On toric varieties of high arithmetical rank, , November 27, 2006

Introduction The arithmetical rank (ara) of an algebraic variety is the minimum number of equations that are needed to define it set-theoretically. For every affine variety

V \subset K^{N}

we have that codim

V \leq

ara

V \leq N

. This general upper bound was found by Eisenbud and Evans [7] . In particular cases a better upper bound can be obtained by direct computations based on Hilbert's Nullstellensatz: this was done for certain toric varieties in [1] , [2] , [3] , [4] , [5] . In all the examples treated there the arithmetical rank was close to the trivial lower bound, i.e., ara

V \leq

codim

V + 1

. In this paper we present a class of toric varieties whose arithmetical rank is close to the general upper bound, namely ara

V = N - 2

. For proving this result, of course, we need a more efficient lower bound:

this is provided by étale cohomology. The same kind of tools was used in [3] .

There they were applied for showing that the arithmetical rank of certain toric varieties of codimension 2 depends on the characteristic of the ground field, and that ara

V =

codim

V

in exactly one positive characteristic. In the present paper, however, we study toric varieties of any codimension, and obtain the same arithmetical rank in all characteristics.

1 The main theorem

Let

K

be an algebraically closed field, and let

n \geq 2

be an integer. Let

e_{1}, . . ., e_{n}

be the standard basis of

Z Z^{n}

. Set

N = 2 n - 1

and consider the following subset of

I N^{N}

T = {e_{1}, . . ., e_{n - 1}, d e_{n}, a_{1} e_{1} + e_{n}, . . ., a_{n - 1} e_{n - 1} + e_{n}},

where

d, a_{1}, . . ., a_{n - 1} \in I N^{*}

. We also suppose that

d

is divisible by two distinct primes

p

and

q

. With

T

we can associate the variety

V

admitting the following parametrization

V : {\begin{matrix} x_{1} & = & u_{1} \\ . . . \\ x_{n - 1} & = & u_{n - 1} \\ x_{n} & = & u_{n}^{d} \\ y_{1} & = & u_{1}^{a_{1}} u_{n} \\ . . . \\ y_{n - 1} & = & u_{n - 1}^{a_{n - 1}} u_{n} \end{matrix},

which is a toric variety of codimension

n - 1

in the affine space

K^{N}

. Our aim is to show the following

Theorem 1 ara

V = N - 2

This will be accomplished in two steps.

2 The defining equations

We show that ara

V \leq N - 2

by exhibiting

N - 2

binomials which define

V

set-theoretically.

Proposition 1 The variety

V

is set-theoretically defined by the following binomials:

\begin{matrix} F_{i} & = & y_{i}^{d} - x_{i}^{a_{i} d} x_{n} & for i = 1, . . ., n - 1 \end{matrix}

\begin{matrix} G_{i} & = & y_{i}^{d - 1} y_{i + 1} - x_{i}^{a_{i} (d - 1)} x_{i + 1}^{a_{i + 1}} x_{n} & for i = 1, . . ., n - 2 \end{matrix}

Proof .It is straightforward to check that for all

w \in V

F_{i} (w) = 0

and

G_{i} (w) = 0

for all indices

i

. Conversely, we have to prove that for every

w \in K^{N}

such that

\begin{matrix} F_{i} (w) = 0 & for i = 1, . . ., n - 1 \end{matrix}

(1)

\begin{matrix} G_{i} (w) = 0 & for i = 1, . . ., n - 2 \end{matrix}

(2)

we have that

w \in W

. Let

w = ({\bar{x}}_{1}, . . ., {\bar{x}}_{n}, {\bar{y}}_{1}, . . ., {\bar{y}}_{n - 1}) \in K^{N}

be a point fulfilling ( 1 ) and ( 2 ). We show that, for a suitable choice of parameters

u_{1}, . . ., u_{n}

w = (u_{1}, . . ., u_{n - 1}, u_{n}^{d}, u_{1}^{a_{1}} u_{n}, . . ., u_{n - 1}^{a_{n - 1}} u_{n})

. This is certainly true if

{\bar{x}}_{n} = 0

: in this case ( 1 ) implies that

{\bar{y}}_{i} = 0

for

i = 1, . . ., n - 1

, and we can take

u_{i} = {\bar{x}}_{i}

for all

i = 1, . . ., n - 1

. Suppose that

{\bar{x}}_{n} \neq 0

. Then, by ( 1 ), for all

i = 1, . . ., n - 1

, we have

{\bar{x}}_{i} = 0

if and only if

{\bar{y}}_{i} = 0

. If

n = 2

and

{\bar{x}}_{1} = 0

, then take

u_{1} = 0

and set

u_{2}

equal to any

d

-th root of

{\bar{x}}_{2}

. Otherwise, up to replacing

n

with a smaller number, we may assume that

{\bar{x}}_{i} \neq 0

for all

i = 1, . . ., n - 1

Set

u_{i} = {\bar{x}}_{i}

for all

i = 1, . . ., n - 1

and let

u_{n}

be a

d

-th root of

{\bar{x}}_{n}

. By ( 1 ) we have that, for all

i = 1, . . ., n - 1

{\bar{y}}_{i}^{d} = u_{i}^{a_{i} d} u_{n}^{d},

which implies that

\begin{matrix} {\bar{y}}_{i} = u_{i}^{a_{i}} u_{n} ω_{i} \end{matrix}

(3)

for some

d

-th root

ω_{i}

of 1. On the other hand, from ( 2 ) we deduce that

\begin{matrix} {\bar{y}}_{i}^{d - 1} {\bar{y}}_{i + 1} = u_{i}^{a_{i} (d - 1)} u_{i + 1}^{a_{i + 1}} u_{n}^{d} \end{matrix}

(4)

Replacing ( 3 ) on the left-hand side of ( 4 ) gives

\begin{matrix} u_{i}^{a_{i} (d - 1)} u_{n}^{d - 1} ω_{i}^{d - 1} u_{i + 1}^{a_{i + 1}} u_{n} ω_{i + 1} = u_{i}^{a_{i} (d - 1)} u_{i + 1}^{a_{i + 1}} u_{n}^{d}, \end{matrix}

(5)

and finally, being

u_{i}, u_{i + 1}, u_{n}

non zero, we have that, for all

i = 1, . . ., n - 2

\begin{matrix} ω_{i}^{d - 1} ω_{i + 1} = 1 . \end{matrix}

(6)

This implies that

ω_{i} = ω_{i + 1}

for all

i = 1, . . ., n - 2

, i.e.,

ω_{1}, . . ., ω_{n - 1}

are all equal to the same

d

-th root

ω

of 1. Set

u_{n}^{'} = ω u_{n}

. Then

{\bar{x}}_{n} = u_{n^{'}}^{d}

and, by ( 3 ),

{\bar{y}}_{i} = u_{i}^{a_{i}} u_{n}^{'}

for

i = 1, . . ., n - 1

Hence

u_{1}, . . ., u_{n - 1}, u_{n}^{'}

is the required set of parameters.

3 The lower bound

We show that ara

V \geq N - 2

by means of the following criterion, which is based on étale cohomology and is cited from [6] , Lemma 3

^{'}

Lemma 1 Let

W \subset \tilde{W}

be affine varieties. Let

d = dim \tilde{W} \ W

. If there are

s

equations

F_{1}, . . ., F_{s}

such that

W = \tilde{W} \cap V (F_{1}, . . ., F_{s})

, then

H^{d + i} e t (\tilde{W} \ W, Z Z / r Z Z) = 0 for all i \geq s

and for all

r \in Z Z

which are prime to char

K

Hence our claim will follow once we have shown

Proposition 2 If char

K \neq p

, then

H^{2 N - 3} e t (K^{N} \ V, Z Z / p Z Z) \neq 0 .

If char

K \neq q

, then

H^{2 N - 3} e t (K^{N} \ V, Z Z / q Z Z) \neq 0 .

Proof .Let

r \in {p, q}

, and suppose that char

K \neq r

. In the sequel

H e t

and

H c

will denote étale cohomology and étale cohomology with compact support with respect to the coefficient group

Z Z / r Z Z

: we shall omit the latter for the sake of simplicity. By Poincaré Duality (s. [8] , Cor. 11.2, p. 276) we have

\begin{matrix} H^{2 N - 3} e t (K^{N} \ V) ≃ H^{3} c (K^{N} \ V) \end{matrix}

(7)

Moreover, we have a long exact sequence of cohomology with compact support

\dots \to H^{2} c (K^{N}) \to H^{2} c (V) \to H^{3} c (K^{N} \ V) \to H^{3} c (K^{N}) \to \dots,

where

H^{2} c (K^{N}) = H^{3} c (K^{N}) = 0

, since

N \geq 3

. Hence

\begin{matrix} H^{3} c (K^{N} \ V) ≃ H^{2} c (V) \end{matrix}

(8)

By ( 7 ) and ( 8 ) it thus suffices to show that

\begin{matrix} H^{2} c (V) \neq 0 \end{matrix}

(9)

K^{n}

fix the coordinates

u_{1}, . . ., u_{n}

and let

X

be the subvariety of

K^{n}

defined by

u_{1} = u_{2} = \dots = u_{n - 1} = 0

. Then

X

is a 1-dimensional affine space over

K

, on which we fix the coordinate

u_{n}

. Consider the surjective map

φ : K^{n} \to V

(u_{1}, . . ., u_{n}) \mapsto (u_{1}, . . ., u_{n - 1}, u_{n}^{d}, u_{1}^{a_{1}} u_{n}, . . ., u_{n - 1}^{a_{n - 1}} u_{n})

and the restriction map

\tilde{φ} : K^{n} \ X \to V \ φ (X),

which is a bijective morphism of affine schemes. For all

i = 1, . . ., n - 1

let

V_{i} = {({\bar{x}}_{1}, . . ., {\bar{x}}_{n - 1}, {\bar{x}}_{n}, {\bar{y}}_{1}, . . ., {\bar{y}}_{n - 1}) \in V | {\bar{x}}_{i} \neq 0}

. These sets form an open cover of

V \ φ (X)

, and

U_{i} = φ^{- 1} (V_{i}) = {(u_{1}, . . ., u_{n}) \in K^{n} | u_{i} \neq 0}

. Moreover, for all

i = 1 . . ., n - 1

, the morphism

ψ_{i} : V_{i} \to U_{i}

({\bar{x}}_{1}, . . ., {\bar{x}}_{n - 1}, {\bar{x}}_{n}, {\bar{y}}_{1}, . . ., {\bar{y}}_{n - 1}) \mapsto ({\bar{x}}_{1}, . . ., {\bar{x}}_{n - 1}, \frac{{\bar{y}}_{i}}{{\bar{x}}_{i}^{a_{i}}})

is the inverse map of the restriction of

\tilde{φ}

U_{i}

. Hence

\tilde{φ}

is an isomorphism of affine schemes, so that it induces an isomorphism of groups

{\tilde{φ}}_{i}^{*} : H^{i} c (V \ φ (X)) ≃ H^{i} c (K^{n} \ X)

for all indices

i

. The restriction of

φ

X

\bar{φ} : X \to φ (X)

maps

u_{n}

(0, . . ., 0, u_{n}^{d}, . . ., 0)

. Hence

φ (X)

is a 1-dimensional affine space and

\bar{φ}

induces multiplication by

d

in cohomology with compact support. Now, as it is well-known,

H^{i} c (φ (X)) ≃ H^{i} c (X) ≃ {\begin{matrix} Z Z / r Z Z & for i = 2 \\ 0 & else \end{matrix}

Since

r

divides

d

, it follows that the induced maps

{\bar{φ}}_{i}^{*} : H^{i} c (φ (X)) ⟶ H^{i} c (X)

are all equal to the zero map. Thus

φ

gives rise to the following morphism of acyclic complexes:

\begin{matrix} Z Z / r Z Z \end{matrix}

\begin{matrix} ≀ | & f \end{matrix}

\begin{matrix} H^{2} c (V) & \to & H^{2} c (φ (X)) & \to & H^{3} c (V \ φ (X)) \end{matrix}

\begin{matrix}  \end{matrix}

\begin{matrix} ↓ & {\bar{φ}}_{2}^{*} ↓ 0 & ≀ | ↓ {\tilde{φ}}_{3}^{*} \end{matrix}

\begin{matrix}  \end{matrix}

\begin{matrix} H^{2} c (K^{n}) & \to & H^{2} c (X) & \to & H^{3} c (K^{n} \ X) \end{matrix}

\begin{matrix} ≀ | & g \end{matrix}

\begin{matrix} Z Z / r Z Z \end{matrix}

Note that

{\tilde{φ}}_{3}^{*} f = g 0 = 0

is the zero map, so that

{\tilde{φ}}_{3}^{*} f

is not injective. Since

{\tilde{φ}}_{3}^{*}

is an isomorphism, it follows that

f

is not injective. This implies ( 9 ) and completes the proof. References

M. Barile, Almost set-theoretic complete intersections in characteristic zero. Preprint (2005). arXiv:math.AG/0504052
M. Barile, A note on Veronese varieties. Preprint (2005). arXiv:math.AG/0504055
M. Barile, G. Lyubeznik, Set-theoretic complete intersections in characteristic $p$ . Preprint (2004). To appear in: Proceedings of the American Mathematical Society.
M. Barile, M. Morales, A. Thoma, On Simplicial Toric Varieties Which Are Set-Theoretic Complete Intersections, J. Algebra 226 (2000), 880–892.
M. Barile, M. Morales, A. Thoma, Set-Theoretic Complete Intersections on Binomials, Proc. AMS 130 (2002), 1893–1903.
W. Bruns, R. Schwänzl, The number of equations defining a determinantal variety, Bull. London Math. Soc. 22 (1990), 439–445.
D. Eisenbud, E.G. Evans, jr. Every Algebraic Set in $n$ -Space is the Intersection of $n$ Hypersurfaces, Inv. Math. 19 (1973), 107–112.
J.S. Milne, “Étale cohomology”, Princeton University Press, Princeton, 1980.

On toric varieties of high arithmetical rank

Margherita Barile † † Partially supported by the Italian Ministry of Education, University and Research. Dipartimento di Matematica, Università di Bari,Via E. Orabona 4, 70125 Bari, Italy

November 27, 2006