1: \begin{abstract}

2: \noindent

3: We study $T^{11-D-q}\times T^q/\Z_n$ orbifold compactifications of eleven-dimensional 

4: supergravity and M-theory using a purely algebraic method. Given the description of 

5: maximal supergravities reduced on square tori as non-linear coset $\sigma$-models, 

6: we exploit the mapping between scalar fields of the reduced theory and directions

7: in the tangent space over the coset to construct the orbifold action as a non-Cartan 

8: preserving finite order inner automorphism of the complexified U-duality algebra. 

9: Focusing on the exceptional serie of Cremmer-Julia groups, we compute the residual 

10: U-duality symmetry after orbifold projection and determine the reality properties 

11: of their corresponding Lie algebras. We carry out this analysis as far as the hyperbolic 

12: $\eten$ algebra, conjectured to be a symmetry of M-theory. In this case the residual 

13: subalgebras are shown to be described by a special class of Borcherds and Kac-Moody 

14: algebras, modded out by their centres and derivations. Furthermore, we construct an 

15: alternative description of the orbifold action in terms of equivalence classes of shift

16: vectors, and, in $D=1$, we show that a root of $\eten$ can always be chosen as the class 

17: representative. Then, in the framework of the $E_{10|10}/K(E_{10|10})$ effective 

18: $\sigma$-model approach to M-theory near a spacelike singularity, we identify these roots 

19: with brane configurations stabilizing the corresponding orbifolds. In the particular case 

20: of $\Z_2$ orbifolds of M-theory descending to type 0' orientifolds, we argue that these 

21: roots can be interpreted as pairs of magnetized D9- and D9'-branes, carrying the 

22: lower-dimensional brane charges required for tadpole cancellation. More generally, we 

23: provide a classification of all such roots generating $\Z_n$ product orbifolds for 

24: $n\leqslant 6$, and hint at their possible interpretation.

25: \end{abstract}

26: