Correlation Between Gravitational and Inertial Mass:

Theory and Experimental Test.

Fran De Aquino
Maranhao State University,
Physics Department,
65058-970 S.Luis/MA, Brazil.

The physical property of mass has two distinct aspects, gravitational mass and inertial mass.
The weight of a particle depends on its gravitational mass. According to the weak form of the
equivalence principle, the gravitational and inertial masses are equivalent. But, we show here
that they are correlated by a dimensionless factor, which can be different of one . The factor
depends on the electromagnetic energy absorbed or emitted by the particle and of the index of
refraction of the medium around the particle. This theoretical correlation has been
experimentally proven by means of a very simple apparatus, presented here.

The gravitational mass, mg , produces δmβ represents also the inertial mass
and responds to gravitational fields. It shift . It can be seen by repeating the
supplies the mass factors in Newton's renormalization of the external
famous inverse-square law of electromagnetic vertex at finite
Gravitation (F12 = Gmg 1mg 2 r12 ). Inertial
2 temperature. On the other hand, it is
easy to see that the inertial mass shift
mass mi is the mass factor in Newton's
is related to inertial Hamiltonian shift
2nd Law of Motion (F = mi a ). δΗ . Thus we can obtain the general
Several experiments1-6, have expression of δmβ by means of the
been carried out since Newton to try to inertial Hamiltonian shift δΗ ,i.e.,
establish a correlation between
gravitational mass and inertial mass. δΗ c δp 2 + mi2 c 2 − mi c 2
δ mβ = = =
Recently J.F.Donoghue and c2 c2
B.R. Holstein7 have shown that the 2
 δp 
renormalized mass for temperature = mi 1 +   − mi [1]
T = 0 is expressed by mr = m + δm0  mi c 
where δm0 is the temperature- where δp is the correspondent
independent mass shift. In addition, for particle's momentum shift.
T > 0 , mass renormalization leads to Consequently, the general
the following expressions for inertial and expression of correlation between
gravitational masses, respectively: gravitational and inertial mass can be
mi = m + δm0 + δmβ ; mg = m +δm0 −δmβ , write in the following form
  2 
δm β   δp  
mg = mi − 2δmβ = 1 − 2 1 +   − 1mi [2]
where is the temperature- 
dependent mass shift given by î   mic  
 
δmβ = παT 2 3mi .
We can look on this change in
This means that a particle’s momentum as due to the
gravitational mass decreases with the electromagnetic energy absorbed or
increasing temperature and that only emitted by the particle ( absorbed or
in absolute zero (T = 0 K ) are emitted radiation by the particle and/or
gravitational mass and inertial mass Lorentz's force upon charged particle
equivalent. due to electromagnetic field).
The expression of δm β In the case of radiation,
obtained by Donoghue and Holstein according to Quantum Mechanics, we
refers solely to thermal radiation. But, can write
 2
δp = N k = N ω /( ω / k ) = U /( dz / dt ) = speed of light is reduced to <0.1m/s,
=U / v [3] the Eq.(7) tell us that the gravitational
masses of the atoms of the Bose-
Where U is the electromagnetic Einstein condensate become negative.
energy absorbed or emitted by the If the absorbed (or emitted)
particle and v is the velocity of the radiation is monochromatic and has
electromagnetic waves, which can frequency f , we can put U = nhf in
be write as follows
Equation(7), where n is the number of
v=
c
[4] incident (or radiated) photons on the
ε r µr 
 1 + (σ ωε ) + 1
 particle of mass mi . Thus we obtain
2

2  
 2 
ε ,µ and σ, are the electromagnetic   nhf  
characteristics of the medium in which mg = mi − 2 1 +  2 r
n − 1m [8]
 î mi c  
the incident (or emitted) radiation is î 
propagating ( ε = ε r ε 0 where ε r is the In that case, according to the Statistical
relative electric permittivity and Mechanics, the calculation of n can
−12
ε 0 = 8.854 × 10 F / m ; µ = µ r µ0 where be made based on the well-known
µ r is the relative magnetic permeability method of Distribution Probability . If all
the particles inside the body have the
and µ 0 = 4π × 10 −7 H / m ). For an atom
same mass mi , the result is
inside a body , the incident(or emitted)
radiation on this atom will be n=
N
a [9]
propagating inside the body , and S
consequently , σ = σbody , ε = εbody, where S is the average density of
µ =µbody. absorbed (or emitted) photons on the
From the Eq.(3) follows that body; a is the area of the surface of a
U U c U particle of mass mi from the body.
δp = =   = nr [5]
v c v c Obviously the power P of the
where nr is the index of refraction, absorbed radiation must be
given by P = Nhf / ∆t = Nhf , thus we can write
2

εµ N = P / hf 2 . Substitution of N into
nr = = r r  1 + (σ ωε ) + 1
c 2
[6] Eq.(9) gives
v 2  
a  P a
c is the speed in a vacuum and v is n=  = D [10]
the speed in the medium. hf 2  S  hf 2
By the substitution of Eq.(5) where D is the power density of the
into Eq.(2), we obtain incident( or emitted) radiation. Thus
  2  Eq.(8) can be rewritten in the following
  U  
mg = 1 − 2 1 +  2 nr  − 1mi
 [7] form:
î   mic  
   2 
  aD  
Recently, L.V. Hau et al.,8 mg = mi − 2 1 +  − 1m [11]
succeeded in reducing the speed of  î mi cvf 
 
light to 17 m/s by optically inducing a î 
quantum interference in a Bose- For σ >> ωε Eq.(4) reduces to
Einstein condensate. This means an
4πf
enormous index of refraction ( nr ≈ 107 ) v= [12]
at ~1014Hz. µσ
Light can be substantially slowed By substitution of Eq.(12) into Eq.(11)
down or frozen completely 9. If t he we obtain
 3
  π 
Rr = (∆z ) [17 ]
2
  aD µσ  
2
 σ i µ i f
3 3

mg = mi − 2 1 + 3
− 1mi [13] 9
 î mi c 4πf   The ohmic resistance of the dipole is 11
î 
∆z
This equation shows clearly that, Rohmic ≅ RS [18]
atoms (or molecules) can have their 2πr0
gravitational masses strongly where r0 is the radius of the cross
reduced by means of Extra-Low section of the dipole, and RS is the
Frequency (ELF) radiation.
We have built an apparatus to surface resistance ,
ωµ dipole
produce ELF radiation ( transmitter and
RS = [19]
antenna) and to check the effects of 2σ dipole
this radiation upon the gravitational
mass of a material surrounding the Thus,
∆z µ dipole f
antenna ( see Fig. 1).
Rohmic ≅ [20]
The antenna is a half-wave r0 4πσ dipole
dipole, encapsulated by a iron sphere
(purified iron, 99.95% Fe; µi = 5,000µ 0 ; Where µ dipole = µ copper ≅ µ 0 and

σ i = 1.03 × 107 S / m ). σ dipole = σ copper = 5.8 ×107 S / m .

The radiation resistance of the Let us now consider the
antenna for a frequency ω = 2πf , can apparatus ( System H ) presented in
be written as follows 10 Fig.1.
The radiated power for an
ωµ i β i 2
Rr = ∆z [14] effective (rms) current I is then
6π
P = Rr I and consequently
2
where ∆z is the length of the dipole and
P (∆zI )  π 
2
εµ [21]
β i = ω i i  1 + (σ i ωε i ) + 1 = D= =  σ i µi f
2 3 3

2   S S 9
ω ε ri µ ri  where S is the effective area. It can be
=  1 + (σ i ωε i ) + 1 =
2
easily shown that S is the outer
c 2  
area of the iron sphere, i.e.,
ω ωc  ω S = 4πrouter = 0.19m 2 .
= (nr ) =   = [15]
2

c c  vi  vi The iron surrounding the dipole

where vi is the velocity of the radiation increases its inductance L . However,
through the iron. for series RLC circuit the resonance
Substituting (15) into (14) gives frequency is f r = 1 2π LC , then when
2π  µ i  f = fr ,
Rr =  (∆zf )2 [16]
3  vi  1 L L
X L − X C = 2πf r L − = − = 0.
Note that when the medium 2πf r C C C
surrounding the dipole is air and Consequently, the impedance of the
ω >> σ ε , β ≅ ω ε 0 µ0 , v ≅ c and Rr antenna, Z ant , becomes purely
reduces to the well-know expression resistive, i.e.,
Rr ≅ (∆zω ) 6πε 0 c 3 . + (X L − X C ) = Rant = Rr + Rohmic .
2
Z ant = Rant
2 2

Here, due to σ i >> ωε i , vi is For f = fr = 9.9mHz the length

given by the Eq.(12). Then Eq.(16) can of the dipole is
be rewritten in the following form ∆z = λ 2 = v 2 f = π µiσ i f = 0.070m = 70mm.
 4
Consequently, the radiation resistance Above this critical current, mgi
Rr , according to Eq.(17), is becomes negative.
Rr = 4.56µΩ and the ohmic The Table 1 presents the
resistance, for r0 = 13mm , according to experimental results obtained from the
System H for the gravitational mass of
Eq.(20), is Rohmic ≅ 0.02µΩ . Thus,
the iron sphere, mg (iron sphere ) , as a
Z ant = Rr + Rohmic = 4.58µΩ and the
function of the current I , for
efficiency of the antenna is
miron sphere = 60.50kg ( inertial mass
e = Rr Rr + Rohmic = 99.56 (99.56% ).
The radiation of frequency of the iron sphere ). The values for
f = 9.9mHz is totally absorbed by mg (iron sphere ) , calculated by means of
the iron along a critical thickness Eq.(24), are on that Table to be
δ = 5 z = 5 π fµ iσ i ≅ 0 .11m = 110 mm . compared with those supplied by the
experiment.
Therefore, from the Fig.1 we conclude
that the iron sphere will absorb
REFERENCES
practically all radiation emitted from the
dipole. Indeed, the sphere has been 1. Eötvos, R. v. (1890), Math. Natur. Ber.
designed with this purpose, and in Ungarn, 8,65.
such a manner that all their atoms 2. Zeeman, P. (1917), Proc. Ned. Akad. Wet.,
should be reached by the radiation. In 20,542.
this way, the radiation outside of the 3. Eötvos, R. v., Pékar, D., Fekete, E. (1922)
sphere is practically negligible. Ann. Phys., 68,11.
When the ELF radiation strikes 4. Dicke, R.H. (1963) Experimental Relativity
the iron atoms their gravitational in “Relativity, Groups and Topology” (Les
masses, mgi , are changed and, Houches Lectures), p. 185.
according to Eq.(13), become 5. Roppl, P.G et. al. (1964) Ann. Phys (N.Y),
26,442.
 2 
 µiσ i  ai  2 
mi [22]
mgi = mi − 2 1 +   D − 1
6. Braginskii, V.B, Panov, V.I (1971) Zh.
2 3  
î 4π c f  mi  
Eksp. Teor. Fiz, 61,873.
7. Donoghue, J.F, Holstein, B.R (1987)
Substitution of (21) into (22) yields European J. of Physics, 8,105.
  2 2 2  8. L. V. Hau, S. E. Harris, Z. Dutton, and C.
  µi σi   ai  
mgi = mi − 2 1+   (∆zI)4 −1mi [23] H. Behroozi (1999) Nature 397, 594.

î  6cS   mi   9. M. M. Kash, V. A. Sautenkov, A.
S. Zibrov, L. Hollberg, H. Welch, M. D.
Note that the equation above
Lukin, Y. Rostovsev, E. S. Fry, and M. O.
doesn't depends on f . Scully, (1999) Phys. Rev. Lett. 82, 5229 ; Z.
Thus, assuming that the Dutton, M. Budde, Ch. Slowe, and L. V.
radius of the iron atom is Hau, (2001) Science 293, 663 ; A. V.
−10
=
riron .1 40 × 10 m airon = 4πriron
; 2
= 2.46×10−19m2 Turukhin, V. S. Sudarshanam, M. S.
and miron = 55.85(1.66×10−27kg) = 9.27×10−26kg
Shahriar, J. A. Musser, B. S. Ham, and P.
R. Hemmer, (2002) Phys. Rev. Lett. 88,
then the Eq.(23) can be rewritten as 02360]
follows 10. Stutzman, W.L, Thiele, G.A, Antenna
{ }
m gi = mi − 2 1 + 2.38 ×10 −4 I 4 − 1 mi [24] Theory and Design. John Wiley & Sons,
p.48.
The equation above shows that 11. Stutzman, W.L, Thiele, G.A, Antenna
the gravitational masses of the iron Theory and Design. John Wiley & Sons,
atoms can be nullified for I ≅ 8.51A . p.49.
APPENDIX: A Simple Derivation of the Correlation between
Gravitational and Inertial Mass.

In order to obtain the general pi = pg = 0 . Consequently, Eqs.[4] and

expression of correlation between mg [6] reduces to
and mi , we will start with the definition (
δH = 2 δp2c2 + mi2c4 − mi c2 ) [7]
of inertial Hamiltonian, H i , and and
gravitational Hamiltonian, H g , i.e., Hi − H g = (mi − mg )c 2 [8]
Hi = c pi2 + mi2c2 + Qϕ [1] Substitution of Eqs[7] and [8] into
Eq.[5] yields
Hg = c pg2 + mg2c2 + Qϕ [2] (m i (
− m g )c 2 = 2 δ p 2 c 2 + mi2 c 4 − mi c 2 )
where mi and mg are respectively, From this equation we obtain
the inertial and gravitational masses at   
2 
δ
rest ; pi is the inertial momentum and  
mg = mi − 2 1 + 
p
 
 − 1 mi [9]
  i 
m c
pg the gravitational momentum; Q is  
the electric charge and ϕ is an This is the general expression of
electromagnetic potential. correlation between gravitational and
A momentum shift , δp , on the inertial mass.
Note that the term inside the
particle, produces an inertial
square bracket is always positive.
Hamiltonian shift, δH ,given by
Thus, except for anti-matter ( mi < 0 ),
δH = ( pi + δp)2 c2 + mi2c4 − pi2c2 + mi2c4 [3] the second term on the right hand
Fundamentally δp is related to side of Eq.[9] is always negative.
absorption or emission of energy. In particular, we can look on the
In the general case of momentum shift (δp ) as due to
absorption and posterior emission, in absorption or emission of
which the particle acquires a δp at the electromagnetic energy by the particle
absorption and another δp at the ( by means of radiation and/or by
emission, the total inertial Hamiltonian means of Lorentz's force upon the
shift is charge of the particle).

δH = 2 ( pi + δp) c2 + mi2c4 − pi2c2 + mi2c4  [4]

2

 
Note that δH is always positive.
We now may define the
correlation between H i and H g as
follows
H i = H g + δH [5]
If δH = 0 , H i = H g , i.e., mg = mi .
In addition from the Eqs.[1] and
[2], we can write:
Hi − H g = pi2c 2 + mi2c4 − pg2c 2 + mg2c 4 [6]
For a particle at rest, V = 0 ;
 Transmitter
9.9mHz

26mm

coiled spring
70mm
r=123mm
iron µ i σ i
counterweight
(dipole + wires)
antenna
Cross section

counterweight
scale

antenna

Front view

Fig.1 - Schematic diagram of the System H

 7

mg (iron sphere )

I ( kg )
(A)
theory experimental

0.00 60.50 60.5

1.00 60.48 60.(4)

2.00 60.27 60.(3)

3.00 59.34 59.(4)

4.00 56.87 56.(9)

5.00 51.81 51.(9)

6.00 43.09 43.(1)

7.00 29.82 29.(8)

8.00 11.46 11.(5)

8.51 0.0 0.(0)

9.00 -12.16 -12.(1)

10.00 -40.95 -40.(9)

Table 1

Note: The inertial mass of the iron sphere is miron sphere = 60.50kg