Skip to main content
Internet Archive's 25th Anniversary Logo

Full text of "Asymptotic Form of Zero Energy Wave Functions in Supersymmetric Matrix Models"

See other formats


Asymptotic Form of Zero Energy Wave Functions in 
Supersymmetric Matrix Models 



J. Fr6hlich(«) , G.M. Graf(°) , D. Hasler^") , J. Hoppe^^'^) , S.-T. Yau^^) 

^"•^ Theoretische Physik, ETH-Honggerberg, CH-8093 Zurich 
^''^ M ax- Planck- Institut fiir Gravitationsphysik, Albert- Einstein- Institut, D-14473 Potsdam 
^'^^ Fachbereich Mathematik, TU Berlin, D- 10623 Berlin 
Department of Mathematics, Harvard University, Cambridge, MA 02138 

February 1, 2008 



Abstract 

We derive the power law decay, and asymptotic form, of SU(2) x Spin((i) invariant 
wave-functions satisfying Q^V = for ^-H = 2((i — 1) supercharges of reduced 
{d + 1) -dimensional supersymmetric SU(2) Yang Mills theory, resp. of the SU(2)- 
matrix model related to super membranes in d + 2 dimensions. 



1 Introduction 

It is generally believed that supersymmetric SU(A^) matrix models in = 9 dimensions 
admit exactly one normalizable zero-energy solution for each > 1, while they admit 
none for all other dimensions for which the models can be formulated, i.e., for = 2, 3, 5. 
For various approaches to this problem see e.g. |ll||-||T3[|. 

In this article, we would like to summarize (and slightly modify/extend) what is known 
about the behaviour of SU(2) zero-energy solutions far out at infinity in (and near) the 
space of configurations where the bosonic potential (the trace of all commutator-squares) 
vanishes. Based on some early 'negative' result concerning N = 2, d = 2 (that used 
rather different techniques/arguments; see |18|) we started our investigation of the 



asymptotic behaviour, in the fall of 1997, with a Hamiltonian Born-Oppenheimer analysis 
of that N = 2, d = 2 case. Some months later, we realized that the rather complicated 
Hamiltonian analysis (Halpern and Schwartz had, in the meantime, derived the form 
of the wave function for d = 9 near oo, by Hamiltonian Born-Oppenheimer methods) can 
be replaced by a simple first order analysis, using only the first order operators Q, and 
first order perturbation theory. One finds that asymptotically normalizable, SU(2) and 
SO{d) invariant, wave functions do not exist for d = 2,3, and 5, in contrast to ci = 9, 
where there is exactly one. 

We close these introductory words by recalling that the models discussed below arise in 
at least 3 somewhat different ways: As supersymmetric extensions of regulated membrane 



1 



theories md + 2 space-time dimensions [14, 18 , as reductions (to + 1 dimension) of ci + 1 



dimensional Super Yang Mills theories |T^-|T^, and, for ci = 9, as a description of the 
dynamics of D-0 branes in superstring theory, [|20| , ^ . In this physical interpretation, the 



existence of a normalizable zero-energy solution is an important consistency requirement. 

The paper is organized as follows. In Section ^ we recall the definition of the models, 
and in Section |] we state our main result about zero-modes. The proof is given in Section 
^ and Appendix 1. We suggest to skip Subsection |]5| and Appendix 1 at a first reading. 
As a warm-up the reader is advised to read Appendix 2, where a simpler model is treated 
by the same method. 

2 The models 

The configuration space of the bosonic degrees of freedom is X = M^*^ with coordinates 

g = (gl, . . . ,qd) = {qsA)s=i,... ,d ■ 

A=l,2,3 

To describe the fermionic degrees of freedom let, as a preliminary, 

Y = bU)a,f3=l,...,s, , {i = l,...,d), (1) 

be the real representation of smallest dimension, called Sd, of the Clifford algebra with 
d generators: {7'', 7*} = 25**1. On the representation space, Spin((i) is realized through 
matrices R G SO(srf), so that we may view 

Spm{d) ^ SO(s,) , (2) 

simply connected subgroup. We recall that 



Sd 



2[d/2] ^ d = 0, 1, 2 mod 
2[d/2]+i otherwise , 



where [■] denotes the integer part. We then consider the Clifford algebra with Sd generators 
and its irreducible representation on C = C^"'*'^^ . On C®^ the Clifford generators 



(61, . . . , = (6aA)a=l,.-,Sd 

A=l,2,3 



are defined, satisfying {QaA^Qps} = ^ap^AB- The Hilbert space, finally, is 

n = V{X,C^^). (3) 

There is a natural representation of SU(2) x Spin((i) 3 {U, R) on H. In fact, the group 
acts naturally on X through its representation SO (3) x SO{d) (which we also denote by 
{U, R)). On C^^ we have the representation 71 of Spin(srf) 9 R 

niRYQaAniR) = R^pQpA , (4) 

where R = R{R) is its SO(srf) representation. Through SO(srf) = Spin(srf)/Z2 and (0) we 
have 

Spin(rf) ^ Spin(srf) , (5) 



2 



and thus a representation IZ of Spin((i). The representation lA of SU(2) 3 U on C®^ is 
characterized by U{U)*QaAU{U) = UAB^aB- 

We shall now restrict to c? = 2, 3, 5, 9, where Sd = 2, 4, 8, 16, the reason being that 
in these cases 

Sd = 2{d-1), (6) 



whereas Sd is strictly larger otherwise. Eq. is essential for the algebra (|^ below 
The supercharges, acting on Ti, are given by the Sd hermitian operators 

Qf3 = Qa- {-hifs^t + ^ g; X g; 7|*„) , (/3 = i, . . . , s^) , 

where 7** = (l/2)(7*7* — 7*7*). These supercharges transform as scalars under SU(2) 
transformations generated by 

Jab = —K^lsAdsB — QsBdsA) — - {QaA^aB — ^aB^aA) = LaB + Mab , 

resp. as vectors in W'' under Spin((i) transformation generated by 



Jst = -Kqs ■ - g; ■ V,) - - e«7f^e^ = i^t + u^t . 

The anticommutation relations of the supercharges are 

{Qai Q13] = ^apH + la^qtA^ABC-JBC ■ (7) 

Here, H is the Hamiltonian 

9 

s=l s<t 

which commutes with both J^b and Jst- The question we address is the possibility of a 
normalizable state ip & Ti with zero energy, i.e., with Hip = 0, which is a singlet w.r.t. 
both SU(2) and Spin((i). Note that on SU(2) invariant states H = 2(5^ > and in fact 
the energy spectrum is ([|T^) cr{H) = [0, oo). Equivalently, we look for zero-modes 

Qp^ = 0, (/3 = 1,... ,Sd). 



3 Results 

The potential X]s<t(^* ^ vanishes on the manifold 

Qs = reEs 

with r > and e ^ = Yls ~ ^- The dimension of the manifold is 1 + 2 + (ci — 1) = 
3d — 2{d — 1). Points in a conical neighborhood of the manifold can be expressed in terms 
of tubular (or "end-point") coordinates |2^ 

Qs = reEs + r-^'^Vs (9) 



3 



with 



f,-e = 0, ysEs = Q. (10) 

A prefactor has been put exphcitely in front of the transversal coordinates yg-, so as to 
anticipate the length scale r~^/^ of the ground state. The change 

{e,E,y)^{-e,-E,y) (11) 

does not affect g^- Rather than identifying the two coordinates for g^, we shall look for 
states which are even under the antipode map (PH). 

We can now describe the structure of a putative ground state. 

Theorem Consider the equations Qpip = for a formal power series solution nearr = oo 
of the form 

oo 

^ = r~-Y^r'-2''^k, (12) 

where: ipk = V'fc(^^, -E", y) is square integrable w.r.t. dedEdy; 
ipk is SU(2) X Spin((i) invariant; 

Then, up to linear combinations, 

• d=9: The solution is unique, and n = 6; 

• d=5: There are three solutions with k, = —1 and one with k = 3; 

• d=3: There are two solutions with k = 0; 

• d=2: There are no solutions. 

All solutions are even under the antipode map (|77|j, 

^fc(e, E, y) = ipki-e, -E, y) , 
except for the state d = 5, k, = 3, which is odd. 

Remarks 1. The equation QfBip = can be viewed as an ordinary differential equation 
in 2; = r^/^ for a function taking values in L^(de d-E dy, C®^) (see eq. ( p^ below). It turns 
out that 2 = 00 is a singular point of the second kind [^. In such a situation the series 
(|I2p is typically asymptotic to a true solution, but not convergent. 

2. The integration measure is dg = dr ■ r^de ■ r'^~^dE ■ r~^''^'^'^~^^dy = r^dr dedE dy. The 
wave function ( [T^ ) is square integrable at infinity if J°° dr r^(r~'')^ < 00, i.e., if k > 3/2. 
The theorem is consistent with the statement according to which only for d = 9 a. (unique) 
normalizable ground state for (^) (which would have to be even) is possible. 

3. Note that the connection of matrix models with supergravity requires the zero-energy 
solutions to be Spin((i) singlets only for d = 9. 

The case d = 2 can be dealt with immediately. We may assume 7^ = as, 7^ = cxi 
(Pauli matrices), so that 

M12 = ^QlA<S>2A , 



4 



with commuting terms. Since, for each A = 1,2,3, (61^62^)^ = —1/4, we see that M12 
has spectrum in Z/2 + 1/4. Given that L12 has spectrum Z, no state with Juip = is 
possible. We mention |l[] that, more generally, for d = 2 no normalizable SU(2) invariant 
ground state exists. 

The proof of the theorem will thus deal with d = 9,5,3 only. 

4 Proof 

We shall first derive the power series expansion of the supercharges Qp. To this end we 
note that 

= r'/'{6,,-E,Et){SAB-eAeB)j^ (13) 
oqtA oysB 

d 1 d 

+r"\eAEt{r— + -y.s- ) + lesEtLBA + ie^^si^st] + Oir'^'^) , 

dr 2 dysB 

with the remainder not containing derivatives w.r.t. r (see Appendix 1 for derivation). 
This yields 

= r'l^Ql + r-\Q\r^^ + Q\) + r-^/^Q^ + . . . (14) 
with r-independent operators 

d 

Ql = -iQaAliai^st - EsEt){6AB - eAeB)T, + ■ (e x yt)Esi± , 

oysB 

Ql = -i{Qa ■ ehif^Et , 

i d 1 ^ 

Qp = QaAlifsi'^BEtlBA + eAEsL^t - - CAEtysBT. — ) + 7;^a ■ ills X ythfa ■ 

^ oysB ^ 

The explicit expressions of QJ^, (n > 2) will not be needed. We then equate coefficients 
of powers of r~^/^ in the equation Qfs-ip = with the result 

QjVn+ (-(/t+^(^-l))Qj + (5y^n-l + (5?^n-2 + ...+Q;^^0 = 0, 

{n = 0,1,...) .(15) 

4.1 The equation at n = 

The equation at n = 0, 

Qi^o = , (16) 
admits precisely the (not necessarily SU(2) x Spin((i) invariant) solutions 

Me,E,y)=e~^^y^'/'\F{E,e)) , (17) 

(with y restricted to ([T0|) ), where the fermionic states \F{E,e)) can be described as 
follows: Let n± be two complex vectors satisfying n+ ■ n_ = 1, e x nj. = ^in± (and hence 



5 



n± - n± = 0, n+ X ?2_ = — ie ). For any vector v G Mf'^ we may introduce Q{v) = QaVa, as 
well as fermionic operators Q{v) ■ n± satisfying canonical anticommutation relations: 

[Q{u) ■ n+, 9(f) ■ n_} = UaVa , ■ "^i) ©("w) • "^^i} = . 

Then, \F{E,e)) is required to obey 

e(f ) ■ n±\F{E, e)) = for ^,7"^ = ±f . (18) 
To prove the above, let us note that 



{g° , Q° } = 5apH' + ^ipEtSABcMABec , (19) 

^ 9_ 

dysA dytB 



= [-i6st - E,Et){6AB - eACB) + ^^'] + i^^T.^ge ■ (9^ x 6^) 



By contracting eq. ([19|) against 5^/3, resp. ■JapEt we see that the equations (|1^) are 
equivalent to the pair of equations 

i/Vo = , EABcMAsec^Po = . (20) 

Here, H'^ is a harmonic oscillator in 2((i — 1) degrees of freedom, with orbital ground state 
wave function e~^=^»/^ and energy 2(d — 1). On the other hand, 

= -E,7^^((e,-n+)(e^-n_)-(0„-n_)(e^-n+)) 

= -Sd + 2P+5(e„-n_)(e;3-n+) + 2P-^(e„-n+)(e/3-ri_) , (21) 

where we used the spectral decomposition EsY = — P ■ In view of (1), the equation 
H^ipo = is fulfilled iff the fermionic state is annihilated by the last two positive terms 
in (0), i.e., if (|T8|) holds. The second equation ( pOf ) is now also satisfied, since 



^SABcMABec = -^e ■ (6^ x e„) 

= ^((e„ ■ n+)(e„ ■ n_) - (e„ ■ n_)(e, ■ n+)) 

= P,-^(e,-n+)(e^-n_)-P+,(e,-n_)(e^-n+) (22) 

annihilates \F{E,e)). 



4.2 SU(2) X Spin((i) invariant states 

We recall that the representation 7l[-] of Spin(c/) on H is (7l[R]ip){q) = lZ{R){jp{R^^q)), 
where TZ{R) acts on C^^. Similarly for SU(2). The invariant solutions among (|T^) are 
thus those which satisfy 

U{U)\F{E,e)) = \F{E,Ue)) , 7^(P)|P(E, e)) = \F{RE,e)) , (23) 

for (U,R) G SU(2) X Spin((i). These states are in bijective correspondence to states 
invariant under the 'little group' (U,R) G U(l) x Spin((i — 1), i.e., to states \F{E,e)) 
satisfying 

U{U)\F{E,e)) = \F{E,e)) , 7^(P)|P(E, e)) = \F{E,e)) , (24) 



6 



for some arbitrary but fixed {E, e) and all U, R with Ue = e, RE = E. The first relation 
holds on all of (|1^). In fact the generator (|2^ ) of the group U{U) of rotations U about 
e annihilates \F{E,e)), as we just saw. To discuss the second relation (|^) we note that 
the generators of Spin(o? — 1) (i.e., of the fermionic rotations about E), are MgtUsVt with 
UsEs = VsEs = 0. We write M^t = + M^, where 

Mi = -(i/2)(e, ■ n+h'MQ^ ■ n-) , mJ, = -(i/4)(e, ■ e)^%{Qp ■ e ) , (25) 



and remark that, by a computation similar to (22), M^UgVt annihilates \F[E,e)). As a 
result, we may study the representation TZ of the group Spin((i — 1) through its embedding 
in the Clifford algebra generated by the 9^ ■ e. 

The operators ■ e leave the space (|TBp invariant and act irreducibly on it. That 
space is thus isomorphic to C, and Spin(sd) acts according to (|D (with QaA replaced by 
0Q ■ e). This representation decomposes (see e.g. [^]) as 

C = (2(^''/2)-i)^ ^ ^2^sj2)-iy (26) 

w.r.t. the subspaces where 9 = 2'^'*/^ na=i ©0'^^= +1; resp. —1. The embedding (|^) and 
the corresponding branching of the representation (but not the statement of the theorem!) 
depend on the choice of the 7-matrices. In order to select a definite embedding, let 

with P, (j = 1, . . . , (i — 2) purely imaginary, antisymmetric, and {P, F^} = 25jfcIsd/2- 
Then ([26| ) branches as (see resp. [jl2|, |13]) 



(44 © 84) © 128 , (c^ = 9) , 

C = <( (5 © 1 © 1 © 1) © (4 © 4) , (c/ = 5) , (28) 

2 ©(!©!), (c^ = 3), 

when viewed as a representation of Spin((i). (The choice 7^^ = Ra'aYa'i3'^i3'i3 "with 
R G O(srf), det i? = —1 would have inverted the branching of the representations on 
the r.h.s. of (|26|) ). The case d = 3 deserves a remark, as there are additional inequivalent 
embeddings Spin((i = 3) ^ Spin(sd = 4), and one has to consider the one appropri- 
ate to (^). In fact R G Spin(3) = SU(2) acts in the fundamental representation on 
C^, the irreducible representation space of the complex Clifford algebra with 3 genera- 
tors. The real representation ( p7D is obtained by joining two complex representations, 
followed by an appropriate change T of basis. The embedding (^) is thus realized through 
R 1-^ T^^{R © 12)7" and the embedding su(2)j^ ^ so(4)j. = su(2)j^ © su(2)j^ is equivalent 

to M 1-^ (m, 0). 

The further branching Spin((i) ^ Spin((i — 1) yields 

:i © 8v © 35v) © (28 © 56,) © (8s © 8c © 56s © 56c) , {d - 1 = 8) , 

C= <( l©l©l©(l©4)©(2+©2_)©(2+©2_) , (d-l = 4), (29) 

^li © l_i) © lo © lo , {d-l = 2). 



7 



The content of invariant states stated in the theorem is now manifest. One should notice 
that for d = 3 the httle group U(l) is abehan and the singlets l-ti do not correspond 
to invariant states. For later use we also retain the fermionic Spin((i) representation to 
which the remaining singlets are associated, 

44 (rf = 9); 1,1,1,5 (rf = 5); 1,1 {(1 = 3), (30) 

together with the corresponding eigenvalue of B: 

e= 1 {d = 9); 1,1,1,1 id = 5); -1,-1 {d = 3) . (31) 



4.3 Even states 

It remains to check which of these states satisfy \F{—E, — e)) = \F{E,e)). Let us begin 
by noting that by (p3D 



\F{-E,-e)) = e'^-*«^^"sV*'^^'-^^^'"|F(E,e)) , 

where u G M^, resp. U E M.^ are unit vectors orthogonal to e, resp. E. The Spin((i) 
rotation can be factorized as e^M^tEsUt-K ^ ^iM^^EsUt-K^iMlEsUt-K _ ^YeXm that e'*^'' ^''^"^ 
\F{E,e)) = a\F{E,e)) with 

a= 1 (rf = 9); 1,1,1,-1 (rf = 5); 1,1 {d = 3) . (32) 

The operator represents a rotation R G Spin((i) with RE = —E in the representation 
(pOl). For d = 9 the latter can be realized on symmetric traceless tensors Tij, = 
1, . . . ,9), where the Spin(8)-singlet is EiEj — {l/9)6ij, implying a = 1. For d = 5, 
the last representation (|30D is just the vector representation, where a = —1. As the 
remaining cases are evident, eq. (0) is proven. A computation using (|27|) and, without 
loss E = (0, . . . , 0, 1), f/ = (0, . . . , 1, 0) shows 

e^'^^td-i-^\F{E, e)) = PJ et^®"'"+^'^®'"+"d/2'"'^"^®"+=<i/2'"+^^®°'""^^''/^|F(£', e}) 

a=l 
Sd/2 

= l[iQ^+s,/2-n+)iQa-n-)\FiE,e}) = \FiE,e}) , 

a=l 

Sd 



= (-l)^^/^en(0c.-n+)(e„+.,/2-n_)|F(E,e)) = |F(E,e-)) 



where we used (0) in the last step. Together with (|32[) this proves the statement of 
theorem concerning the invariance under (0). 



4.4 The equation at n > 



We next discuss the equations (|T5 
onto the states (|T^ 



with n > 1. Let Pq be the orthogonal projection 
i.e., onto the null space of Q^. We replace them with an equivalent 



8 



pair of equations, obtained by multiplication of (|T5D„+i with Pq, resp. of (|T5D„ with Q^, 
which is injective on the range of the complementary projection Pq = 1 — Pq: 

Po{-{k + ^n))Ql + Ql)PoiJn = -Po(QpoV'n + + ■■■ + Q^Vo) , 

(n = 0, 1, . . . ) , (33) 
(n = l,2,...) (34) 

(we used PqQ/j-Po = 0). Here, and until the end of this subsection, no summation over (3 
is understood. The equation (|33|) at n = reads 

PoQl^o = f^PoQl^o (= ^^Ql^o) • (35) 

We shall verify this by explicit computation later on. Since a similar issue will show 
up in solving the equation at > 0, let us also present a more general statement, 
whose proof is postponed to the next subsection. 

Lemma Let Tp he linear operators on the range of Pq, which transform as real spinors of 
Spm{d) and commute with the antipode map. Then, for each invariant state we have 

TpijjQ = nQ^ipo , (36) 

with K, depending only on the associated representation j^Mj). 

We now assume having solved the equations (^ ^) up to n — 1 for Spin((i) invariant 
ipi, . . . ipn-i (which is true for n — 1 = 0), and claim the same is possible for n. Since Q^^ 
is invertible on the range of Pq, eq. (|^)n determines Poipn uniquely. The fact that the 
solution so obtained is independent of (3 and is Spin(c?) invariant may deserve a comment, 
because the equivalence of the equations Q/^ip = and {Q/3)^ip = 0, which holds on (H), 
does not apply in the sense of formal power series (|T2]). Consider the expansion ([T^), i.e.. 



oo 



k=0 



as well as its formal square 



oo 

3l 



k=0 



Notice that {Q/bY is, by (0), independent of (3 and Spin((i) invariant as an operator on 
SU(2) invariant power series. Similarly, let [Qi3ip]k (given by the l.h.s. of (^)) and 
[{Qi3Y4']k be the coefficients of the corresponding series. By induction assumption we 
have [Qf34']k = for A; = 0, . . . , n — 1. Since QpiQpij) = {QpY'ip, we obtain 

n „ 

[(Q/3)V]n = J]gJ[g/3V^]„-fc-(«: + -n-2)gj[Q^V^]._i = g°[g^^]„, 

A:=0 

[(g/3)V]n = (gS)'V'n + , 



where ipn-i (determined by ipO: ■ ■ -i^n-i) has the desired properties. The equation (0)n, 
i.e., Q°[Q/3^]n = is thus equivalent to {Q^^Yipn = — ^n-i, which exhibits the claim. 

On the other hand, invariance requires Potpn to be a linear combination of invariant 
singlets. For the ansatz Po^n = Xn'4'oy sq. (|33D^ reads 

^nXnQliJo = -PoiQlPoi^n + + • • • + Q^+Vo) , 

because of (0). Again, by the lemma, this holds true for suitable A„. Indeed, this solution 
for Poipn is the only one. 

4.5 Proof of the lemma 

The vectors Tpipo, (/3 = 1, . . . ,Sd) transform under Spin((i) as real spinors, although they 
might be linearly dependent. By reducing matters to the little group as before, any 
representation of that sort is specified by the values |F^(i^^,e)) of its states (see (|l3)) at 
one point {E, e), which are required to satisfy 

Rp^{R)\F"{E,e}) =n{R)\F^{E,e}) 

for R with RE = E. Pretending the states \F^{E,e)) to be linearly independent, the 
branching Spin((i) ^ Spin((i — 1) yields 

16 = 8,©8e id = 9); 4 © 4 = (2+ © 2_) © (2+ © 2_) {d = 5) ; 

2 ©2 = (li©l_i)©(li©l^i) {d = 3). 

For d = 9,5 each term on the r.h.s. occurs as often as in (^), and ipo can indeed be 
chosen so that the Sd vectors Qjjipo are independent. Not so in the last case, where the 
vectors Tpipo just belong to li © l_i. We continue the discussion for different values of d 
separately. 

• d = 9. Any linear transformation K commuting with a Spin(9) representation as 
above is thus of the form K = Ks® Kc- If K also commutes with the antipode map, then 

= = Applying this to the representation Q^Vo and to the map K : Q^V^o ^ TpipQ 
yields the claim. 

• d = 5. Let us regroup (2+ © 2_) © (2+ © 2_) = (2+ ® I2) © (2_ © I2). Then any map 
K commuting with the representation is of the form 

K = {1®K+)®{1®K_) , 

where K_ is conjugate to if K commutes with the antipode map. This allows for a 
four dimensional space of such maps K. To proceed further we shall again assume that 
= (0, . . . , 0, 1) and introduce creation operators 

a; = -^[(e„-e) +i(e,+4-e)] , (a = l,...4) 

which then define a vacuum through aQ,|0) = 0. We next choose an orthonormal basis 
{ipl, ... , V'o} 4-dimensional subspace of singlets in the range of Pq by specifying 



10 



the values of the corresponding fermionic parts (see (|T^) at {E,e): 



\F,\E,e)) = -L(|0)-ata;aK|0)), 
\F^{E,e)) = ^f>>;|0) = ^(7^fW(e«-el(e;3-e-)|Fo^(E,e)), (^ = 1,2,3), 

where 

with S G 0(4) and det S = — 1. Note that tpQ is the singlet belonging to the 5-dimensional 
fermionic representation of Spin(5). One can verify that the four maps 



Ql%, (^ = 1,2,3), 

l'p^E,Qi^^,, (z = 4), 



besides being of the kind just discussed, are linearly independent. Therefore any map 
K of the above form is a linear combination thereof. In particular this applies, for any 

{x,X4) G R^+^, to the map K : Qj^ipl i— > XiTpipQ + x^'j^^^EtTaipQ, hence 

This defines a linear map n : (x, X4) iy,yi) on R^+^. We claim that 

K : {Rx, X4) t-> {Ry, 2/4) (37) 

for R G S0(3), which implies n = diag(Ki = 1^2 = 1^3, 1^4,) and hence (|3^) . Eq. (|37D can be 
proven using RijipQ = 'R-x^'l for 71 G Spin(8) projecting to i? G Spin(3) C Spin(5) ^ S0(8). 
This in turn follows from (^) and from TZipQ = ipo- 
• d = 3. Analogously to = 9. 

4.6 Determination of k, 

Since Jab'iPo = Jsti'o = we may replace by 

Ql = QaAlip {-CBEtMBA ' CAEsM.t - ^ eAEty^B^) + ■ (ys X ythga ■ (38) 

^ oysB ^ 

We discuss the contributions to ( P^D of these four terms separately, 
i) With 



^bMba = ((0/3 ■ e)e^A - Q(iA{Q(3 ■ e)) 



we find 



O^aCbMba = i{{Qa-n+){Qp-n-) + {Qa-n~){Qp-n+)){Qp-e) 

PoQaAeBMBAi'O = i(0a " 6 )V'0 , 



11 



since only the term with (3 = a survives the projection Pq. Hence 

-PoQaA-fipeBEtMBA^o = Qli^o (39) 



contributes 1 to n. 
ii) Similarly, 



where MJ^ is given in (|3T|) . For the r.h.s. we then claim 

-(©a ■ eHpEsMl^o = k'QI^Po (40) 

with 

(rf = 9) , 

«:'=<[ 0,0,0,4, (rf = 5), (41) 

,0, (d = 3). 

This is clear in the cases where the representation in (PO) is already a singlet, i.e., when 
k' = 0. To prove the two remaining cases we first establish 

-{Q^ ■ ehi^EMl^, = -^Yaf,E,[Q^ . e , mJ,m1]^o - i^^(©a ■ ehlf^E.^P^ , (42) 
or the equivalent equation obtained by multiplication from the right with Eu'y^'- 

-(e, ■ e )(7V)./5i?«i?sM]l ^0 = ■ e , mImI]^,, - i^^(0/3 ■ e )^o • (43) 

To this end we note that, by the invariance of ipQ, its fermionic part \F{E, e)) a,t E E S'^^^ 
is invariant under rotations of Spin((i) leaving E fixed: {d^s — EuEg) MsviS^t — EyEt)ipo = 0, 
i.e., 

{mIe^E, + MlE^EMo = Miv^o • (44) 
Using 7*7" = —7"* + and the observation just made we rewrite the l.h.s. of (^) as 

-(e,-e)(7V)a/3K^.M]l^o = {ea-eha^E^EMli^o 

= \{Qa-e)^ll{EuEsMi-EtEsMl)^o 



The commutation relation 

i[e„-e,Mi] = ^7:^^0/3 -e") 
follows from (0) or by direct computation. It implies 



i[e« ■ e , mImI] = llaU^p ■ e , Ml} = 7^^(6/3 ■ e )Mi - ^1:^^^^ ■ e , M^] 

12 



Solving for the first term on tlie r.h.s. proves PB| ) and lience (0). Let us now note 
that for (i = 9 the fermionic part of ipo, resp. of (9^ ■ e)ipo belongs to the 44, resp. 128 
representation of Spin (9) (see (PH])). Eq. then implies 



-(6, ■ e)ji^EsMiiJo = (C(44) - C(128) + 9)QJ^o = 9QJ^o , 



where we used the values [^] of the Casimir: C(44) = C(128) = 18. In the case d = 5 
the fermionic part of ipo, resp. of (9q, ■ e)'4'o belongs to the representation 5, resp. 4 © 4. 
We conclude that 

-(e„ ■ ehipEsMl^Po = (Cib) - C(4) + ^)gJV^o = 4QjV^o , 

given that C(5) = 4, ^(4) = 5/2. 

We remark that the proof of (^Tj) can be shortened by using the lemma, according to 
which (^Of ) holds true for some k'. Thus, contracting with Qpipo and summing over (3, we 
find 

-/€'(Vo, QlQl^o) = -i(^o, (67 ■ e)7"/3^n(e„ ■ e)^ipEsM!,^lJo) 

= 2(^0, mI{mIEuEs + MlE,Et)^o) = 2(^-0, mIMM • 

In the step before last we relabeled indices in half the expression; in the last one we used 
(HI). Using Q^Qp = -Sd/2 we obtain {sd/2)K' = 2 • 2 • C, i.e., k' = 8C/sd, where C is the 
Casimir in the representation (ppj). The above values of C(44) {d = 9) and of C(5) {d = 5) 
yield again (|41|). 

iii) Using de~^ ^^/dy = —ye~^ we get 



1 d 



V'o = —ysBVsB'ipo = ^{y^B - ^)^o -\-2{d- 1)^0 , (45) 



- ^^^^ 2 2 _ . 

where the sum, consisting of second Hermite functions, is annihilated by Pq. 
iv) The last term in (|38D, when acting on ipo, is similarly annihilated by Pq. 

Collecting terms (|39| , ^51) we find 

^ [6, (f/ = 9), 

«:=l + /€'--(rf-l)= J -1,-1,-1,3, (rf = 5). 



2 



0,0, (rf = 3) 



Appendix 1 



To prove (|l^) we shall compute the partial derivatives in 



d 



dr 



d 



+ 



dEs d dysB d 



+ 



dqtA dqtA dr dqtA des dqtA dEs dqtA dyss 



(46) 



13 



We regard r, e, E, y as functions of q defined by e ^ = J2s -^s = 1 (P, 0) and solve 
for their differentials by taking different contractions of 



Using that 



dqtA = {eAEt - ^/^?/tA)dr + rEtdcA + rcA^Et + r ^^MT/tA 



e^dytA + ytA<^eA = , EtdytA + VtA^Et = 

CAdCA = , Etd^i = , 



5 



the contractions are: 

CAEtdqtA = dr 

{Sba - eBeA)EtdqtA = rdes - r'^^'^ytAdEt , (47) 

e^(5,t - EsEt)dqtA = rdE^ - r~^'^ysAdeA , (48) 

{Sba - eBeA){5st - EsEt)dqtA = --r'^^'^yssdr + r'^^'^{dysB + ^BysAdcA + E.ytBdEt) . 
We solve (0, ||) for de^, dE,: 

dr = CAEtdqtA , 

dcij = {m~^)Bc{r~^{ScA - eceA)Et + r"^/^|/iC'eA)dgM 

= {r-\6BA - 6^6^)^^ + 0{r-''/^))dqtA , 

dE, = (M-i),„(r-i(5,t - E^Et)eA + r-^/^^.A^^Odg^ 

= {r~\5st - ^s^t)eA + 0(r-^/2))^^^^ ^ 

d^/sB = [^^^^^((^BA - eBeA){Sst - EsEt) + -r~^eAEtysB]dqtA - esysAdcA - E^ytBdEt , 

where m, M are the matrices 

rriAB = Sab - r'^ytAVtB , Mst = Sst - r'^ysAUtA ■ 
We can now read off the partial derivatives appearing in (^) and obtain 

^ = r'/\6st - EsEt){6AB - eACB)^ + r-^[eAEt{r— + -y.s- )] 

dqtA dysB dr 2 dysB 

d d 
+r'^{.5Ac - (^Aec)Et{5cBT\ ^BysC 



OeB OVsB 

+r'\5ut - E^Et)eA{5usj7w- - E^yuB^ ) + 0{r-''/^) , (49) 

dEs dysB 

with the remainder not containing derivatives w.r.t. r. Finally, we insert this expression 
into 

d d 

iJ^BA — qsBT\ QsA 



dqsA oqsB 

d 

[{5 AC - (iAec)ysB - {^BC - eBec)ysA 



dysc 

/<: d d . d d 
+eB[0ACT, ecysAT. — ) - eAyoBCT. ecVsB 



oec oysc oec oysc 

14 



(with no higher order corrections, as Lab is of exact order 0{r^)) and then into 

d d 

ir'^esEtLBA = r'^^iSAc - eAec)Et{5cBi^ (^bVsC^ ) • 

ocb oysB 

Similarly, we have 

d d 
ir~^eAE,Lst = r"^(5„t - EuEt)eA{SusT7T=r - EsVuB- ^ 



dEs dysB 
Together with (H), this proves (p^). 



Appendix 2 

Consider 



H={-d.-^-d,' + xVn+i 4 4 ) ' (50) 



which is the square of 

dy + xy 



Q 



dy - xy -da: 



Just as in (P), the bosonic potential V (= x'^y'^) is non- negative, but vanishing in regions 
of the configuration space that extend to infinity (causing the classical partition function 
to diverge). Quantum-mechanically, just as in (^, the bosonic system is stabilized by the 
zero point energy of fluctuations transverse to the flat directions; the fermionic matrix 
part in (|SD|) exactly cancels this effect, causing the spectrum to cover the whole positive 
real axis [0. As simple as it is, it has remained an open question (for now more than 10 
years) whether ( pH ) admits a normalizable zero energy solution, or not. The argument, 
derived in a few lines below, gives 'no' as an answer and provides the simplest illustration 
of our method: as x — +00, Q'^ = has two approximate solutions, 

^+ = e-^^ J ^ and ^_ = ( ) ' ^^^^ 

the first of which should be chosen for \t'o in the asymptotic expansions 

^ = a;-"(*o + ^i + ...) • (52) 
In this simple example, the sum Q = Yl'^=o Q^^^ terminates after the first two terms, and 



yields (as already anticipated, cp. (^ 



dy- xy 



dy + xy 



15 



and 



Multiplying (^) by \1'J and integrating over y one sees that 

/ + 00 ^ 2 

-oo 

has to vanish, implying in particular 

/+00 / 2 \ 
1 

~4 ' 

which proves that ( ^OD does not admit any square-integrable solution of the form (|52|). 
A different approach has recently been undertaken by Avramidi p6|. Finally note that, 
calculating the '^n>o from (|53|), yields the asymptotic expansion, x +oo. 




where fo = I = go, fi = = gi, and the 5'n(s) are the (unique) polynomial solutions 

n n 

fn{s) = ^fn,iS\ gn{s) = ^ gn,iS' 

i=0 1=0 

of 

2s/; + (1 - 2s)/„ = {l-2s~Qn)g^ + Asg'^, 
, /3 s 3n\ ^ ^, 

Acknowledgments. We thank A. Alekseev, I. Avramidi, V. Bach, F. Finster, H. Nico- 
lai, C. Schweigert, R. Suter, P. Yi for useful discussions. We also thank the following 
institutions for support: the Albert Einstein Institute, the Fields Institute, the Erwin 
Schrodinger Institute, the Institute for Theoretical Physics of ETH, the Mathematics 
Department of Harvard University, the Deutsche Forschungsgemeinschaft. 



= 

K = 



References 

[1] J. Frohlich, J. Hoppe, On zero-mass ground states in super-membrane matrix mod- 
els. Comm. Math. Phys. 191, 613-626 (1998); piepth/9701119 . 

[2] P. Yi, Witten index and threshold bound states of D-branes, Nucl. Phys. B505, 
307-318 (1997); |hep-th/9704098 . 



16 



S. Sethi, M. Stern, D-Brane bound state redux. Comm. Math. Phys. 194, 675-705 
(1998); |hep-th/ 9705046. 



M. Porrati, A. Rozenberg, Bound states at threshold in supersymmetric quantum 
mechanics, Nucl. Phys. B515, 184-202 (1998); ^iep-th/9708TT9| . 



J. Hoppe, On the construction of zero energy states in supersymmetric matrix models 
I, II, III, |hep-th/970"9l32| , |liep-th/970"92l7| , |hep-th/97lT033| . 



M.B. Green, M. Gutperle, D-particle bound states and the D-instanton measure, 
JHEP 9801 (1998); |hep-th/97in07| . 

J. Hoppe, S.-T. Yau, Absence of zero energy states in reduced SU(N) 3d supersym- 
metric Yang Mills theory, |hep-th/9711169| . 

M.B. Halpern, C. Schwartz, Asymptotic search for ground states of SU(2) matrix 



theory. Int. J. Mod. Phys. A13, 4367-4408 (1998); |hep-th/97 12133 



W. Krauth, H. Nicolai, M. Staudacher, Monte Carlo approach to M-theory, Phys. 
Lett. B431, 31-41 (1998); |hep-th/9803Tr7| . 



G. Moore, N. Nekrasov, S. Shatashvili, D-particle bound states and generalized 
instantons, [hep-th/9803265 . 



A. Konechny, On asymptotic Hamiltonian for SU(N) matrix theory, JHEP 9810 
(1998); [hep-th/9805046| . 

G.M. Graf, J. Hoppe, Asymptotic ground state for 10 dimensional reduced super- 
symmetric SU(2) Yang Mills theory, |hep-th/9805080 . 



J. Hoppe, S.-T. Yau, Absence of zero energy states in the simplest d = 3 {d = 5?) 
matrix models, [hep-th/9806152 . 



J. Goldstone, unpublished. 

J. Hoppe, Quantum theory of a massless relativistic surface, MIT Ph.D. Thesis 
(1982); Proceedings of the workshop Constraints theory and relativistic dynamics, 
World Scientific (1987). 

M. Claudson, M. Halpern, Supersymmetric ground state wave functions, Nucl. Phys. 
B250, 689-715 (1985). 

R. Flume, On quantum mechanics with extended supersymmetry and nonabelian 
gauge constraints, Ann. Phys. 164, 189-220 (1985). 

M. Baake, P. Reinicke, V. Rittenberg, Fierz identities for real Clifford algebras and 
the number of supercharges, J. Math. Phys. 26, 1070-1071 (1985). 

B. de Wit, J. Hoppe, H. Nicolai, On the quantum mechanics of supermembranes, 
Nucl. Phys. B305, 545-581 (1988). 

B. de Wit, M. Liischer, H. Nicolai, The supermembrane is unstable, Nucl. Phys. 
B320, 135-159 (1989). 



17 



[20] E. Witten, Bound states of strings and p-branes. Nuclear Phys. B460, 335-350 
(1996); |hep-th/9510135| . 

[21] T. Banks, W. Fischler, S.H. Shenker, L. Susskind, M theory as a matrix model: a 
conjecture, Phys. Rev. D55, 5112-5128 (1997); |hep-th/ 96 10043 . 

[22] E.A. Coddington, N. Levinson, Theory of ordinary differential equations, Krieger 

(1987) . 

[23] R.S. Palais, C. Terng, Critical point theory and suhmanifold geometry, Springer 

(1988) . 

[24] B. Simon, Representations of finite and compact groups, American Mathematical 
Society (1996). 

[25] W.G. McKay, J. Patera, Tables of dimensions, indices, and branching rules for rep- 
resentations of simple Lie algebras, Dekker (1981). 

[26] I. Avramidi, On strict positivity of some matrix-valued differential operators, (work 
in progress). 



18