IC/8U/3

REFERENCE
INTERNATIONAL CENTRE FOR
THEORETICAL PHYSICS

SPONTANEOUS COMPACTIFICATION

IN QUANTUM KALUZA-KLEIN THEORIES

M.H. Sarmadi

INTERNATIONAL
ATOMIC ENERGY
AGENCY

UNITED NATIONS
EDUCATIONAL,
SCIENTIFIC
AND CULTURAL
ORGANIZATION
1984 MIRAMARE-TRtESTE
 ic/614/3
I. Introduction
The Kaluaa-Klein idea [1,2] of enlarging U-dimensio-nal space-time to
International Atonic Energy Agency explain internal symmetries is an attractive means of unifying gravity with other
and known forces. In this approach the ground state is taken to be the product of
United H: . Ions Educational Scientific and Cultural Organization U-dir.ensional space-time and soir.e compact homogenous space whose isometry is the
gauge symmetry of non-gravitational interactions. It is being realized that the
INTERNATIONAL CENTRE FOP. THEORETICAL PHYSICS proper approach to supergravity is hy means of spontaneous coir.pactification from
higher dimensions in which these higher dimensions are not just a technical tool
of reduction for obtaining the desired model in lower dimension but are rather
believed to have physical significance. In most of the models which incorporate
these (i.e. supergravity and Kaluza-Klein compactification) one is faced with
SPONTANEOUS COMVACTIF I CV-TIOE TK QUANTUM FAUiZA-KLEIN THEORIES the difficulty of predicting a very large cosraologicalconstant for ^-dimensional
space-time. For instance the N = l, eleven dimensional supergravity admits solutions
of the form Ads x = where Ads (anti de-sitter) and the compact internal space
M.H. Earmadi have sises of the same order (see [3] for references). It is conceivable that
International Centre for Theoretical Physics, Trieste, Italy, quantum effects may have a bearing on this problem, i.e. including these may
enable It-dimensional Minkowski space to be a ground state of such models.

Recently, it was shown by Candelas and Weinberg [I*] that the quantum
fluctuations in a matter coupled gravity theory consisting of a suitable mixture
ABSTRACT
of scalars and spinors in h+N dimensions make spontaneous compactification to the
We study one loop effective action for pure gravity in higner odd product of ^-dimensional Minkowski space and the N-dimensional sphere possible for
dimensional spaces with a cosmological constant. We develop a method for N =5,9,13 and If- Their neglect of graviton loops was justified by taking the
computing the effective action for backgrounds -which are a product of lt-dimen3iona2 number of matter fields to he large. The inclusion of these loops becomes neces-
space-time and an odd-dimensional sphere S . For N=l in harmonic gauge the y&ry in supergravity where the number of matter fields is constrained by super -
potential has an unstable stationary point at which the one loop corrected symmetry.In fact, as we will see below, the contribution of graviton loops are
Newton constant has a wrong sign. Yor N > 1 in harmonic gauge there is an relatively large. Therefore, we think it is necessary to ask if it is possible to
imaginary contribution to the .effective potential fror- the Fa&leev-Popov ghost. get spontaneous compactification to Minkowski x S in a theory of gravity in U+R
The results for spheres up to N=17 are presented. dimensions without any matter fields. Our approach is to see if this space (i.e.
Minkowski x S ) minimizes the one loop effective action (however see [5] for a
different approach). In addition we require the one loop Newton constant to fee
positive for the minimum to be acceptable. Intuitively, it may be puzzling how a

MIRAMARE - TRIESTE small fluctuation of order Planck constant could have the enormous effect of curling up

January I98I1 N of the dimensions to such a small distance as the Planck length. As we will
see below the ^-dimensional cosmological constant should be adjusted. It is
proportional to the square of the Planck mass. All the dimensional parameters are
given in terms of the only scale in the action which is the Planck length.

We shall restrict our study to the case of S = odd dimensions. For this
case because the total dimension D = k + N is odd there are no logarithmic Ultra-
violet divergences at one loop. Therefore by using dimensional regularization one

To be submitted for publication.

-2-
obtains a one-loop effective action free of J ; tra-Violet divergences.

In the next section we work out the relevant operators and their eigen-
values. In Sees, III and IV we give the expressions for the effective potential
and the one lot corrected Nevton constant-Section V contains the results and
discussion. (h)

II. THE RELEVANT OPERATORS

We start with the Einstein-Hilbert action in D = k+H dimensions including where f^ (g,h) = 0 is a background gauge fixing condition and "FT* i s t'rle
a cosmological constant A : Faddeev-Popov matrix obtained by considering the transformation of the h under
an infinitestimal coordinate transformation -
S = - \fz f (1)
(5)

where G~ is the D-dimensioal gravitational constant, R is the scalar curvature

a
nd g D =det C g ^ ) - We lookfor a ground state solution M x S where the ^-dimensional
space-time M is flat and the extra IJ dimensions are compactified to an N-dimensional
while the background is held fixed. The Legendre transform of W [ J , R ] gives the
sphere. Obviously there is no such solution at the classical level. However there effective action f e f f [|. I J:
is the possibility that the one loop e'fective action [6] :

(2) (6)

which includes the one loop term T7 may have such a solution.

For T^f we use the standard gauge invariant effective action [?]. One first
From (It) it, follows that P e f f satisfies the following functional equation:
splits the metric into :

(3) -J
where h
h is the fluctuation about a background metric 3KB . This "background
metric need not be a solution to the classical equations of motion. One then
defines the generating functional W [ j , g] ty:
Starting from this expression one can write a perturbative expansion for f [7].
Be Witt's gauge invariant effective action is obtained by setting g = g in
r_ff.[g,i]. At one loop it is given by ;
*) Our notation is the following : capital letters are used for D-dimensional space-
time indices, Greek indices for ^-dimensional space-time and the mid Latin alphabets exp
for S . The Riemann curvature is given by E =JBf,p „ -i^^y „+• - • - and the
Ricei tensor by R.^= R.rI, • For the covariant derivatives,jfj =(B |t i7.J.

(3)

-3- -k-
 (si**!

( £>*+ 7*
7*+ ^ ) 7. j do)

and

A K
*9 is the Faddeev-Popov ghost field and a is the radius of S . Functional
integration over h. and *£ gives "p* as the sum
We emphasize that this expansion is valid even when " S
(for details see Bef.[7]). 83Aa

For gauge fixing we use the harmonic gauge condition

(ID

(9) where C = 1 except for ghosts when C - -1 and C are the operators
appearing in (10). The mixing between h, + and hytf is a problem. Actually
for the calculation of the potential we can set 3^= "L, - It is only for
where ^ is the background covariant derivative. This choice of gauge is obtaining the induced gravitational term that we need L^ V ^ *?jii>
particularly convenient because as we will see the mixing terms are minimal. But to simplify this ve choose ft;>y = -fr K+d^v which will leave us with
In this gauge the quadratic part, which is needed to calculate (8) is given by: h and + mixing only. This is handled by writing :

$„ -
D«f
rur -'"-" f"*- », i f'

liere the imaginary partn are the price we pay for having operators linear in D
rather th.in quadratic. But because the two factors are complex conjugates of
each other, their totnl contribution to the potential is real.
 If the functional integral is performed in Euclidean space then the negative
sign in one of the two factors in (12) is dealt with by using Hawkins's prescription ii) Transverse vectors carry the representation (l,0,0,..v0) of S0(N) and
[8] of rotating the integration colour for the corresponding field from real to the representation (Al,O,...,0) of S0(H + l ) ;
f
*)
imaginary
To find ^determinant of the relevant operators in (10) we need the eigen-
values of the Laplacian acting on scalars, vectors and symmetric^second rank -> 3 , (16)
tensors on S H . For this we use the method of Kef.9. Writing 5 * softf)
the eigenvalues £ of -V* are given by the quadratic Casimir invariants C,,
of the representation as :
with the multiplicity

(17)
(13)

Using G e l ' f a n d - Z e t l i n n o t a t i o n we know t h a t t h e r e p r e s e n t a t i o n (m ,m^ ,.,. ,IR.)

of SO(£k + l ) i s contained in t h e r e p r e s e n t a t i o n (/,»A >•--»/l 1
k+l' iii) Transverse traceless symmetric tensors carry the representation
of S0{2H-2) i f and only i f j>( > m, > 4X > mt > --.> j ^ k i / • Thus on spheres
(2,0,0,.../)) of S0(l) and the representation (£,2,0,..^) of S0(H + l ) ;
of dimension M = 2k + 1 :

i ) s c a l a r s carry the r e p r e s e n t a t i o n ( 0 , 0 , . . . , 0 ) of SO(N) and the

representation ( i , 0 , . . . , 0 ) of SO (n + j _ ) ;
(18)
a*

with the multiplicity

with the multiplicity

(15)
re*)
•) This will introduce the factor (i) -^ where ^ (0), which is the regularized
dimension of the operator, is given by the s = 0 value of the generalized zeta S0(2k
For spheres of even dimensions S2 k the representation
function : 2
(m 1T r.2,...,r.k ) of S0(2k) is contained in the representation ( i ^ ^ , . . . ^ )
of S0(2k + 1) if and only if ^ > m^^l m£. . . > ^ > | m J . The eigenvalues
and multiplicities are the same as above.

III. THE POTENTIAL

To find out if Minkowski x S is a solution we need to calculate the

But as we will see below, I (s-n/2) 2_ T>i (Mj )*~ does not li.ive a pole as
effective action in (2) for the background metric
=
s -» 0 and n-» li. Therefore 5(^^ ^ because of the factor -.

-fl-
 V
(20)

125)
where g., is the metric for S H . The minimum of the effective potential

To compute the one loop potential V = - we use the eigenvalues

of the relevant operators in (10).Integrating over tne eigenvalues of d , after

(£1)
the analytic continuation of space-time dimension to n has been done, we arrive
at the form:

jAff fi L _t
Z(ir.)

will give the radius of S . Here GQ is the tree level Hewton constant
(26)
N-11
for the potential. Here k ^p-,i=0,l,2 refers to scalars, transverse vectors
and transverse symmetric traceless 2nd rank tensors on S' , and the !( which

G. = (22) can "be read from (10) are given in Table 1. First we work out the contribution
o f s s " scalar. (The contribution of vectors and tensors can be worked out similarly.)
The sum over I for this case starts from zero. However T)L which can be written
As we will see shortly, V (
(a,A) has the form: in the form:

(S3)

where we have defined ^he dimension!ess quantity x =\af. The effective <V.jr (27)
dimensional cosmological constant is given by the value of the effective potential
at its minimum. Thus for the four dimensional space to be flat, A should be vanishes when i ranges from -k to -1 . ^hus we can start the sum from -k
chosen such that the effective potential vanishes at its minimum; which then can be replaced by a sum ranging from -» to +*> . Therefore we can
write

A
A.
ItsG.

(28)
The minima of the potential are found from, the condition ^ ^ = o Oheld fixed) (0)
where the coefficients d are defiend by ;
which in terms of F {x) reads: 3B, k ji

-10-

-9-
 C (29)

In (28), ±n order to combine all the ft dependent terms in one factor, we have 2" T1
done a binomial expansion of (& ) in powers of (iT + tfQ) and (-&). Using the
integral

Thus the potential is reduced to a finite sum of the functions Sj. When E e
and also for large values of vit is suitable to use the series representation which
converges very rapidly. In other cases it is more appropriate to use the
integral representation:

(30)

and the identity:

(31)
$* (*)
J t- zx
(35)

which can he obtained by using the integral {3G).

For a vector the sum over J, starts from one. After a decomposition into
the two irreducible representations of S0(W + l ) , namely (£, 0,...j0) and
(i, l,0,...,0), the multiplicity of the latter can be written in the form:

(36)
(32) It may seem that the above steps do not go through because Di does not vanish
K for £ = _i_ Hovever there is also the missing terin£=O for the (X,0,0,. . .,0)
for the contribution of a 5 scalar to F (x). Here ._ (z) is a Bessel function
2+m part of the vector; this cancels exactly with the % «-l term of (/,l,0,,.,0)
which is given in terms of elementary functions; 2
such that ve can replace both of the sums by sums ranging from -*> to +•» .
For the traceless tensor <f> , whose irreducible representation of SO{N+1) are
(i,0,0s...,0) , (i,l,0,...,0) and {£,2,0,...,0) the multiplicity of the i =2
/ *
(33) piece of it can be written as:

Hov the infinite sum over 1 can be recognized to be the Poly-Logarithmic function

-12-

-11-

t f
" - 3 • • :
*-\$,u*.
 The higher powers of R are neglected on the ground that their contributions
are down by inverse higher powers of Planck mass. As we stated earlier, in
order to simplify the mixing problem we choose a background H^y =j- B, gL # .
p
To calculate P for different fields in (10) we use the veil known result
(37) [10] that fora scalar with the action

Again the sums over / can range from _» ^o +•»• a f t e r recognizing the following
D'-Q.) f (1*0)
cancellations . I = 0 of i = 2 with i=l of i = 1, i = -1 of i = 2 with i? = 1
of i = 0 and i = -1 of = 1 with 1= \ of 1 = 0.

To summarize, we have found the following expression fOr the one loop P is given by

<*y
3t iC J J

In Table 2 we give the result o f a simple calculation of the coefficients f

for different components of h and 1A . In fact if w e calculate the one loop
(38) action for (l) from a D=h + H dimensional point of view w e find that

which we will use for our numerical computation.

(»«
IV. OHE LOOP NEWTON COUSTMT

As we will see 'below, the tree level Newton constant has the wrong where a computation similar to the above gives

sign for some of the minima. But the relevant quantity is the sum of this and
the one loop induced Newton constant which of course should be positive. This
(it3a)
requires the calculation of the induced four dimensional gravitational R term
whose coefficient will give the one loop correction to G Q . For this we need the

effective action for a background 1^ . The one loop action V ((g^ .g

is given by: Is • ii

(It 3c)

H- R, 1.39)

-13-
 Inserting the expansions ( M a - c ) in Eq.^59) and using Table 2 for the coefficients
/ ° ^ ' we see that (39) and (1*2) are identical. This can be considered as a check
But (39) and (1*2) should be Identical. In fact it is not difficult to see this.
We first find the following expansions for small s : of Table 2.
Finally, after a trivial integration over s in Eq.(39)we obtain the

effective action:

where
(i*i*b)
.«•>

r
V
Mow we can read off the one loop Newton constant from (1*7): (1*8)
and

J Here we have used the radius at the minimum

(1*9)

(it lie)
These expansions are found simply by first writing (50)

For the minimum to be an acceptable one we require to be positive.

The numerical computation of E is done similarly to F
JM H
(1*5)

and then using the identity in (31) from which it follows that: RESULTS AW DISCUSSION
V.
As we mentioned earlier the contribution of graviton loops are relatively
large. In Table 3 we have given the contribution of different components of /)
and "[ and their total contributions to F_(x) at x = 0. The x = 0 behaviour
is particularly important because the sign of J^tO) determines vhether the
effective potential goes to +•= or -=» as a-+0. For comparison, in this Table
we have given also the corresponding quantity for a scalar in higher dimensions

-15-

-16-
 coupled to gravity by the action:
x = -.10668 for which F (x ) = .000378. In Fig.l we have plotted the effective
0 1 2
potential V _.(x) in units of ( ,-?-? ~~\ • As this figure shows the extreiaum is
ef f lbltu
a maximum and therefore it is unstable. Because \<(o and F 1 ( x Q ) ^ 0 from Eq.(2l)
we see that the tree level Newton constant G should be positive. However the
computation of ft gives p, = - .65 which means the one loop corrected Newton constant
is negative and therefore there is antigravity at the extremum x .
Of course for the scalar, F^ is independent of x. Note that for gravitons the
••i;?n of F K (o) is given by (-l)k. This can be understood from the behaviour of Fcr U>1 the - component of Faddeev- Popov ghost has M ^ 0 independent of

1 fcr large /. It is easy to show from (360 that x. This gives a constant imaginary contribution to F (x):

z * kl (53)

There are also tachyonic modes in some of the components of (^ itself. They
occur when x ) 5 for H = 3 and when x^(U-l)(H-2) + h for N> 3. These modes
Even for small k because of the presence of the exponential damping factor give an x-dependent imaginary contribution to F (x). \ic do not know yet if they
in (32) this asymptotic behaviour is still good for determining the sign. indicate instability of the background or they are gauge artifact. However we
have looked at the effective potential in light cone gauge[12] vhich is free
To find the extreraa of the effective potential ve have looked for the
from Faddcov-Popov ghost and therefore it does not have the constant imaginary
zeros of the function on the left-hand side of (25) vhich ve denote by U(x):
contribution let (53). What we find is that the potential is real for some range
of x. Therefore it might be the case that the imaginary contribution from the
Faddeev-Popov ghost is not an indication of instability. For this reason we look
for the exterma in the real part of the potential but in the range of x where
(51)
there are no negative modes in the components of yl, . The results for K = 3
to 17 are summarized in Table h. We have plotted the real part of the effective
For the case N = 1 the potential has an imaginary part when x > 0 . This is
potential in figures 2 to 9- There are minima for N = 5, 7, 9, 13, and 17.
because of the existence of tachyonic reodes. The mass of each mode is given by:
However because the value of the function F (x) at the minima is negative, for
this to be consistent we should have G / o . The result of the calculation of fa
given in Table 4 gives G = -j- 2 (o for all the cases ; for all the minima we have
anti-gravity which is not desirable.

If there are matter fields in higher dimension it is possible to obtain

acceptable solutions. For example with 1000 scalars minimally coupled to gravity
A glance at Table 1 shows that for x ) 0 some of the modes have M 2 < 0 . Thus
there is an acceptable minimum for s at X- =13.72 for which .3153 and
restricting ourselves to x <0 we find that the potential has an extremum at
Ps = 6.228. With the same number of scalars for 3^ the minimum is at X = 32.6 with
F(x ) = . 3271 and #/=12.037. This case (N=T) is particularly of interest in
*) The potential for 1J = 1 case with A = 0 was calculated in [llj .
connection with eleven dimensional supergravity which at the classical level has

-17- -18-
the solution Ad?xn . In this model there is no cosmologies! constant but there
NOTE ADDKD
p w
a totally an',- ••'mmetric gauge field Ap,cj) hii-h can h-ive u r.oij zert background
value After conr!i>ting this work we received a preprint by A. Chodos and
i:.:-:yers [lit] who discuss the s a w problem. Their net hod ?s different from oui •

= m rhey have performed the computation for ;••; = i w htrh agrees with ours.
I thank Professor Chodos for pointing out i sign error in an earlier draft
of this paper.

This parameter m plays a role similar to the cosralogical constant in the model
we have been discussing . The classical, solution does not fix m and therefore the
curvature of the space-time and the internal space are not determined. If the
effective action admits a solution with flat It-dimensional space time then the
parameter m and therefore the gauge coupling constant can be determined -
Considering our negative result it will be interesting to see if the presence of
*) ACKNOWLEDGMENTS
the gravitino and the gauge field in this model gives an acceptable solution

I am grateful to Professor J. Strathdee and Dr. A. Hamazie for many

helpful and stimulating discussions. I thank Drs. C M . Mukku, K.S. Marain
and R. ?andjbar-Daemi for useful discussions. I would like to thank
Professor Abdus Salam, the International Atomic Energy Agency and UNESCO
for hospitality at the International Centre for Theoretical Physics, Trieste.

*) This has teen discussed in Ref.[l3] but no explicit calculation has yet
performed.

-20-
-19-
 REFERENCES

[1] Tto. Kaluza, Sitzungsber. Press Akad. WIss, Kl (1921) 9&6;

0. Klein, Z. Phys. T7_ (1926) 895.
[2] B. De W i t t , Dynamical Theory of Groups and F i e l d s (Gordon and Breach,
New York, 1965);
A. Trautman, Rep. Math. Phys. 1_ (1970) 29;
Y.M. Cho, J . Math. Phys. 1§ (1975) 2029;
E. Witt en, Uucl. Phys. Bl86 ( l 9 S l ) 1*12;
Atdus Salam and ,T. S t r a t h d e e , Ana, Phys. lja (1982) 31b;
S. Randjbar-Daemi, Abdus Salam and J . E t r a t h d e e , Mud. Phys. B21lt (.1983)
1491
[3] M.J. Duff, E.E. Nilsson and C.N. Pope, Imperial College, London, p r e p r i n t ,
ICTP/82 -83/29 (1903).
[h] P. C'andelas and S. Weinberg, University of Texas, preprint, UTTO-6-83 (1983).
[5] C. Wotterich, Phys. Lett. 113B , (1982), 377.
[6] S. Coleman and E. Weinberg, Phys. Rev. D7. (1973) 1880;
K. Jackiw, Phys. Rev. D £ (191h) 1666.
[?] B. De Witt, in "Quantum Gravity 2", Eds, C.J. Isham, R. Penrose and D.W. Sciama
(Oxford University Press, 1981),
Al
[8] S. Hawking, in "General Relativity, An Einstein Century Survey", Eds.
S.Hawking and W. Israel (Cambridge University Press 1979).
[9] Abdus Salam and J. Strathdee, see Ref.[2],
J. Strathdee, ICTP, Trieste, preprint Ic/83/3 .
[10] B. De Witt, see Ref.[2].
[ll] T. Appelquist, and A. Chodos, Phys. Rev. Lett. 50 (1983), lltl;
T. Appelquist, A. Chodoa and E. Myers, Yale University preprint, YPT 83-04 (1983);
M.A. Eubin and B.D. Roth, Nucl. Phys. B226 (1983) hh; Phys. Lett, 127B (1983), 35
[12] S. Handjbar-Daemi and H.H. Sarmadi; in preparation.
[13] P. Candelas and D.J. Raine, preprint, September 1983.
[ll*] A. Chodos and E. Hyers, Yale University, preprint, TTP 83-18 (1983).

-21-
-22-
 TABLE 2

The coefficients p- of the induced F.^ term for different components

of h , ^ and n A (see Eq,. 39)

h •
h, + * h1 n
\
u

(p) 2T l . l , /sk-1 7 10
"5" 3"2 "V 2k+3 T 3
" 6
1
~ 6

TABLE 3

F for
The values of fl (°) different components of hA _ and nA , for their total and for

a scalar with minimal coupling in higher dimension.

Total
u scalars
H h plus ,j. n p (real part)

0511 ,
3 -9.5325X10" 1 7.8l49OxlO"3 -6.0549x10 -1.217l*xlO~2 -1.1261J 7.568x1O^5
2.2809xl0~ 2 5.15505x10 14.283X1O"11
5 2.914906x10 1.77705x10 It. 83122

7 -14.38081+xlO2 6.67672 -!4.67396xlO2 -1.8592OxlO2 -6.52708x1O" 3 -14.02901+xlO"1 -1.O8483X1O3 8 . 1 5 9 x 1 0 ^

9 l*.375601O3 -5.21525x10 7.5O955xlO3 3.97542X1O3 -9.97115xlO" 3 -1.81+327 1.63079x10" 1.1339xl0"3

ll -4.66255x10 3.11004x10* -9.140828x10 -6,21*052x10 -1.063^5xlO- 2 -4.87291 -2.02803xl0 5 1.3293xlO" 3

13 4.06534X1O5 -1.35777xl0 3 l.oi5l»5xio 8.10632xl0 5 -1.09922xlO" 2 -1.00633x10 2.23125x10 1.3TltOxlO~3

15 -3-3325WO 6 2.02285xl0 3 -9-92852x10 •9.27186x10 -1.00199xl0~ 2 -1.796728x10 -2.25309xl0 7 1.2525xlO" 3

L
17 2.6H* 5 W 14.31920x10 9.O5O53xlOT 9.68179x107 -7.64732X10" 3 -2.910808x10 2.13511x10 9.559X1O"1*
 TABLE 1* FIGURE CAPTICHS

The position of stationary points of the effective potential

and the values of and at, these ertrema: Fig.l - The effective potential V f f W for II = 1

Fig. 2 - The effective potential VeJ,f(x) for M = 3

Maximum
Stationary Fig.3 - The effective potential V e f f (x) for N = 5
or t\.(x n ) R (defined bv ^h
points X,•Q mimimuir,
Fig-li - The effective potential V e f f (xj for N = 7
Fig.5 - The effective potential VgffCx) for U = 9
-0.1067 0.000378 -0.65 Fig.6 - The effective potential V f f ( x ) for N = 11
Fig.T - The effective potential V e f f (x) for H = 13
X2.29 min -0.1161* 3.99 v x for
Fig.8 - The effective potential eff( ' M = 15
2T-1+1+ max -0.1893 3.8? Fig.9 - The effective potential V e f f (x) for N = 17
32.05 min -0.1*863 2.1*9
48.06 rain -0.289 7-6
50.16 max -0.3225 5.68
85.85 max -0.1099 3.1+8
,.13 103.18 min -2.176 T.33

131.7k max -2.31*6 2.06

183.207 max -3.795 2.05
C 1T
178.58 min -6.21*8 9.61
2*41.601 £.1(2

-25-
-26-
-.36 -.32 .28 -.24 -.20 -.16 -.12 -.08 -.04 0

8Q00
VBtttX) N-3
a0( P|00 OBO 160 240 320 400 4B0 560 640 7.20

-saoo
-16000

-240.00

32000

40000

-48000

Flg.1
Rg.2
 14Q00 VJX) N-7
\4,f(X) N-5
13000
10000
moo
60JOO
90.00
2Q00
70.00
2S|00/26XX) 27.00 28J0O 29.00 30.00 31.00 3 2 0 0 3 3 0 0 34:00
25
2000
5000
-6000
30.00
10000
1000
1(100 1080 lifiO 1240 1320 1400 14.80 1560 1640 17.20 -140.00
-10.00

Fig. 4
Fig. 3
 8O00
64O00 6400 6800 7200 76.00 80D0 8400 8800 92J0O 9600 1O0J00
N-9
480.00

320.00

160.00

000
4C00 46B0 4760 4840 49.20 50.00 50.80

160.00

32000

48Q00

Flg.6
Ro.5
 N=15
(X) N-13 aoo
16|7OO16aO017100 173JOO 175JD0 177.00 17&00 18100 18a0018500

aoo

7fcO084O0 9200 10000 10800 H6O0124.00' 13200 140.00148.00

Fig.8
 480.00

40000
. v

32000

o 24000

160.00

8000
\ /
1 .1 ^ - 1 '—^^1 1 • ' r - I 1 1

10OJOO17OXX) 180001900020000 21000 22000230002400025000

Flg.9