proof also has the merit that the nondegeneracy of the pairing is clear, at least when q is
m generic. This is because (X, g), X m gn = m (q )(0, q 2n ) vanishing for all m, n implies that

vanishes too. Reducing non-degeneracy to the q-derivative or higher-dimensional braidedderivative[21] appears to be a useful strategy of proof which I have not seen elsewhere. It is now an easy matter to compute D(Uq (b )) = H op H as generated by H = Uq (b ) as above and H op = Uq (b+ ) with generators X, g , g1 say. The result from (1) is gX = q 2 Xg, gX = q 2 Xg, X X q 2 XX =
g g1 q 2 1

X = X 1 + g X,

X = X 1 + g X

(16)

with the corresponding counit and antipode. The generators g, g are group-like. We want to make two comments about this quantum double. The rst working over formal powerseries C[[t]] in a deformation parameter t, and the second in Section 5 working over C. In the rst case we change variables X = q 2 X ,
H H X = q 2 X+ ,

g = q H q C ,

g = q H q C

to new generators X , H, C then we get the usual algebra relations of Uq (sl2 ) along with C central and primitive and X = X q 2 + q 2 q C X . Following Drinfeld one can now derive the universal R-matrix of Uq (sl2 ) from the one of D(Uq (b )) after projecting by C = 0. On the other hand, we can consider the 1-dimensional quantum group Uq1 (C) with generator C and C = C 1 + 1 C, R = q 2 C C
1 H H

This is called the quantum line in[22]. On the Hopf algebra Uq (sl2 ) Uq1 (C) we dene = q 2 C H and check that it is a 2-cocycle. Then [C H, ( ) X ] = 2C( ) X tells us that the coproducts of X in the double are just the conjugation by of the usual coproduct of Uq (sl2 ), i.e. we have D(Uq (sl2 ))Uq (sl2 ) Uq1 (C) = as twisted by this cocycle. With a bit more work we have this as a twisting of quasitriangular Hopf algebras.
1

Quantum double as a cocycle bicrossproduct

Now we recall another aspect of abstract Hopf algebra theory, which is the notion of a cocycle cross product. Thus a (right) cocycle action of a Hopf algebra H on an algebra A is linear maps : A H A, where 1h = 1(h), a1 = a, (ab)h = (ah(1) )(bh(2) ) (17) : H H A

(h(1) g(1) )((ah(2) )g(2) ) = (a(h(1) g(1) ))(h(2) g(2) )

(h(1) g(1) f (1) )((h(2) g(2) )f (2) ) = (h g(1) f (1) )(g(2) f (2) ) (1 h) = (h 1) = (h) for all h, g, f H and a, b A. In this situation, there is a right cocycle cross product algebra H A on the vector space H A with product < (h a)(g b) = h(1) g(1) (h(2) g(2) )(ag(3) )b. 10 (18)

That this product is associative follows from the assumptions (17). A cohomological picture of such cocycle cross coproducts is due to [23]. They can be interpreted as trivial quantum principal bundles in quantum group gauge theory [24], with a new quantum number not present classically. In the dual setting we have of course the notion of a cocycle coaction. So a left cocycle coaction of a Hopf algebra A on a coalgebra H for example means maps : H A H, : H AA

obeying some axioms dual to those above. See [11]. We will not need the most general form
with the dual-cocycle . In this case we have a usual cross coproduct: if (h) = h(1) h(2) is a

coaction of A on the coalgebra H and respects its structure in the sense (id ) (h) = h(1) (1) h(2) (1) h(1) (2) h(2) (2) ,

(id ) =

(19)

(a comodule coalgebra) then there is a cross coalgebra H> A built on H A with (h a) = h(1) h(2) (1) a(1) h(2) (2) a(2) ,

(h a) = (h)(a).

(20)

Putting these ideas together we suppose next that H acts on A by a cocycle action as above and A coacts back on H. We can demand conditions on these such that the resulting cross product algebra and coalgebra t together to form a bicrossproduct Hopf algebra H A. This was introduced for general Hopf algebras by the author in [2] (without cocycles) and [11] (with cocycles), with some examples in [2][25] and [26] respectively. We refer to these papers for the precise compatibility conditions which ensure that the bicrossproduct is a Hopf algebra. A general feature of the bicrossproduct is that A H A H by the obvious inclusion and the obvious projection via . There is a theorem that every extension of Hopf algebras like this which is invertible in a certain sense (the extension should be cleft and cocleft) is of this bicrossproduct type, possibly with cocycles. For a recent application of this theory of extensions to physics, see [27]. We want to note only that that quantum double of H = Uq (b ) in the form computed in the last section, is exactly such an extension
CZ D(Uq (b ))Uq (sl2 ).
p

11

Here p is dened by p(X) = F q 2 X ,

p(X) = E q 2 X+ ,

p(g) = p() = q H g

(21)

where the E, F, q H can be considered as generating a version of Uq (sl2 ). It is a sub-Hopf algebra of the usual form since we do not use q 2 themselves. With this form of Uq (sl2 ) understood, we have p as a surjection. Next, it is clear that the invertible element K = g g1 generates
CZ D(H) as a commutative sub-Hopf algebra. Moreover, it is clear that D(H) q (sl2 ) CZ =U
H

as vector spaces by X m X n gk K l F m E n q Hk K l . Finally, one veries that this isomorphism respects multiplication by K from the right, and the coaction of Uq (sl2 ) from the left in the manner required for a Hopf algebra extension. This and some technical cleftness conditions which one can also verify means that our quantum double must be a cocycle bicrossproduct. It is D(Uq (b ))Uq (sl2 ) CZ, = 1K ( F ) = 1 F , (E F ) = H 1 q 2 q 1 q H
E K E

with = on the other generators of Uq (sl2 ). The action in the bicrossproduct and the dual-cocycle are both trivial. We identify X = F 1, X = E 1, g = q H 1 and K 1 K. Then the product from (18) as modied by the cocycle gives XX = (E 1)(F 1) = E (1) F (1) (E (2) F (2) ) 1K = EF 1 + g2 (E F ) = EF 1 + g2 1 q 2 1 g2 1K = q2F E 1 + 1 + g2 2 1q 1 q 2 2 K 2 1g 1g K = q 2 (F 1)(E 1) + = q2 X X + 1 q 2 1 q 2 which are the relations (16) for the quantum double. The coalgebra structure (20) as modied by is X = (E 1) = E 1 1 1 + q H (E) 1 = X 1 + gK X as required in (16). This completes our bicrossproduct description of D(Uq (b )). We see that as an algebra it is a central extension of Uq (sl2 ) by the cocycle , and as a coalgebra it is a cross coproduct. Since CZ = C(S 1 ) we can think of this as a quantum principle bundle over S 1 with bre SUq (2) in the geometrical picture of [24]. For a full geometrical picture one should consider the -structure also, with the above as the chiral part of the full picture.

12

Quantum double as a bosonisation

For completeness, we conclude here with a brief mention of another abstract result on the quantum double already published in [3][5][7]. This applies to the case when A is a dual quasitriangular Hopf algebra or bialgebra (such as A(R)). According to the theory of transmutation[12] it has a braided group version B which lives in the braided category of representations of the corresponding quasitriangular Hopf algebra H (such as Uq (g)). For example, the transmutation of the quantum matrices A(R) is just the braided matrices B(R) introduced in this way in [28] with matrix generators u and relations R21 u1 Ru2 = u2 R21 u1 R. It turns quantum 22 matrices Mq (2) into braided (hermitian) 2 2 matrices BMq (2) or q-Minkowski space. Likewise at the Hopf algebra level it turns SUq (2) into the braided group BSUq (2) which is the mass-shell in q-Minkowski space[7]. On the other hand, there is a general theory of bosonisation [29] which turns any braided group B in such a representation category into an ordinary quantum group B>H. For the algebra we make a cross product as in (18) but with a left action (say) rather than a right one, and without a cocycle. By denition, B is covariant under H and we use this action for the cross product. For the coproduct we use the induced left coaction[3] : B H B, (b) = R21 b

and make the cross coproduct (20). This is a generalisation to braided groups of the JordanWigner transformation for superalgebras. When applied to braided planes, this bosonisation gives us q-Poincar Hopf algebras[6]. On e the other hand, when applied to the above-mentioned braided group obtained from A dual to H, it gives the quantum double[3][5] B>H D(H). = Cross product algebras can also be considered as quantisation of position functions B by momentum quantum group H, which is the interpretation developed in [7]. We can also consider the quantum double in this form as a trivial quantum principal bundle[24]. For example, BSUq (2)>Uq (su2 ) is a bundle over the mass-shell in q-Minkowski space with bre SUq (2) .

References
[1] V.G. Drinfeld. Quantum groups. In A. Gleason, editor, Proceedings of the ICM, pages 798820, Rhode Island, 1987. AMS. 13

[2] S. Majid. Physics for algebraists: Non-commutative and non-cocommutative Hopf algebras by a bicrossproduct construction. J. Algebra, 130:1764, 1990. From PhD Thesis, Harvard, 1988. [3] S. Majid. Doubles of quasitriangular Hopf algebras. Comm. Algebra, 19(11):30613073, 1991. [4] S. Majid. Representation-theoretic rank and double Hopf algebras. Comm. Algebra,

18(11):37053712, 1990. [5] S. Majid. Braided matrix structure of the Sklyanin algebra and of the quantum Lorentz group. Comm. Math. Phys., 156:607638, 1993. [6] S. Majid. Braided momentum in the q-Poincar group. J. Math. Phys., 34:20452058, 1993. e [7] S. Majid. The quantum double as quantum mechanics. J. Geom. Phys., 13:169202, 1994. [8] N.Yu. Reshetikhin and M.A. Semenov-Tian-Shansky. Quantum R-matrices and factorisation problems. J. Geom. Phys., 5:533, 1988. [9] P. Podles and S.L. Woronowicz. Quantum deformation of Lorentz group. Comm. Math. Phys, 130:381431, 1990. [10] U. Carow-Watamura, M. Schlieker, M. Scholl, and S. Watamura. A quantum Lorentz group. Int. J. Mod. Phys., 6:30813108, 1991. [11] S. Majid. More examples of bicrossproduct and double cross product Hopf algebras. Isr. J. Math, 72:133148, 1990. [12] S. Majid. Braided groups and algebraic quantum eld theories. Lett. Math. Phys., 22:167 176, 1991. [13] S. Majid. Quasitriangular Hopf algebras and Yang-Baxter equations. Int. J. Modern Physics A, 5(1):191, 1990. [14] S. Majid. Braided groups. J. Pure and Applied Algebra, 86:187221, 1993. [15] S. Majid. Quantum and braided linear algebra. J. Math. Phys., 34:11761196, 1993. [16] L.D. Faddeev, N.Yu. Reshetikhin, and L.A. Takhtajan. Quantization of Lie groups and Lie algebras. Leningrad Math J., 1:193225, 1990. 14

[17] S. Majid. Cross product quantisation, nonabelian cohomology and twisting of Hopf algebras. In Proc. Generalised Symmetries, Clausthal, Germany, July 1993. World Sci. To appear. [18] V.G. Drinfeld. QuasiHopf algebras. Algebra i Analiz, 1(6):2, 1989. In Russian. Translation in Leningrad Math J. [19] D.I. Gurevich and S. Majid. Braided groups of Hopf algebras obtained by twisting. Pac. J. Math., 162:2744, 1994. [20] S. Majid. q-Euclidean space and quantum Wick rotation by twisting. J. Math. Phys., 35, 1994. [21] S. Majid. Free braided dierential calculus, braided binomial theorem and the braided exponential map. J. Math. Phys., 34:48434856, 1993. [22] V.V. Lyubashenko and S. Majid. Fourier transform identities in quantum mechanics and the quantum line. Phys. Lett. B, 284:6670, 1992. [23] Y. Doi. Equivalent crossed products for a Hopf algebra. Comm. Algebra, 17:30533085, 1989. [24] T. Brzeziski and S. Majid. Quantum group gauge theory on quantum spaces. Comm. n Math. Phys., 157:591638, 1993. [25] S. Majid. Hopf algebras for physics at the Planck scale. J. Classical and Quantum Gravity, 5:15871606, 1988. [26] S. Majid and Ya. S. Soibelman. Bicrossproduct structure of the quantum Weyl group. J. Algebra, 163:6887, 1994. [27] S. Majid and H. Ruegg. Bicrossproduct structure of the -Poincar group and none

commutative geometry. Phys. Lett. B, 1994. [28] S. Majid. Examples of braided groups and braided matrices. J. Math. Phys., 32:32463253, 1991. [29] S. Majid. Cross products by braided groups and bosonisation. J. Algebra, 163:165190, 1994.

15