plasma column : isotope separation with a hollow cathode
discharge
Citation for published version (APA):
Boeschoten, F., & Komen, R. (1977). On the possibility to separate isotopes by means of a rotating plasma column : isotope separation with a hollow cathode discharge. (EUT report. E, Fac. of Electrical Engineering; Vol. 77-E-72). Technische Hogeschool Eindhoven.
Document status and date: Published: 01/01/1977
Document Version:
Publisher’s PDF, also known as Version of Record (includes final page, issue and volume numbers)
Please check the document version of this publication:
• A submitted manuscript is the version of the article upon submission and before peer-review. There can be important differences between the submitted version and the official published version of record. People interested in the research are advised to contact the author for the final version of the publication, or visit the DOI to the publisher's website.
• The final author version and the galley proof are versions of the publication after peer review.
• The final published version features the final layout of the paper including the volume, issue and page numbers.
Link to publication
General rights
Copyright and moral rights for the publications made accessible in the public portal are retained by the authors and/or other copyright owners and it is a condition of accessing publications that users recognise and abide by the legal requirements associated with these rights. • Users may download and print one copy of any publication from the public portal for the purpose of private study or research. • You may not further distribute the material or use it for any profit-making activity or commercial gain
• You may freely distribute the URL identifying the publication in the public portal.
If the publication is distributed under the terms of Article 25fa of the Dutch Copyright Act, indicated by the “Taverne” license above, please follow below link for the End User Agreement:
www.tue.nl/taverne Take down policy
If you believe that this document breaches copyright please contact us at: openaccess@tue.nl
providing details and we will investigate your claim.
Isotope separation with a hollow cathode discharge
F. Boeschoten R. Komen
AFDELING DER ELECTROTECHNIEK GROEP ROTE~ND PLASMA
DEPARTMENT OF ELECTRICAL ENGINEERING GROUP ROTATING PLASMA
ON THE POSSIBILITY TO SEPARATE ISOTOPES BY NEANS OF A ROTATING PLASMA COLU~
ISOTOPE SEPARATION WITH A HOLLOW CATHODE DISCHARGE
F. Boeschoten
R. Kamen
This work was performed under the terms of the agreement between the
Tecqnische Hogeschool Eindhoven and the association Euratom, to
con-duct jOint research in the field of plasma physics.
TH Report 77- E- 72 Narch 1977 ISBN 90 6144 072 6
ABSTRACT 1
1. INTRODUCT ION 2
2. APPARATUS AND PLASMA 5
3. DIRECT SEPARATION 9
4. PLASMA MANTLE 12
5. PERPENDICULAR INJECTION 15
6. AXIAL EFFECTS - THERMAL DIFFUSION 17
7. CONCLUSION 18
ABSTRACT
An investigation. was made on the possibility to separate isotopes by means of a rotating plasma column. In spite of the high rotational velocities, direct separation in the plqsma did not seem promising because of the high temperature of the plasma. Therefore this study was mainly concentrated on the possibility to drive a neutral gas with the plasma in a set up which has some similarities with an ultra centrifuge.
The poor state of plasma physics if i t has to come to explanations and predictions on the behaviour of laboratory plasmasJrequired to make an
extensive study on the rotation of a plasma. The positive c:')lumn of a holle'",,·
cathode discharge was chosen as an object of s~ch a study as i t was expected to provide for a relatively simple rotating plasma column.
It is difficult to evaluate the role of neutral particles in the discharge. Even in r.elatively small concentrations, neutral partic3..es may be of great importance in a highly ionized plasma. An unforeseen axial tem-perature gradient in the plasma of the hollow cathode discharge ( caused by
charge exchange collisions) leads to domination of the axial separative effects over radial effects and to breakdown of the proposed scheme.
In pure argon gas static separation fqctors cf maximal 1.3 were
found. This is not unfavourable in comparision with other methods of isotope separation, but not enough in view of the costs of generating the plasma. Even if substential technical difficulties are left out of consideration i t seems that isotope separation by means of a rotating plasma is not of practical interest.
1. INTRODUCT ION
The high rotational velocities which a,re observed in various plasma experiments lead to the question whether rotating plasma col~mns may be used for isotop~ sepa~ation. The same question arises as the ultimate consequence
of a method of i~otope separation I1hich was proposed by Slepian in 1942
( Lit. 1). His so called" Ionic CentrifugeH
was based on the difference in
centrifugal forces which are experienced by ions of different mass,rotating
in crossed electric and magnetic fields. In the first calculations a single
particle model was used, i.e. the assumption was made that each ion moves in
the existing electric and magnetic field as if i t were alone. Then the required large ~oncentration of ions and the resul ttng collisions between the ions ... 'ere taken into account, leading to a fluid of ions. Loose reference was made to electrons which are needed for neutralisation of the space charge of the ions.
But by ~dding the required electrons to the ~on fluid we automatically land in the realm of plasma physic~ and the original scheme becomes meaningless due to the fact that the high dielectric constant of the plasma prevents the penetration of the required d.c. electric fields. Unjustifiable assumptions on electric fields present in the plasma were also made by Smith, Parker and Forrester (Lit. 2)and this fallacy has reappeared in the litterature till
the present qay ( see e.g. Lit. 3).
If plasmas are concerned i t is obviously better to use the two
fluid model ( see e.g. Lit. 4) for the calculations. Instead of concentr;ting
on the life of siI'gle particles the question becorr . ....;s how different density profiles for the isotopes may be obtained. Exte~nal electric fields which are used in single particle models cannqt be applied at liberty. Instead
we have to deal with electrostatic space charge fields which set up auto-matically in the plasma and which may flot be controlled at will.
Unfortuna-tely so far no gas discharge operated with magnetic field is studied dnd un-derstood well enough that the density profiles and the electric fields in the plasma may be calculated theoretically. Such a knowledge is needed when the discharge is to be used for the isotope separation.
In this context we decided to make an extensive experimental and
theoretical study of the positive column of the hollow cathode discharge.
turn~d out that the. relation between the measur~d values of the plasma parame-ters is properly described by the particle conservation equation (" classical
diffusion I'), the equation of motion of the ions, and the " classical II
power balance equ~tion of the ions. ~he electrons do not obey the stqtionary power balance equation, a large power surplus ~s drained off by strong low
frequency oscillations ( Lit. 7)* The origin of the radial electric field ( which is related to the rotation of the plasma column) lies probably in the radial gradient in the electron temperature. A theoretical expression for the radial density profile is still lacking, but it could be shown that the measured density profile agrees well with the angular velocity ( Lit. 8).
Thou<:lh of much help, a detailed knowledge of its pos.'.tive column is
still not sufficient to give an unambiguouq answer to the question how far a
continuo~sly operated hollow cathode discharge may be used for isotope sep<':!.ra--tien. For this the cathode region must be known as well. More research in this part of the arc may yield also interesting information on heating mechanisms of plasma particles.
Most experiments were made with argon, b~t helium, neon, nitrogen
and hydrogen were also used. The facilities did not allow for operation of
uranium arcs. The central plasma column of the hollow cathode discharge in argon makes up to 105 revolutions per second, which is ten times more than what may be obtained in ultra centrifuges. However ¢ue to the high tempera-ture of the plasma and to the way that the centrifugal force is balanced, the
high rotational speed of the plasma column does not lead directly to a hlgh separative power. The radial density profile does not differ sufficiently for the light and the heavy isotope to make direct separation in the p[asma attractive. The attainable separation factor is only 1.1, whereas a value of at least 2 is required.
From the beginning our hope was directed towards the possibility to
use the high rotational velocities of the plasma for driving a neutral gas.
It was proposed (Lit. 9) to use a rotating plasma in the form of a hollow cylinder with the neutral gas inside ( " P[asma mant[e method "). To that
,', This power loss is a great disadvantage for the economical outlook of
purpose hollow cylindrical plasma columns were generated with special de-signed hollow cathodes. It turned out, however, that for cathode diameters above about 2 em the circumferential velocity of the plasma does not increase, but even decreases. Correspondingly the separation factors were not higher than for the "norrQalll 1,3 em diameter cathoqe. (\180 charge exchange collisions pre-dominate over elastic collisions to such extend that the larger part of thE rotating neutrals have s~ffered charge e~change collisionsj thus they have
not only a high rotational velocity, but also a high temperature. Moreover
the plasma mantle is not as ~'tight" i;l.S was hoped for. Additiopal experiments made with a beam of neutral particles directed transversely at the rotating plasma column gave further information on the charge exchange processes and
confirmed the calCufations made previously.
Contrary to what was believed to be the case, the plasma column of the
hollow cathode discharge is not homogeneQus in the direction of the magnetic field. The measurements showed that the temperature of the plasma drops in axial direction from about 10eV at the cathode to about leV at the anode side of the discharge. It turned out, unexpectedly, that the separation power of
the arc is not so much dete+mined by radial forces (as was assumed in the
pro-posals), but more by axial forces. Not foreseen thepmodij'fusive effects in
combination with the cuY'rent are responsible for the measured enrichment of
the light isotope at the anode side. Also i t cannot be excluded complete!_y
that separative effects take place in the cathode region.
In this report the plasma physical aspects of the work are only touched marginally. These are discussed in detail in the annual reports (Lit.S and L~t.6).
The economY requires that practically all the gas which passes through
the arc is 100 per cent enriched. In other words, separation factors of at least
two should (maximal 10
be found (Lit. 9) _ This number folloWS directly from the gas feed 3
cm NTP/s) and the energy consumption of the discharge (about 7kW) _ As shown in Lit. 6 the high energy consumption of the discharge is not only due to ionization qf the gas (about 1 kW), but also to heat radiation from the catho-de (about 1 kW) and to heating of the plasma with successive energy transport by
the e+ectrons (about 5kW). The magnetic field which is needed for confinement of
the plasma particles does not loose energy to the plasma; i t may be generated oy
permanent magnets or by superconducting coils and has not to be incorporated tn the energy consumption.
2. APPARATUS AND PLASMA
The hollow cathode disGharge as was used in this investigation is an arc sustained in a vacuum environment in the p~esence of a magnetic field. Gas - in t~is case argon- is introduced through an incandescent tantalum tube at the cathode side and pumped away by vacuum pumps, which maintain a pressure yarying from 10-3 Torr to 0.1 Torr along the axis. The
gas is very efficiently ionized and heated by passing the cathode region. A picture of the set up is shown in Fig. 1; a schematic in Fig. 2. The mass spectrometer is of the Quadrupole type (Balzers QMG lOlA). The vacuum in
the analysing chamber is sustained by a turbo molecular pump. The use of
-8
gold rings provides for an ultimate pressure better than 10 Torr with
no detectable traces of parU'cles with mass 36. The gas which entered at
h d . d 1 d f 1 -5 10- 6 P ' .
t e ano e 81 e was ana yse at pressures a 0 to Torr. rOVl.Sl0nS
are made for freezipg of impurities on surfaces cooled with liquid nitrogen
- i t turned out/ however, that this does not influence the reading of the mass spectrometer at masses 36 and 40.
The plasma density n (r/z)/ the ion
e
perature T (r,z), the mass velocity
e
temperature T. (r,z), the electron
tem-1
v(r/z) and the plasma potential ¢(r/z) /
were measured as function of the discharge parameters - arc current I
3
( 10 - 300A) / gas feed
12
(0.2 - 8 cm NTP / s) / magnetic field strength B ( 200-' 5400G; homogeneous within 5%)/ arc length L ( continuously variable between 25 cm and 250 cm) and core diameter d ( cathode tubes of 6/9/13/20/ 30 ~ninner diameter and some special models). The discharge is operated continuously. Under" standard" conditions L = 140 em, d= 1.3 em, B = 3400 G, I = 10QA
3
and Q
=
4.5 cm NTP/s.The dependence of the plasma density, n , on radius is shown in Fig.
e
3 ( measured with Thomson scattering and Langmuir probes). In radial
direc-tion a distincdirec-tion may be made between the core region ( O~ r <~d), the
re-gular region ( ~d ~ r < r
k), and the turbulent region ( rk ~ r < R) .In the
core region the plasma carries the current; the plasma is highly ionized ~n the
core >95%) and its conductivity is found to be " normal" within a factor
of two. In the regular region the radial density profile is approximately
21 2 -r q
n ( r ) = n ( o ) e (1)
e e
No theoretical expression for q2 is known, but from experiments we find
2 -1 -l,
that q is proportional to B T. and depends only weakly on the ion
2
2
mass. In argon gas q 0 6.4 cm at B = 3400 G. The turbulent region
pro-vides for a transition of the plasma to the wall.
that n e
density
Both Thomson scattering an Langmuir probe measurements indicate
practically does not vary with z ( less than 30% per mI. The radial profile and its logarithmic derivative dlnn
lar
also show littlee
variation with z.
The ion temperature, T. ( derived from Doppler broading of the A II
~
4806 ~ line), is found to be constant ( within 5%) with radius ( up to r =
7 em). The electron temperature, T (measured with Thomson scattering and
e
Langmuir probes) drops sharply in the region next to the core as may '·'8 seen in Fig. 4 which was measured at z = 60 cm ( the middle of the column) . The axial variation of T. and T. at the axis ( r = 0) is shown in Fig. 5.
e ~
Approximately T (0) ~ 2 T. (0) and T. depends weakly exponentially on z.
e ~ ~
The angular mass velocity ~ = ve/r ( derived from Doppler shift of
the AII
R
line, pendulum and directional probe measurements) is stronglysheared Fig. 6). Like the temperat'.lre the rotational velocity decreases
with z ( Fig. 7). The be about 6 x 104 cmls
axial component of the mass velocity V
z ,is found to same diagnostic means): the plasma flow is direl.t:ed
from the cathode to the anode. The
2
radial componen~ of the mass velocity is near the core v ~
r 3 x 10 cmls ( indirectly determined and apparently
in agreement with the " classical" diffusion rate) .
Fig. 8 shows the variation of T. and ~ with magnetic field
strength-~ 0
up to B 0 3000 G they increase monotonously with B. Fig. 9 shows the
Varia-tion of Ti and rlo with arc current I. It is possible to vary no within certain
limits, but i t has always to be kept in mind that the plasma parameters
cannot be varied independently.
In and near the core region the plasma behaves as may be expected
from the equations of the two fluid model of a plasma in the stationary state; only the energy equation of the electrons cannot be balanced without taking
plasma ( Lit. 7). The relationship between the experimental determined d.c. values of n
e, Ti , ve,Er and Bz is properly described by the equation
of motion of the ions. This is illustrated in Fig. 5 where is shown that the relation.
( 2)
holds. Q
=
ve/r is the mass angular; QE is the angular 'frequency asso-ciated with the electric drift ( c/rx(Er/B) and Q
di is the angular frequen-cy associated with the diagmatic ion current: ( c kT. / reB x ( dlnn / 3r),
~ e
RelatIon (2) follows from the equation of motion of the ions by neglect of the centrifugal force ( see e.g. Lit. 4). The radial electric field was
de-termined from floating potential measurements with Langmuir probes ( a
pro-per procedure in the region r > 2 em). Its origin lies probably in the
ra-dial gradient of the electron temperature in combination with the finiteness of the Larmor radii of the ions.
The fact that the plasma of the positive column is highly ionized does not mean that the neutral particles do not play an important role in the discharge. E.g. the decrease of the ion temperature along the axis is
caused by collisions which the ions suffer with an influx of cold particles
from the ,,'all. As a matter of fact we are dealing with three fluids and the neutral gas may only be neglected in some restricted areas of the discharge.
Unfortunately i t is impossible to obtain about the neutral particles "as much information as about the plasma particles.
The density distribution of the neutral particles is very difficult
to measure as ionization gauges can only be placed far away from the plasma,
and the neutral pressure, p , which is measured at a certain place in the
o
vacuum chamber does not give information on the pressure elsewhere. Besides on the gas flow the distribution of the neutrals depends on the interaction of the plasma with the neutral gas, like ionization ( mainly in the active
zone), recombination ( mainly at the wall and at the anode) and charge ex-change ( pumping action of the discharge). Fig. 10 shows p (z) measured at
o
the various portholes at a distance of 50 cm from the centre, beyond the wall
of vacuum chamber ( R
=
16 cm). neutral particle density in the3
part/ cm . In the neighbourhood
-4
about 7X 10 Torr and n ~ 2 x
n
From the spectroscopial observations the
core region is estimated to be n (0)
~
10.12n
of the wall the neutral gas pressure is
13 3
va-lue of n at intermediate radii was obtained by introducing a tantalum
n
tube (
¢
=
6 mm) into the plasma and by measuring the pressure at the other end of this tube, p , as function of the radial position of the snout"'. Fig.n
11 shows p (r), together with the most probable variation of n with r
be-n n
tween its values in the core and at the wall. Outside the core n (r) drops n
exponentially with an e-folding l.ength
Of
about 0.6 cm. The arc clearly acts as a" pump.3. DIRECT SEPARATION
At first one might think abo~t separation of the isotopes direct-ly in the plasma. This cannot be realised so easidirect-ly because of the high temperature of the plasma. The plasma may be heavily disturbed because of the introduction of a foreign body into the plasma and vice versa the extraction gadget mall., go to pieces in the plasma. Only the physical aspects of the separatiqn will be treated here and not the possible technical diffi-culties. The experiments were made with natural argon gas and the enrichment in the 36A isotope ( natural abundance 3,4 per mil) was measured.
In order to see how far radial forces in the plasma ,nay be used for the separation of ions of different mass, it was at first assumed that the plasma parameters do not change along column. The axis of the plasma column points in the direction of the magnetic field. Like in the ultracentrifuge the separative ~ower of the plasma column is expected to depend strongly on a difference in density profiles of the isotopes. According to eq (1) the radial density profiles of the plasma is approximately Gaussian:
n(r) ~ n (0) exp _ r2/q2 2 -1 -~ q is found to be proportional to B T ion mass m .. At ~ in the hydrogen 2 B = 3400 Gauss, q ~ 6.4 arc.
to depend only weakly on the 2
in the argon arc and q ~ 3.0
The fact that the ion temperature is constant over the radius, indi'::
cates that the ions at different radii are in thermodYnamic equilibrium.
This leads to a separation factor between centre and radius r: Cl
=
n, (0)/n2(0)=
ex r2(!
~!)n1 (r)/n
2(r) p ql 2 q2 2
(3)
It may be expected that the light isotope (q1) is more concentrated near the axis and the heavier one (q2) is more abundant in the outside regions. With
2 2 2 2
q2 - q1 )
=
6 q «q1 equation (3) yields approximately2 2 4
Cl ~ exp r 6 q / q (3a)
There is some ~esemblance to the expre~sion for the ultracentrifuge where we
f ind uDlW , 2 / 2 kT instea d of 'q2/ q4. F Ll or na ura t 1 argon gas ( 1 ) . a - loS expecte d to be about 1/10 ( r = 4 cm; 6q2
~
0.3 cm2).The separative power is given by:
1f 2
:2
pDl.(a - 1) LThe perpendicular diffusioncoefficienti' 4x 10- 4n
Dl. ~ B2
IT
e(4)
3
expressed in units of part/ em I B in Gauss and T in eV. With
llq2 2 e wt:ere n is (0.-1) ~ ~ r we find: q 6u ~ 21f. 10- 4 m(r 2n)2 B2/T e (4a
l
Thus the separative power
tional to
T~.
Its maximumis expected to be independent of Band
propor-value was attained for r == q. Exp.1~ession (4a)
does not yield a promising starting point for direct separation in the plasma. The corresponding separation factor (3a) is too low to be
measu-red by the means at our disposal, permitting an accuracy in the
measure-ment of a of about 10% (see also Lit. 10).
The angular frequency is not constant over the radius but is strongly sheared (see Fig. 6). Its radial profile is related to the radial density profile and to r
k,
~t a distance of 4
which depending on the magnetic field strength, B, lies
to 5 em from the axis (see Fig. 3). For r ~
of the plasma particles with neutrals become of importance and
r
k collisions the transport across the magnetic field d~ffers qualitatively from that at the centre.
Cor-respondingly, the radial density profile which is approximately Gaussiap in the regular plasma region shows a kink at radius r ~ rko It was found that in this regiot of the arc the gas is enriched in its light component. The dependence of the stattc separation factor on the radius is shown in Fig. 12.
In order to improve the situation one may envisage the use of a
car-rier gas of much lighter mass than the isotopes which have to be separated (Lit. 9). In the hydrogen arc the e-folding length squared has half the
va-2
lue of q found in the argon arc (at the same value of B). By mixing a
rela-tivelY'small amount of argon to the hydrogen arc the separation factor for
the argon isotopes is expected to be quadrupled (4a). The increase in
sepa-*
This .j.s the so called II classical II coefficient which is found in bur experiments.ration factor goes at the cost of the over all efficiency as the hydro-gen arc may be loaded with not more than 5% argon in order that its
pro-perties are not too much changed by the presence of the argon atoms. In-deed higher separation factors are found by this method, but again in the
outside region of the arc where the theory cannot be applied (Fig. 13).
Calculations on isotope separat~on with rotating plasmas were also
made by Bonnevier (Lit. 11) and by Nathrath et al. (Lit. 12,13). The for-mula which they derived for the separation factor is basically the same
as in an ultra centrifuge:
" - exp (5 )
In experiments with a rotating neon arc at medium gas pressure (10
Torr) a values of 1.1 were found in agreement with what could be expected
(Lit. 14). The experiments of Bonnevier with hydrogen- deuterium mixtures (Lit. 15) are harder to interprete as they were made in a complicated toroidal geometry - the plasma rotating like a car tire. Also the pulsed operation
(duratiory about 1 msec) does not allow an interpretation in stationary state
terms, and the plasma parameters are not known locally. Similar experiments,
but with better known plasma parameters were ~ater made with neon by Cairns (Lit. 16), who found separation factors of about 1.1.
4. PLASMA MANTLE:
Because of the low separation factors attainable directly in the plasma, our efforts to separate isotopes with the hollow cathode dischar-ge were mainly directed at the use of the plasma as an intermediate which drives a neutral gas. It was expected that in this way high rotational
velo-cities could be combined with relatively low temperatures of the neutral gas. It was proposed ( Lit. 9) to generate a rotating plasma in the form of a hollow cylinder with the neutral gas inside ( Fig. 14). The axial magnetic field should be strong enough to compensate for the pressure of the neutral gas and the plasma mantle must be thicker than a mean free path length of a neutral particle in the plasma.
Doppler spift measurements of the solid rotating core of an argon arc under normal operation revealed that the neutral atoms acquire indeed the rotational velocity of the argon ions:
(6)
However, their temperature also equals the temperature of the ions:
T (r=O) '" T. (r=O)
n l. (7)
Apparently the plasma -neutral interaction is mainly determined by charge exchange processes, which after all turned o~t to playa very important role in the hollow cathode discharge. E.g. they are also responsible for the axial variation of the ion temperature. (Lit. 17).
Rotating hollow plasma cylinders were realised by using cathode tubes, consisting of two concentric cylinders (Fig. 15). Two sizes Were used: a cathode
(Cl) made of tantalum tubes
¢
20 x 1,5 mm and¢
13 x 1 mm, and a cathode (C2) with a¢
30 x 2 mm outer tubethe cylinders are respectively
and a
¢
20 x2
1.4 cm and area of the standard cathode
¢
13 x 1 mm.1,5 mm inner tube. The area between 3.0 cm , compared to the 1.3 cm 2 2
Like observed before with the normal cylindrical cathodes, the ion temperature, T., is constant with radius, but decreas~s in axial direction
l.
towards the anode. The temperature of the neutrals also equals the ion tempe-rature. The values
drical cathodes if
of T. and T are close to the values found with the
cylin-l. n
gas flux.
Fig. 16 shows the intensity of the A I I 4806
R
lime and the much weaker A I 4201R
line as function of the radius for the large cathode, at z=
0 cm for I=
100 A and I=
200 A ) and at z=
60 cm ( I=
A) .Similar profiles were found wi th the Cl cathode. The intensity pro--files are clearly hollow, in contrast to what is found with a normal catho-de. As the intensity of the spectral line varies - ne exp - ( l/kTe) the intensity profile would represent the n (r) profile if T were constante e
with radius. A hollow radial intensity profile would, however, also be found in case the electron temperature drops somewhat at the centre. A de-finite choice between these two altenatives could not be made, as we were not able to make Thomson scattering measurements on large diameter cathodes.
The results of the Doppler shift measurements are shown in Fig. I;; .. Inside the core region ve is not proportional to r, ( as is the case for the normal cylindrical cathodes), but
n
approaches zero at the centre of the arc column. As mentioned in the previous section, the origin of the plasma rotation lies probably in the radial gradient of the electron temperature. The ve (~) profile is about as expected when Te is constant with radius in~side the cathode diameter and changes rapidly outside.
In order to increase the neutral gas density inside the hollow plas-ma'cylinder independently of the main feed, a ceramic tube with separate gas
feed was mounted at the centre of the Cl and the C2 cathodes ( Fig. 15- dot-ted lines). Additional central injection of relatively large amounts of neu-tral gas caused a decrease in the rotational velocities and in most cases a decrease in the measured separation factors. Numerous attempts made with various gas discharge parameters showed that the separation which may be ob-tained with rotating hollow plasma cylinders is not larger than found with solid rotating plasma cylinders ( a ~ 1.3) .The enrichment in the light iso-tope is likewise near the anode and in the outside regions of the arc. Appa-rently.the proposed scheme ( Fig. 14) does not work because of charge exchan-ge collisions, which are also responsible for the inhomoexchan-geneity of the
plas-rna in axial direction. Moreover the plasma mantle might be somewhat
"leaky" ,-: as the m.f.p. length of a neutral particle is comparable to the thickness of the plasma mantle.
In order to decrease the m. f.p. length for ion _ neutral collisions some experiments were also made with a medium pressure ( Po
low cathode arc. The temperature of this arc was much lower so were the rotational speeds. We did not manage in running
~ 1 Torr) hol-T,40 leV), but
~
the arc stably for longer times (~ 1 min)· and no separative properties were observed.
A different proposal to use a rotating plasma for driving a neutral gas was made by Lehnert ( Lit. 18). It is not clear, however, how in his proposal the required d.c. electric fields may be introduced into the plas-ma ( see introduction). Moreover in Lehnerts proposal the neutral gas and the plasma are mixed in the same volume, whereas it seems advantageous to have
the neutral gas as much as possible separated from the plasma.
,', The interaction of the neutral gas with the plasma in the hollow cathode discharge takes place in such a multifarious and complicated way that thin-king in such terms should be regarded with the required caution.
5. PERPENDICULAR INJECTION
The findings which are described in the previous sections were ob-tained with a cylindrical plasma column and axial injection of the neutral gas. The gas discharge ( of which the positive column is only a part) re-sults as an interplay between the applied voltage and the gas. The created plasma does not only depend strongly on the amount of gas which is fed to the discharge, but also on the way the gas feed ( and pumping) takes place. Ir.l order to know more about the role of the neutrals in the discharge a beam
of neutral particles was directed transversely towards the core of the po-sitive column.
A flow channel for gas injection is easily obtained by replacing the Langmuir probe in the probe mount shown in Fig. 8 of Lit. 5 by a 6 mm i.d.
tantalum tube. This opens the possibility of injecting a beam of neutral particles into the arc at various radii and axial positions ( or alternati-vely to suck off neutral gas, e.g. in order to analyze i t in a mass spectro-meter) •
3
Perpendicular injection at z= 60 cm of 2 cm NTP/s argon gas at a
dis-tance r=2 em from the centre of the arc causes a drop in the ion temperature
~t z= 60 cm from T. = 3.6eV to
5 ~ 5
10 rad/s to 1.2 x 10 rad/s
T
=
1.7eV; in this casen
drops from 2.4 xi 0
The same value of no and Ti are measured when the gas stream is directed obliquely upstream ( remember the plasma streams
4
from the cathode towards the anode with a velocity of 6 x 10 cm/s). When di-rected downstream the effect of the neutral particles on Ti and no is negligible. These measurements indicate that the ion temperature depends on ion -neutral collisions and support the calculations which show that the axial gradient in Ti in due to charge exchange collisions. An exact evaluation is complicated by the fact that with extra injection of neutral particles their density raises everywhere in the vacuum chamber.
So far these experiments were only made for charge exchange studies, but in principle it might be possible to separate isotopes in this way. Argon particles of different mass emerge at the other side of the arc with somewhat different directions. Due to the inhomogeneities in longitudinal direction this effect could only be measured if two pumping baffles were placed at both
arc
beam
neutr~======~====;;;;i:~::~---particle~
Fig. 17
sides of the region where the neutral particle beam intersects the arc ( Fig. 17). It did not seem worth while to make such efforts with the argon arc, but it could be considered in experiments with an uranium arc. Sampling should be relatively simple in that case.
6. AXIAL EFFECTS - THERMAL DIFFUSION
The tendency of the light isotope to concentrate at the anode side of the are, together with the fact that the temperature in the arc drops from the cathode toward the anode raises the surmise that thermo diffusive effects could be partly responsible for the measured isotope separation.
If the suffix 1 is used to denote the light isotope 36A and 2 the heavy one
4~,
the flux density of species 1 is ( n2» nj):
nj Vj ~ - Dt2 ( V n1 + \i n1 V tnT) ( 8)
where v is the thermal diffusion factor ( see e.g. Lit. 19). In steady state n
1 v j = 0 and the separation factor 0; is found to be:
in 0; = -v in (T(Z)/T(O») (9)
With the measured values of the mass separation in the argon arc v is found to be about 0.1. ( o;.p n
1 (an);! n1 (cath) ~ 1.3; T(an)! T(cath)~ 1/1Jl).
Contrary to what is generally the case in a gas, in the hollow cathode dis-charge the light isotope has the tendency to concentrate at the cold side. This may be due to the fact that the friction with the electrons which carry
the current is larger for the light isotope (T i, e - m;>.
An effort was made to Use this effect by deviding the region around the arc in axial direction in two parts. To this purpose a water cooled dia-phragm was mounted between two sections of the vacuum chamber ( see Fig. 18).
Indeed somewhat higher separation factors than before were measured (OC~ 1.3) if the anode was placed close to the diaphragm.
Placing a diaphragm around the core of the arc affects inevitably the gas discharge. The effect of the diaphragm on the plasma is described in Lit. 6. The radial electric field decreases and the rotational velocity of the arc changes correspondingly.
7. CONCLUSION
The idea behind this work was to see whether the high rotational
speeds observed in plasma columns may be used to separate isotopes. This seems attractive as the angular velocity is ten times larger than in an
ultra centrifuge and is not limited by mechanical problems.
It turned out that the gas is enriched considerably in the light isotope 36A at the anode side of the hollow cathode discharge operated
with argon. The same enrichment is found with hollow plasma cylinders,
which were realized in the hope to transfer the high rotational veloci-ties of the plasma to a relatively cold neutral gas ( plasma mantle method) . This scheme did not work as expected because of the predominating role of
charge exchange collisions. These collisions are also responsible for an
axial temperature gradient in the discharge, which obscures the effects due to forces working in radial direction. After all it turned out that the axial temperature gradient in combination with the current is responsible for most of the separation which occurs in the arc.
Evaluation of the results mentioned in this report may be made by
comparision with the ultra centrifuge. Attention is paid to the physical
aspects only; other requirements, like technical feasibility, are not con-sidered. Taking the power consumption and the capital cost of the discharge into account i t is necessary that:
1. Separation factors of at least two are obtained. Assuming that the expression for the separative power is comparable to that in an ultra
centrifuge ( eq. 4) this would correspond to ten times more separation per unit length.
2. The separative power is attainable in practise, i.e. it must be possible to obtain separation factors of two in at least half of the amount of gas which flows through the arc.
·In pure argon gas separation factors higher than 1.3 were never found. As even the first condition was not fulfilled, there where no efforts made to check on the second. Separation factors of two could be obtained by adding less than 5% argon to hydrogen ( carrier gas). But it is clear that in this case the second condition is not fulfilled.
At the basis of our experience it may be stated that isotope
separation by means of a rotating plasma column does not seem to be of prac-tical interest.
ACKNOWLEDGEMENTS
The authoxs would like to acknowledge the skilful assistance of A. van Iersel in performing the experiments, the Thomson scattering ex-periments made by W.F.H. Merck and A.F.C. Sens, and the helpful discus-sions with L.H.Th. Rietjens.
REFERENCES
1. Slepian J. in Wakerling R.K. and Guthrie A. (Eds), " Electromagnetic separation of Isotopes in Commercial Quantities", TID - 5217 ( 1949).
2. Smith L.P., Parkins W.E. and Forrester A.T., Phys.Rev. ~, 989 (1947).
3. Hashmi M. and Van der Houven Van Oordt A.J., Int. Conf. Uranium Isoto-pe separation, paIsoto-per 7, London ( 1975).
4. BraginskiI S.l., "Transport Processes in a plasma", Rev. Plasma Physics
!,
Consultants Bureau, New York (1966).5. Euratom- THE Group, "Experiments with a large sized hollow cathode dis-charge fed with argon II",TH Eindhoven - Report 75-E-59 ( 1975).
6. Euratom- THE Group, " Experiments with a large sized hollow cathode discharge III", TH Eindhoven- Report 76-E-67 ( 1976).
7. Boeschoten F., Kleyn D.J. and Komen R., Bulletin Am. Phys. Soc. Series II,
.?!.
1167 (1976).8. Janssen P.A.E.M., to be published.
9. Boeschoten F., " Rotating Plasma", Euratom 15.~36 I XV I 69, Brussels (1969) .
10. Boeschoten F., Int. Conf. Uranium Isotope Separation, paper 11, London (1975) •
11. Bonnevier B., " Diffusion due to ion - ion collisions in a multicompo-nent Plasma", Arkiv for pysik~, 255 (1966).
12. McClure J.J., Nathrath N. and Schwenn R., Proc. 10th Int. Conf. in ionization Phenomena in Gases 209 (1971).
13. Nathrath N. et aI, Int. Conf. Uranium Isotope Separation, paper B London ( 1975).
14. Heller H. and Simon M., Phys. r.etters SOA , 139 ( 1974).
15. Bonnevier B., " E~perimental evidence of element and isotope s~paration
in a rotating plasma ", Royal Institute of Technology - Report, Stockholm ( 1970).
16. Cairns J.B.S., Int. Conf. Uranium Isotope Separation, paper 9, Lon-don ( 1975).
17. Boeschoten F., Komen R. and Sens A., Proc. 12th Int. Conf. on ioniza-tion Phenomena in Gases, 98 (1975).
18. Lehnert B., Physica Scripta 2, 106 ( 1970).
19. London H. ( Ed.) " Separation of Isotopes ", George Newnes Ltd, London ( 1961).
Fig. 1. Experimental set-up
3 movable anode support 4 magnetic field coil
5 laser diagnostic equipment 6 Fabry-Perot interferometer 7 high speed camera
8 mass spectrometer
9 cdrrelator and spectrum analyser
pump needle valve
l.==i)<l::::::::l
flowmeter ar cnamoer pumpcold t cold tra
==-ig.2 Isotope separation with to!.,· '-'8ClJU1" 3re
t"l c:: XI •
'"
o o'"
'"
OJ C
1
r -4-...
I'\.
\
\
S -I ;_.
-, -- --. 01 1 .. 5 --,.- 1--- .- --1012•
1,
d cathode 2 - - - - -_. -i\
i,
}r -.
I
I
I
,
I1
I
t
:\
,
- 1--- --I~
,
I
-1
I
I
--I
I
--- - .t
( I I I 4 I II I I
B=
3400 G - - - - ..._-
-I=
100 A 3 Q ~ 4.5 em NTP/s Po=
10- 3 Torr L=
120 em d=
13 mm z=
60 em f--- -. - _ . -- L-I0 Thon son catt ring
I
•
Lans muir prob s-
t--
-I - ---- I--- -- -- -_ ..._-~
... - . . __ . -_ .._.-~
-- ----... .. -.• . . - -_._.-1----s
(, 7..
r (em)Fig.3 Radial density profile at the middle of the positive eolumn of a
9
0 Thorn on s catt ring
"'-
•
Lang .uir prob s""
X Dopp er t road ning of 1\ I 46 06)( peet al line\
7\
6 ,T 5 4 Ti\
3\
\
\
'\~
2~
1...
1 23
45
6
•
r (em)Fig.4 Radial profiles of electron- and ion temperature (stAnriarri condi tiona,
> Qj
~
!~
I
10'\
I
I
I\.
, IX : ave, age ver ~!Lm asur ment~
'\
",ITo+
··11+
i•
-- ave, age ve r p me sure ents I
-tr
: star dard,
devifi
tiol-i
~--t--- - -~--1""
IJ-~~
I--1-
I - -- --I
I
!..
""
i
II
! , i I I,
, ,-I
""
!
I'\.
I ,i
~
, I i -1---. -~t-""
;~
i
I I Errc r in z I I~
i
,
i'"
T.I
•
i ; 1 ,I
•
,
+--- T--~i
1\.
i
I
I~++'\~-it
--~
~+---""
~
I
I
,
I"-...j----....
. '\
~~---.t-...
l'\.
r-
~
'\
Ar,o e 5so
100 - - - . . . , .... z (em)Fig.5 Axial variation of electron- and ion temperature; r =: 0; standardc.:onditions. T measurements at lower values of z,
are impeded by the fact that the amount of scattered light in each wave lengthe "channel" becomes smaller by
further broadening of the Doppler profile (6A~Te~) and finally drowns in the background fluctuations.
'"
o
o
'"
2.1)
1\
x Dopp ~er hUt of A I 48b6J1
pect al.
LOne--\
---I
•
Lang Inuir prob s 2\
•
Dire i::tio al L ngrnu r pr obes\
Pend ulum+
1 .5\
I ~ I+--
,
\
- - - --l ------
! i -..
-t
' . .ClE Er or -- . . 1\\
~
'\
~
.-!\
0.5\
'-...,
~o
1 2,
~
~ ;f.
cr-\cm~.
...
V
.! D1 --0.5 " , , -i.-.11. - n.~ -1Fig.6 ·Angular mass velocity •
.n.
at the middle of the positive column of ahollow cathode discharge as function of radius, measured from Doppler
shift of the A;L;L 48061\ spectral line,S·,·."·\3,,d oon'l'tlun,·.; z t.,(\ ~m.
~i (diamagnetism of ions)
.n
E (electric drift) from Langmuir probe measurements.
n=A +AE
It
"-"-
"-
"
"-
'"
r...
"
...~
~o
3
2~
~
...i'-..
"
I'---
~
~ 1".
- -..
Fig.? Dependence of an~lar frequencyJl
o' and ion temperRture Ti • on axial distance. Normal operation of the arc.
8
1
6
It 2 ';:' t>l ~.,
'"
o o 00'"
If T (. IT) •
.]
.0.
0 ( 05r die.,/'
,....
".\ / lV
1
,
lA(
~
2 2./'
•
./
/
V
1 1/
/
-B(56
~ G),
I ( 1p , ~. D", ~..
,,~."".
bf .. ,...
,,.
.
.. fr, ,auer icvn.an. ior til" b .. ra' ,ure T
i on m ~gn .. ie ield str n~h
B.
~orm 1 0 erat ~on f.n.
arc.11.(
05r dis 0 T (e V'.
~
...
. /
6
~
/-"
5 ~V
, -~nJ-l
~
4 2~
V
~
•
1'1'.
~
~ 1 2l1
ICA)
5
1 0 150 2C 0 2 03PO
~ig. De. ende ~c. f ar 19u1a r fr lquer c,..n •• an iOI t ... " .. ra lire Ti cn a Ire: c rrel t I . !lor! al. c p.r. ion of t
IBe
a re..-...-r'"
~r
p2P-"3
..
CI thOd"~ _t
tAnodl...
-
. _.- ---r.,
I
I
II
1
-j
I
I
I
!
I
I
p ,-, (Torr) 10-1 --- - - f--_i( 4 2 1 -4'0
--L-
I
i ---lCT
--_._---- --I~
-I
- - - _ . P1n~. 10a. A,.~on gas
P2
P,
~i
, //
/
10 20/
pr.""",,. ..
m .... sured :.ith i(')niz .. tion ~"up;es) at vari(,)UR plac .. s in the vacuum ch .. mber. G .. ~ reed through the cathode; RtRnd .. rd con~ttions./
/ '
T(A) 30 40 50
F't;.10b.Arll:On lI;as pressure inaide hollow anod., all function of lire current I. G •• reed throu!l;h cathode; standard conditions.
---r-0.2 0.1 3 2 1
v--
...
z=1 bOcm Errolr
7
/
. /
"""
'''= c;r .~17
7
/
[7
7
i 1 23
4 5 6 r( em)Fi~.12
Static enrichment in 36A isotope as function of radius. Pure argon arc;Q=4.5cm 3NTP/s;L=120 cm.z=1 pOcm
1F. ....
n~V
V
/"
z=5 pcmV
V
1.--V
I
1 2 4 5 6 r(cm)Fig.13 Static enrichment in 36 A isotope as function of rsdius. Hydrogen arc + argon. Q=10 cm 3NTP/s hydro-gen + 0.5 cm3NTP/s argon; L=120 cm.
..
ROtATING HOLLOW PLASMA CYLINDER
_.~ ______ gas depleted
gas to be
---:::~
..
~:.:::::::---separated
... ---in light component
..
cathode
j
gas depleted in light componentFig.1!, Scheme of "PlaRma Mantle" method
gas enriched
--~===~.:
in lighteu
insert'fa shield; Tungsten wi re
Breed
reed
wtI_::=H:~-~---~
A B
~u wed
Fig.15 The C2 cathode with additional injection of neutral gas in the centre.
PI c:
""
•'"
o a 00 C7\(::~:::~~~t-':!LIl:""'i
It-'-:!..S.II
~.J.ln't"IX:
••_~
i-
jr
1 _
...
t~-+-"-t---I---t-~\---,.-+,
-+--·+--+D+,
-t-+--+-+-120~ __+-__
b-~~r+ __ ~ __ ~~+-+-+-__
~~__
~.
II
10<UU"'r--t--lHt
T---l\H--t-
i
~--tJ!;~JJ---lrIf\
-tt ..
-+!
--+1--~:
v
-..
4 \ \ \ 2 + .. ~.--.----+.~~-+--~f---+~\..-+-V .
1. ___
~
__
+_~+---+~~+
____
I_;\:---+--+-•
40
5n
60 70 110 seannin~ distanc~ (mm)e
! ?()(': A; z..;..!' em;T;d;.P.
eV.,
!=10f" A;z7O cm:T,' 3.f.
eV. •~ r:~ '0(\ A· • z.": hO"'m· ., '-' , i . T _-.., eV.
- - . - - lntensHy Af 4201 ~ line(x10n);T =3.1) eV.
n
~i~.16h. Hadial prD~i Ie of th~ azimuthal mass
velocity, Vt' of an arp:on arc operated
w1th tr.p C? cathode.
Fi@t.16a. Radial profil .. of the int .. nsity of the All 4Qo{.~ lin .. in an ar~on arc operated with th .. C2 cathode
