A AEC/E567
AUSTRALIAN ATOMIC ENERGY COMMISSION
RESEARCH ESTABLISHMENT
LUCAS HEIGHTS RESEARCH LABORATORIES
A REVIEW OF LASER ISOTOPE SEPARATION OF
URANIUM HEXAFLUORIDE
by
J.W. KELLY
April 1983
ISBN 0 642 59772 3
AUSTRALIAN ATOMIC ENERGY COMMISSION
RESEARCH ESTABLISHMENT
LUCAS HEIGHTS RESEARCH LABORATORIES
A REVIEW OF LASER ISOTOPE SEPARATION OF
URANIUM HEXAFLUORIDE
by
J.W. KELLY
ABSTRACT
There is continuing world-wide interest in the possibility of enriching
uranium by a laser process which uses uranium hexafluoride. Since no actual
commercial plant exists at present, this review examines the key areas of
related research. It concludes that such a process is feasible, that it must
employ an adiabatic cooling system, with UFg the minor constituent in a
predominantly monatomic or diatomic carrier gas, that the necessary infrared
and/or ultraviolet-visible lasers are in a state of development bordering on
the minimum required, and that the economics of such a process appear highly
promising.
National Library of Australia card number and ISBN 0 642 59772 3
The following descriptors have been selected from the INIS Thesaurus to
describe the subject content of this report for information retrieval
purposes. For further details please refer to IAEA-INIS-12 (INIS: Manual for
Indexing) and IAEA-INIS-13 (INIS: Thesaurus) published in Vienna by the
International Atomic Energy Agency.
COOLING; EXCITATION; EXPANSION; GAS LASERS; INFRARED RADIATION; INFRARED
SPECTRA; LASER ISOTOPE SEPARATION; MATRIX ISOLATION; ULTRAVIOLET RADIATION;
ULTRAVIOLET SPECTRA; URANIUM HEXAFLUORIDE; URANIUM ISOTOPES; VISIBLE
RADIATION; VISIBLE SPECTRA
CONTENTS
1. INTRODUCTION 1
2. SPECTROSCOPY 2
2.1 Infrared Spectrum of Gaseous UFg 3
2.2 Ultraviolet-visible Spectrum of UFg 4
3. NON-EQUILIBRIUM COOLING BY ADIABATIC EXPANSION 5
3.1 Introduction 5
3.2 Ducted Nozzle Expansion of UFg 6
3.3 Supercooling and Condensation 7
3.4 Practical Considerations of Adiabatic Cooling 8
4. INFRARED LASERS FOR LIS IN UFC 9b
4.1 Introduction 9
4.2 Difference Frequency Lasers 10
4.3 Parametric Oscillators 0
4.4 Raman Processes in Liquids and Gases 11
4.5 Optically Pumped and Energy Transfer Systems 12
4.6 Other 16 ym Lasers 15
5. EXCIMER LASERS 6
5.1 Introduction 6
5.2 Developments in Rare-gas Halide Lasers 17
5.3 Developments in Halogen Lasers 20
5.4 Tuning UV Radiation to UFg Transitions 20
6. LASER IRRADIATION OF UFg 22
6.1 Matrix Isolated UFg 2
6.2 Gas Phase UFg 3
7. CONCLUSIONS 27
8. REFERENCES 8
Table 1
Table 2
Table 3
Table 4
Major absorption bands of infrared spectrum of UFg
Gas dynamic supersonic nozzle parameters
Performance data on some excinier lasers developed in period1976-1980
Cost structure of 9 x 106 SWU/y capacity enrichment plants(1976)
37
37
38
39
Figure 1
Figure 2
Figure 3a
Figure 3band 3c
Figure 4a
Figure 4b
Figure 5
Infrared absorption spectrum of UFs at room temperature 41(after Jetter 1975)
Population of ten lowest vibrational states vs. temperature (K) 41(after Jensen et al. 1976)
Absorption spectrum of UFs cooled by expansion through anozzle (after Jensen et al. 1976)
Absorption spectrum of Sf^ at 55 and 300 K (after Jensen
et al. 1976)
Onset of SF5 condensation vs. temperature
Effect of nozzle design on condensation pressures(after Wu et al. 1978)
UV/VIS absorption spectrum of UF5 (after De Poorter andRofor-de-Poorter 1978)
42
42
43
43
44
1. INTRODUCTION
Information on the separation of the uranium isotopes by laser isotope
separation (LIS) processes using uranium hexafluoride (UFg) emanates primarily
from the various institutes of the USSR Academy of Sciences (Ambartzumian et
al. at the Spectroscopy Institute, Basov et al. at the Lebedev Physics
Institute), the Los Alamos Scientific Laboratory (LASL) in the United States
of America (Robinson et al.), from numerous patents mainly of German origin
and most recently from the Exxon Laboratories who are better known for their
studies on LIS using atomic uranium.
However, so significant are the commercial and political aspects of this
area of work that the Americans and, to a far greater extent, the Russians,
have talked of their work on the LIS of UFg only in general terms, and patents
have disclosed little of quantitative value, being concerned mainly with the
enunciation of principles. Apart from the general assertion that the LIS of
UFg must work, the only quantitative support for this belief, is a three line
statement from the Los Alamos Laboratories in 1976 to the effect that they had
separated the isotopes on a laboratory scale using an unspecified two-laser
process and a comment* from Los Alamos staff that they were passing a two-
laser process over to the Oak Ridge Laboratories for pilot plant development.
Hence, this assessment of the LIS of UFg is not based on an evaluation of
direct data from a demonstrated commercial process, there being no such
process at present. Instead, it uses three criteria (the most important of
the many relevant criteria) to evaluate the significance of work in a wide
spectrum of research areas for a possible UF,. based process.
The three criteria used here for an efficient, economic UFfi-based LIS
process are that there must be
(i) a system with appropriate spectral conditions: for all practical
purposes this means that the separation must be carried out in the
gas phase where isotopic differences in the absorption spectrum are
most likely to be resolved and preserved;
/'
(ii) lasers of appropriate wavelength, efficiency, power, spectral
purity, repetition rate, reliability and (low) capital cost; and
* Made to Dr A Ekstrom, AAEC, during overseas visit, June-July 1980.
(iii) an efficient photochemical process which preserves the laser-
induced selectivity under practical conditions.
In Section 2 of this report is a review of both the most recent work on
the infrared spectra obtained under conditions of adiabatic cooling, and
ultraviolet-visible (UV/VIS) spectra obtained under conditions of matrix
isolation, as well as the more classical aspects of this subject. Section 4
deals with the adiabatic expansion (flow cooling) technique. While flow
cooling has been known for many years, its use for improving spectroscopic
resolution is quite recent, and the problem of achieving high gas densities
without condensation in supersaturated streams has only recently been
investigated*.
Section 4 treats infrared lasers, particularly those in the 16 ym
spectral region, and assesses the many totally different approaches, studied
or under study, to a practicable, high powered, high repetition rate, 16 pm
laser. Since excimer lasers are by far the most promising candidates for a
suitable UV/VIS laser (_< 410 nm) these are discussed in Section 5. Finally,
Section 6 deals with the laser irradiation of UFg and Section 7 contains the
conclusions of this assessment.
2. SPECTROSCOPY
Advantages in the use of UFg as the process gas for LIS include: (a) UFfi
has a significant vapour pressure at and below room temperature, (b) isotopic
spectral differences are due entirely to the uranium atom since fluorine is
monoisotopic, and (c) the technology of producing and handling the gas on an
industrial scale exists. It should also be noted that a very high degree of
enrichment of the sulphur isotopes has been obtained by Basov et al. [1975]
for the analogous SFg molecule. For these and other reasons, the spectroscopy
of UFg continues to hold great interest.
* In 1976, Kelly and Struve discussed the technique from a spectroscopicviewpoint following the Los Alamos disclosure [Nucleonics Week, 6 February
1976]: In Section 3 of this report, it is the gas dynamics of thetechnique which is of primary concern.
2.1 Infrared Spectrum of Gaseous UFg
The infrared spectrum of UFg has been determined by a number of authors
[e.g. Bar-Ziv et al. 1972]. The rain absorption bands are listed in Table 1.
The v^ (16 ym) absorption band, being the strongest, is the centre of interest
for LIS studies of UFg.
32 34Unlike the SF,- and SF,- vo absorption bands which are fully resolved
235 ?38at room temperature, the v
3 bands of UFg and "?UFg are not. Figure 1
shows that, although the isotopic shift of x 0.55 cm is clearly visible at
room temperature, it is much smaller than the FWHM (full width at half
maximum) of the band.
The fractional population of the 10 lowest-lying vibrational states of
UFg is shown as a function of temperature in Figure 2 for 30 <T< 170 K, from
which it is clear that < 10 per cent of the molecules are in the ground state
at 170 K. Less than 1 per cent are present in the ground state at room
temperature. By comparison, SFg has ^ 30 per cent of the molecules in the
ground state at room temperature. Thus, the room temperature spectrum of UFg
consists of thousands of superimposed hot bands, and lowering the gas
temperature is an obvious way of simplifying the spectrum. Hot band
conditions in UFg at 120 K are comparable with those in SFg at room
temperature.
The spectrum of UFg cooled to 50 K by expansion through a nozzle is shown
in Figure 3a, together with the room temperature spectrum. Because of signal
to noise limitations, only the Q branch of the spectrum can be seen. The
spectra of SFg with and without cooling in this way are shown for comparison
in Figures 3b and c. These spectra were all taken with a conventional
spectrometer. Higher resolution spectra of UFg cooled by nozzle expansion
were obtained with tunable diode lasers and show Coriolis splitting of
individual lines. The temperature of Q band resolution is such that a
substantial fraction (?*, 30 per cent) of the molecules can be made to interact
with a single laser frequency, provided the effective laser band width is
comparable to the width of the Q band.
Nozzle expansion of UFg is clearly a means by which the gaseous isotopic
spectra of UFg can be fully resolved at pressures having a practical
significance (Section 3). If, by contrast, UFg is cooled conventionally
(equilibrium conditions) to temperatures where similar spectral resolution is
attained, it becomes a solid with negligible vapour pressure. Thus, only
non-equilibrium dynamic cooling is practical for LIS methods which rely on a
temperature-assisted resolution of the absorption spectra.
2.2 Ultraviolet-visible Spectrum of UF.-1 (3
The UV/VIS absorption spectrum of gaseous LIFg has been studied in a
number of laboratories for more than 35 years, e.g. Lipkin and Weisman [1942],
Hurst and Wilson [1972] and De Poorter and Rofor-de-Poorter [1978], but none
of the data obtained had sufficient structural detail to justify more than a
superficial analysis of the spectrum.
The first detailed analysis of the structure of UFg was published in 1976
by Lewis et al. at the LASL and based on their observations of complex
vibronic structure on the broad absorption features of UFg which was matrix-
isolated in argon at 14 K. They were not concerned with isotopic structure
and their results did nothing to dispel the usual notion that if UFg is to be
used for LIS, then a two-photon process (infrared to discriminate, UV/VIS to
dissociate) is essential.
Subsequently, similar work by Grzybowski and Andrews [1978] at lower
solute/solvent ratios found strong, sharp, highly structural fluorescent
spectra in samples where the solute/solvent ratio (UFfi:A) was 1/1800. They
also observed that this structure coalesced into broad bands when the
solute/solvent ratio was greater than 1 in 800. Their results demonstrated a
strong solute-solute interaction and also explain why Lewis et al. [1976]
failed to observe highly structural spectra, their solute/solvent ratio being
1 in 20. The results obtained by Grzybowski and Andrews have been confirmed
by Miller et al. [1979] who further noted that the line width was < 3 cm" in
argon and krypton matrices at 12 K. These results strongly suggest that the
first excited electronic state of UFg is bound (stable). Corroborative
evidence for such a state is provided by Andreoni et al. [1977] and Benetti et
al. [1976] who observed fluorescence in gaseous UFg at room temperature.
These latter studies, which explored the dependence of the fluorescence decay
and quantum yield on pressure, temperature and excitation wavelength also
provided valuable information on the photochemistry of UFfi.
The existence of a bound (stable) excited state is highly significant.
In the past, isotopic resolution of the electronic spectrum of UFg has not
been possible, and this can now be attributed to spectral congestion resulting
from hot band transitions rather than spectral broadening associated with an
unbound excited state. We believe that reducing this spectral congestion
through adiabatic cooling to low temperatures (see Section 3), enables the
spectrum of gaseous UFC to be isotopically resolved. It is intriguing that0
high resolution UV/VIS spectra of UFg under these conditions have never been
reported in the literature. Consequently, there is a distinct possibility
that uranium enrichment can be achieved by a process in which a single UV/VIS
photon will both isotopically discriminate and photo-dissociate a UF..
0molecule.
3. NON-EQUILIBRIUM COOLING BY ADIABATIC EXPANSION
3.1 Introduction
Greene and Milne [1970] pioneered the use of adiabatic cooling techniques
in their low temperature work on simplification of the microwave spectra of
condensable gases. This was followed by Smalley et al. [1974] who made a
laser spectroscopy study of the high resolution fluorescence spectrum of N0?
in supersonic molecular beams generated by free jet expansion. In 1976, in a
brief note in the February issue of Nucleonics Week, Los Alamos workers
claimed that they had isotopically resolved the infrared spectrum of UFfi at
50 K in a Mach 5 gas stream. This claim was analysed by the present authors
(unpublished work) who concluded that these temperatures could not have been
achieved in a Mach 5, free jet expansion of pure UFfi. Subsequently, Jensen et
al. [1976], also from Los Alamos, revealed that such low temperatures at
these modest Mach numbers were achieved by the co-expansion of UF,. in ao
carrier gas under ducted flow conditions.
Apart from cooling, the expansion process must simultaneously generate
gas densities sufficiently high to be of commercial interest (the vapour
pressure at temperatures low enough for isotopic resolution (< 60 K) ison
negligibly small (< 10 Pa))- In free jet expansion, gas densities
1p O< 10 c cm"0 are unlikely to be exceeded, and the resulting long irradiation
paths and high pumping costs make such an approach uneconomic. However, the
situation appears to be quite different for ducted flow for which gas
densities < 10^ cm"^ at 50 K have been claimed by Los Alamos workers
[Robinson et al. 1976]. Such claims appear to be reasonable in view of the
published work on SFC.b
3.2 Ducted Nozzle Expansion of UF,-
The basic ducted nozzle is a converging/diverging channel and the general
equations (for example, see Shapiro [1952]) governing the flow of a perfect
gas in such a channel under abiabatic, isentropic conditions are
(2)
and
where y = the ratio of specific heats, M = Mach number, T p and n are the
temperature, pressure and gas densities above and T, p and n the same
parameters below the throat of the nozzle.
Two sets of values of these parameters are given in Table 2 covering the
range M = 2 to 7. The upper set (y = 1.67) corresponds to the flow of
monatomic gases and the lower set (y = 1.40) to the flow of diatomic gases
(also in the table are At and A, the area of channel at and below the throat
of the nozzle, and the stagnation pressures P and R (see Section 3.4)).
These values show clearly that temperatures of the order of 50 K reported
by Jensen et al. [1976], are readily obtained in mono and diatomic gases at
modest Mach numbers (M = % 5), which are well within the capability of
present-day wind tunnel design, while the pressure ratios p /p are markedly
less than those considered necessary (-v 10 ) in free jet expansion. Further,
since the gas density, not the pressure is significant in a separation
process, it is important to note that the ga:
value of M is lower than the pressure ratio.
s density ratio (n /n) for a given
Since the ratio of the specific heats of UFg is close to unity (y = 1.06
at room temperature), equations (1) and (2) predict that very high,
impracticable, pressure ratios would be required to achieve a beam temperature
of 50 K. Hagena and Henkes [1960] established an experimental value of
T = 120 K for a Mach 5 expansion of pure UFg which is in reasonable agreement
with that predicted by the above equations when allowance is made for the
increase in ywith decreasing values of T. This difficulty can be avoided by
the co-expansion of UFg in either a monatomic or diatomic carrier gas at small
molar fractions of UFg. A 1 per cent molar fraction of UFg reduces the
Y value of a monatomic and a diatomic gas to l-Slg and 1.382 respectively,
i.e. by less than 4 per cent, so that the temperature (T) achieved is
essentially that produced by the carrier gas alone. A 10 per cent mixture
reduces y to 1.357 which is still sufficient for effective cooling.
3.3 Supercooling and Condensation
The minimum gas density attainable in a supercooled beam of UFg contained
in a carrier gas is set by various condensation processes. While condensation
of pure gases in supersonic nozzles has been studied over many years (for
example, see Wegener and Wu [1977]), very little information is available on
nucleation within dilute gas mixtures.
In 1978, Wu et al. published the results of experiments in which dilute
mixtures of SFg in argon were cooled by expansion at low Mach numbers and the
onset of condensation determined by Rayleigh scattering of light from a He-Cd
laser downstream of the nozzle. They showed (Figure 4a) that the condensation
pressure p, had a temperature dependence similar to that of the equilibrium
vapour pressure (pm) and that condensation densities of up to 10 cnf and 4
x 10 cm could be obtained at temperatures of 100 and 50 K respectively.
Figure 4b shows the effect of nozzle design on the condensation pressure.
Similar work on the expansion of SFg (and presumably UFg) is known to have
been carried out at Los Alamos (see references in Kim and Person [1978a]) but
the results have not been published. Wu's work was limited to M = 2.8 (max.),
while the Los Alamos workers presumably used values of M = -\. 5.
The factors governing the onset of condensation are not clearly
understood but it is known that the degree of supercooling attained is
dependent on the rate of cooling, with the higher cooling ratios favouring
condensation processes, and independent of the molecular weight of the gas, at
least to the first order. Thus, it is probable that condensation pressures
comparable to and possibly higher than those of Wu are attainable by the
nozzle expansion of dilute mixtures of UFg in argon at Mach 5. Expansion in
He should give higher p. values than for expansion in argon because of the
higher supersonic beam velocity. Thus, the results of Wu et al. provide
valuable corroborative evidence for the claim made in the Los Alamos patent of
Robinson et al. [1976] that a 10 per cent mixture of UFg in helium could be
cooled to 30 K at a gas density (of UFg) of ??. 5 x 1016 cnf3.
Note also that (i) higher gas densities are possible if sufficient
spectral resolution is available for a process at higher temperature, and (ii)
higher cooling rates can be achieved by the use of strongly divergent nozzles
and/or a low molecular weight carrier gas.
3.4 Practical Considerations of Adiabatic Cooling
In the apparent absence of any detailed estimates of the cost of a UF,
laser LIS process, there has been agreement that the pumps (and pumping costs)
were likely to be a major cost component; higher values are expected for P0/p,
since the pumps would have to encompass both the laminar and molecular flow
I /? oregimes. The realisation that gas densities of n = x 10 cm at p /p values
of 100 to 500 at Mach 5 are attainable has clarified the situation and makes
it clear that pumps need to be considered for only the laminar flow regime.
In a practicable adiabatic expansion system, gas dynamic temperatures of
the (carrier) gas downstream of the expansion nozzle would be used to minimise
pumping costs. Under these conditions, it would be the stagnation pressuresi
(P upstream) and P (downstream) of the system which determine the
compression ratio of the gas recirculation pumps. Table 2 lists values of
P /P for a system in which the expansion nozzle is followed by a divergent
diffuser, and the system operated under conditions of normal shock recovery.
For M = t 5 and T = -u 50 Ks the compression ratio is of the order of 10 (more
sophisticated systems give even lower values).
Roots blowers fulfil these requirements; they operate in the laminar
flow regime and have compression values of this order of magnitude. In many
ways they are ideal for this application; they operate in an oil-free
environment, have minimum wear because the pumping surfaces are not in contact
with the pump walls, and are commercially available in a form suitable for
handling UFg. A full analysis of pumping costs is beyond the scope of this
report but a cursory estimate suggests it would be less than $1000 p.a. for a
facility employing a 1 watt laser.
Finally, it is important to consider the effect of gas density on the
optical path required for efficient use of the laser irradiation. The optical
path is inversely proportional to both gas density and adsorption cross
section and, for free jet expansion, it was estimated to be impracticably1 /- o
long. For ducted expansion in which n - . 10 cm~? and 63 per cent of the
p *3 C
radiation is absorbed by UFg molecules, the optical path length was
estimated to be ?v 10? m. This is still long but not impracticable; it can be
decreased by operation at higher values of n but the upper limit of n is a
complex function of many parameters, including mechanical, spectroscopic,
photochemical, etc.
4. INFRARED LASERS FOR LIS IN UFg
4.1 Introduction
Granted that the spectral conditions for LIS in UFfi can be provided by
adiabatic croling, the next important requirement is a laser that will operate
reliably for long periods at the required wavelength, band width, repetition
rate and power level. For commercial application, the presently favoured
two-photon approach (see Section 6) requires infrared lasers of at least 100
watt average power (at 16 ym), UV/VIS lasers of -v 40 W average power and
repetition rates of ?*. 10 kHz. For multiphoton dissociation, energy fluences_o
of 1 - 10 J cm are required and at high repetition rates.
Semiconductor diode lasers (see, for example, those described by Hinkley
et al. [1978]) have been used extensively for the generation of infrared
radiation, but are restricted to spectroscopic studies and laboratory
experiments concerned with 'demonstration of principles'. Many other optical
pathways have been explored in the search for an infrared laser suitable for
LIS of UFg: these may conveniently be grouped under four headings:
(i) difference frequency generators,
(ii) optical parametric oscillators,
(iii) coherent resonant Raman devices, and
(iv) optically pumped devices.
Because of the large number of laser transitions which overlap known UFfi
absorption bands, Doppler fine tuning may be sufficient for efficient optical
coupling with (iii) and (iv) type devices.
10
4.2 Difference Frequency Lasers
These lasers use non-linear optical effects in solids to generate
continuously tunable infrared radiation. The limitation of this class of
laser can be illustrated by reference to the Los Alamos work [Piltch et al.
1975] on the mixing of CO and COj, radiation in single crystal chalcopyrite
(CdGeAs,,) and which produced 40 \iJ of 16 urn radiation. The device suffered
from the fundamental difficulty that the CO laser is inherently a long pulse
device (the 5.8 ym line having a characteristic pulse width of 3 ys) whereas
the C02 laser pulse is typically of -v 0.5 ys duration. Thus temporal overlap,
and hence efficiency, is low. Development of a short pulse-length CO laser is
required if reasonably high efficiency is to be obtained. If the COp pumping
laser is operated under normal pressure, the output is line tunable; high
pressure operation of the CCL laser allows continuous tuning.
Another basic problem with this class of laser 7"s the difficulty of
obtaining large mixing crystals of good optical quality- In general, inherent
crystal inhomogeneities cause broadening of the tuning curves and orders of
magnitude differences in mixing efficiency from point to point across the
crystal faces. Recent studies by Feigelson et al. [1980] indicate that the
problem can be ameliorated by self-annealing, but it remains unlikely that
this class of laser will be scalable to pulse energies beyond a few hundred
microjoules, while bulk heating will probably limit operation to low average
powers.
4.3 Parametric Oscillators
The first optical parametric oscillator in the infrared was that of
Giordmaine and Miller [1965] who used a crystal of LiNbO., pumped at 532 nm to
generate radiation from 0.96 to 1.15 ym. Byer et al. [1974] of Stanford
University subsequently used an Nd:YAG laser to pump LiNb03 and so produce up
to 1 mJ of radiation continuously over the range 1.4 to 4.4 ym.
The first parametric oscillator at 16 ym was that of Menzel and Arnold
[1976]. They used the 2.87 ym line from an HF oscillator-amplifier to pump a
CdSe crystal and produce idler outputs of 100 yJ, with band widths of 1.8 to 9
cm over the tuning range and pulse lengths of ?*. 100 ns.
Later, Andreou [1977] and Hyer et al. [1979] separately described a
similar device consisting of a Nd:YAG driven, LiNbO^ parametric oscillator
11
with non-linear mixing of the signal and idler wavelengths in CdSe. The
radiation had a peak power of -\. 5 kW, pulse energy of -v. 5 pJ when grating
tuned, linewidth of -- 2.6 cm" , and pu
operated at repetition rates up to 20 Hz
 pulse duration of -v 10 ns. The laser
Since a prime motivation for Andreou's work was the generation of narrow
(x 0.5 cm) linewidths without the use of high loss etalons, it is apparent
that this goal is difficult to obtain. In addition, the tuning curves v;ere
very sensitive to small changes (10 ) in the refraction index so that the
tuning curve varied within and between crystals. Parametric oscillators
therefore, like differance generators, cannot be readily scaled up.
4.4 Raman Processes in Liquids and Gases
In 1976, Kung and Itzkhan generated 8.5 and 16 pm radiation by stimulated
electronic Raman scattering from potassium vapour in a heat pipe excited by
two dye lasers operating at 404.6 and 582.2 nm, and produced an estimated
10 mW of 16 pm radiation at 1 kW pumping power.
Frey et al . [1977] utilised the large Raman gain in liquid nitrogen,
pumped by a ruby excited dye laser to generate, by stimulated Raman scattering
(SRS), a continuously tunable output up to 1.7 mJ/pulse, having a bandwidth of
x 1 cm" and a photon scattering efficiency of 20 per cent.
The potential for high energy conversion efficiency and scale-up is the
main reason for interest in SRS from gases. Byer [1976] proposed such a
system using rotational Raman scattering from FU gas at 77 K pumped by a C0?
laser, and predicted a 10 per cent energy conversion efficiency. Similar
predictions were made by Konev et al . [1977].
Efficient conversion from pumping radiation to SRS radiation requires
very careful optical design because of the low Raman gain of FL at 16 ym and
because pump thresholds are higher than those predicted theoretically (see
Laser Focus, April 1978). However, these problems have been overcome by
Rabinowitz et al. of the Exxon Laboratories who, in 1980, revealed that they
had generated 1.6 J, 20 MW pulses at 623 cm"1- by Raman shifting (XL radiation
in hydrogen at an efficiency of 55 per cent (maximum theoretical efficiency =
64 per cent). Thus, with a C02 laser efficiency (electrical input to optical
output) of ^ 20 per cent, the overall efficiency is % 10 per cent as
predicted. This is a major step forward. The principal remaining problem is
12
Athe need to increase the repetition rate from 1 to about 10 Hz.
4.5 Optically Pumped and Energy Transfer Systems
Optical pumping of gaseous molecules in the infrared is analogous to
pumping of dyes in the visible, and is one of the simplest and most direct
means of inducing new laser transitions in mid-infrared. Highly reliable and
efficient lasers (CO, (XL, HF, etc.) are available for pumping; the main
problem is that of establishing coincidence between the pump and absorption
lines of gases with suitable laser emission. Optically pumped lasers are
almost always line rather than continuously tunable. Where a suitable laser
gas cannot be directly excited, optically pumped lasers have been developed
which depend on collisional energy transfer.
4.5.1 Sixteen urn lasers pumped by HF lasers
One such laser system has been assembled by Buchwald et al. [1976] who
used an HF laser for direct pumping of various isotopes of C0?. Numerous
lasing lines in the 4.3, 10.6 and 17 ym regions were obtained, and the system
was said to be capable of scale-up. However, while the HF laser is a high
power source emitting up to 10 MW/pulse, its output between 2.6 and 2.0 ym
contains up to 25 lines, and it does not operate efficiently on a single line.
Consequently, HF pumped systems are limited to non-critical matching systems.
4.5.2 Sixteen ym lasers pumped by HBr and CO
Osgood [1976] used an HBr laser to pump a resonant absorber (HBr)
followed by collisional V-V transfer to C0? to generate 'useful' 16 ym
radiation. However, HBr pumped lasers appear to have limited potential for
scale-up because of the relatively modest (-v 100 mJ) outputs currently
available.
Pulsed CO pumping lasers can deliver a few joules output but, as noted in
Section 4.2, their long pulse lengths lead to poor temporal overlap and hence
low overall efficiency.
4.5.3 Sixteen ym lasers pumped by COp lasers
The COp laser can be operated at high powers on many single wavelengths
in the 9 and 10 ym bands, and so can be used to pump many molecular
13
transitions. It is the most important pumping laser developed to date for the
generation of 16 um radiation.
Efficient energy down-conversion (to 16 ym) requires at least four
conditions to be met. They are:
(i) saturation of the lasing transition by a high-powered pump matched
to the absorption line,
(ii) establishment of the resonance condition,
(iii) creation of a suitable absorption path, i.e. optical cavity, and
(iv) constraints on the optical density of the medium arising from
collisional deactivation.
Laser systems which have been developed and which attempt to meet these
conditions may be diviued for convenience into two categories, the single-
photon and the two-photon pumped.
Single-photon pumped, 16 ym lasers
Although not a 16 um laser, the (XL-pumped 12 ym, NH~ laser must be
included because of its potential use in UFg irradiation. By pumping one of
the lines of the fundamental, v2 vibrational band of NH- with a CCL-TEA laser
(R16 line of 9.4 ym band), Chang and Magee [1976] induced population inversion
with respect to the ground state and generated, at ?*. 12 ym, 5 kW pulsed output
at 1.5 MW pump power. This is the greatest power yet achieved in an optically
pumped molecular laser. An energy conversion efficiency of 11 per cent has
been reported for this laser [Tiee and Wittig 1978a] which has since been used
to generate 16 ym radiation in an InSb spin flip Raman laser. Thus, a single
C02 laser, which could be scaled up to achieve gigawatt levels via current
state-of-the art technology, can generate radiation close to the 16 ym (vo)
and 12 ym (v3 + v5) absorption bands of UFg.
However, the most important C02 molecular infrared laser available was
first described by Tiee and Wittig at the 1978 Atlanta Conference [1978a,b].
They excited the v~ + v^ level in CF., producing an (erratic) multiline output
of >_ 100 mJ for a 10 J input and achieved a line tunability between 612 and
653 cnf and a conversion efficiency of 5 per cent in their basically simple
14
configuration. This was, at that time, the most powerful 16 ym laser
available, and as such it generated intense developmental effort.
Experimental efforts to overcome the instability problem are well under
way. Stein et al. [1978] have used an intracavity low pressure gain cell in a
high pressure C02 laser cavity to induce frequency, narrowing of the pump.
This, together with telescopic matching of the pump outlet to the lowest order
mode of the 16 ym resonator, has increased the energy output per pulse from
0.5 to 15 mJ on the 615 cm transition, with a quantum yield of 15 per cent.
The output energy increase depends in part on the frequency displacement of
the narrowed pump line for the particular absorption feature. In multiline
operation of CF,, P and R branch lasing is much stronger than Q branch lasing
[Jones et al. 1978].
In addition to the instability problem, other aspects of this laser that
must be considered, if it is to be commercially used for uranium enrichment,
include:
(a) The ability to fine-tune with respect to a UFg transition (cf. the
spectral studies of Kynazev et al. [1979] on weakly forbidden
vibration-rotation transitions in CF^ in relation to continuous
tunability in the 16 ym region);
(b) Cavity design, including the method of introducing the pump beam
into the resonator (these configurations usually involve multiple
paths because of the low absorption coefficient of combination bands
and the need to avoid long cavities; for example, see Telle [1979]);
(c) The repetition rate, which governs the average power and hence the
laser capital costs of the enrichment process (Baranov et al.
[1978] studied the dependence of output energy on pump energy, gas
pressure and composition at repetition rates up to 100 Hz);
(d) Development potential for increasing output (Baranov et al. [1978]
have observed saturation of the output at 0.1 J with a peak
efficiency of ^ 1 per cent. Since the CF. laser is a low gas
pressure laser, there is reasonable doubt that significantly higher
outputs at high repetition rates are possible).
15
While the CF^ laser is, clearly, a simple and effective device, its
development into a commercial model will, if possible, require a substantial
effort. Similar- considerations apply to two other 16 urn molecular lasers, the
nitrogen-sulphur-fluorine (NSP) laser of Fischer [1980] and the perchloryl
fluoride (FC103) laser of Rutt [1980].
Two-photon pumped. 16 urn lasers
The first experimental demonstration of two-photon pumping to generate
16 ym radiation was that by Barch [1975] who used the non-linear SF,. molecule.o
In this case, off-resonant pumping of the intermediate state was possible
because of the several SFg lines existing within the 720 MHz bandwidth of COp
laser 10 P(14) pump line. Power output was not given, but Barch noted that it
decreased rapidly with increasing pressure, indicating that lasing was from an
excited combination band. High efficiencies were predicted for an optimised
system.
Another two-photon laser based on the excited vibrational-rotational
levels of NH^ was described by Rummer et al. [1978], following earlier work
by Jacobs et al. [1976]. Increased attention was paid to the optimisation of
the system by changing pump intensities, relative beam polarisation, pulse
length, NHg pressure and irradiation path length. Four wavelengths were
generated at 12.11, 13.72, 13.85 and 15.95 ym. The first two can be
suppressed by suitable intracavity absorbers, such as (CH-CU)?- Off-
resonance pumping by as much as 300 MHz was evaluated using the optical Stark
effect. Output was 1 mJ at >_ 10 per cent conversion efficiency and a maximum
of 30 per cent was predicted. This system appears to have the advantages of
potential for scale-up and high pulse repetition rates, and to be capable of
average powers of ?*. 1 W.
4.6 Other 16 urn Lasers
Free electron lasers have attracted considerable interest recently
because of their potential for reaching high power levels and because they
can, in theory, be continuously tuned by changing the electron energy. These
lasers rely on an electron beam which passes through a spatially varying
transverse magnetic field where energy is extracted by scattered photons. To
date, most work has been concerned with the physics of such lasers but
amplification has been observed at 10.6 ym, oscillation at 3.5 ym and
stimulated super-radiant emission at 400 ym (see Laser Focus, August 1980,
16
p.72). It is not unlikely that studies of a 16 ym laser of this kind are in
progress, but none has so far appeared in the literature.
A laser of a totally different kind has been described by Wexler and
Wagnant [1979]. They studied an electric discharge gas dynamic laser in which
an N?/He mixture was supersonically expanded and then mixed with COg to
generate 16 urn radiation in a 5-pass optical cavity configuration. They
obtained a 5 W output at 4800 Hz (the highest repetition rate yet obtained) at
an output energy per pulse of 1.1 mJ. This system was reliable and with
further development to improve the efficiency (0.1 per cent) and peak power
obtained, could meet the requirements for the LIS of UF . In a separate
theoretical study Suzuki et al. [1980] predicted that high power levels at
high repetition rates can be obtained from a conventional gas dynamic C0?
laser in which a fast 9.4 ym pulse is injected downstream of the throat nozzle
in order to populate the upper (16 ym) level. Clearly, there is room for
further development of this class of laser.
5. EXCIMER LASERS
5.1 Introduction
Copper/copper halide lasers which have the very desirable characteristics
of high output power, high efficiency and high repetition rate, and which are
used in the atomic LIS process have, unfortunately, too long a wavelength for
UFg. Generally, UV/VIS lasers for use with UFg must operate at wavelengths _<
430 nm. Of the various lasers which operate in this region, the nitrogen
laser (x = 337 nm) can be immediately excluded because of its low overall
efficiency (?*. 0.01 per cent). Likewise, flash lamp pumped devices are not
discussed since their development seems to have foundered on the problem of
adequate lifetime at high powers and repetition rates of interest.
The only class of UV laser which has the potential for adequate
efficiency (_>. 1 per cent) at high powers and high repetition rates is the
excimer laser. With few exceptions, these lasers are line tunable only so
that, barring a fortunate line coincidence, they will most likely serve as the
pump for some other conversion device, e.g. a dye laser or possibly a
stimulated Raman scattering device for up or down conversion of the pump
radiation.
17
The three criteria which determine the overall efficiency of a UV/VIS
excimer laser are that
(i) the radiation life-time of atoms or molecules must be between 10~9
and 10"6 s,
(ii) the rate of decay of the lower level must be high, since it sets a
limit to the intracavity flux that can be utilised before a
'bottlenecking1 occurs, and
(iii) the laser discharge load should be well matched to the network
that produces the electric pulse for good electrical coupling;
this places a strong preference on laser media which can be
discharge-pumped in a stable, non-arcing manner.
These criteria have led workers to look for systems in which the lower
level is rapidly depopulated. This can occur in a number of ways, including
collisional quenching, pre-dissociation and dissociation. This last is the
fastest, and characterises excimer lasers where the lower level disappears in
] 0a time of the order of 10 s. Thus large cavity fluxes can be used for
efficient extraction of the energy from the upper laser level and lead to the
maximum efficiency obtainable within the limits of kinetic branching ratios
and the quantum efficiency of the pathway (s) used to formulate the upper
level.
The discussion that follows is devoted primarily to excimer lasers formed
from rare-gas halides. Excimer lasers based on halogen molecules (I? laser at
342 nm, Br^ at 292 nm and C12 near 260 nm) form a small complementary set and
are of little use for the LIS of UFg since their characteristic wavelengths
lie in the vacuum UV. An understanding of the chemistry that leads to excited
rare-gas halides or halogen molecules requires a knowledge of competing
kinetic paths, as well as the mechanism whereby each path varies with mode of
excitation, pressure and gas composition. This aspect is not treated (in
depth) in this report which is concerned largely with the results attained for
given laser systems rather than the detailed mechanisms of them.
5.2 Developments in Rare-gas Halide Lasers
Most of the early systems, including the first excimer laser of Ewing and
Brau [1975a], were pumped by relativistic E-beams. These lasers were
18
necessarily high pressure devices in which excitation was caused by three-body
collisions. However, the complexity of such lasers hindered their development
and application, so that there has been increasing interest in the development
of electric discharge devices. These are of two basic types:
(i) E-beam controlled discharges, and
(ii) avalanche mode discharges.
E-beam controlled discharges, which use low intensity electron beams to ionise
the gas volumetrically.stabilise it and thus prevent it from arcing, have long
pulse lengths and high energy outputs per pulse. Avalanche devices appear to
offer the fastest route to medium powers, short pulse lengths (few tens of ns)
and high pulse repetition rate (kHz).
The KrF laser has a sharp emission band bear 250 nm. It is one of the
most efficient excimer lasers and has received the greatest attention in
theoretical studies based on alkali halide models. It has been developed
mainly in the E-beam pumped configuration and, to a lesser extent, in the
discharge pumped configuration. A large KrF laser has been ordered by Los
Alamos, reportedly for its two-photon UF^ separation program (Laser Focus,
September 1978, p.979).
While the fluorescence efficiency of Ar/Kr/F^ mixtures is high, their
laser efficiency is even higher [Bhaumik et al. 1977]. Using a high
intensity relativistic electron beam, Ewing and Brau [1975b] obtained an
efficiency of 15 per cent (600 mJ out for 4 J deposited in the cavity) from
0.3 per cent F2, 6 per cent Kr, and 93.7 per cent Ar, at a total pressure of
300 kPa. Nitrogen trifluoride was also used as a source of fluorine (F) but
the overall efficiency was lowered, probably because of branching which leads
to lower yields of KrF*. Longer laser pulses have been obtained in KrF lasers
with high cavity fluxes, thus demonstrating the absence of lower energy level
bottlenecks. The ultimate efficiency of relativistic, electron beam pumped,
rate-gas halide lasers is limited by the amount of energy, W, required to form
rare-gas ions and metastables. Thus the maximum efficiency of an electron
beam pumped, KrF laser in an Ar buffer is ^ ' = 24% (av.). The maximum
achievable efficiency has been estimated to be < 8 per cent and, in practice,
?vhe 'wall plug1 efficiency is 1-3 per cent, sufficient for an economic LIS
p. "icess.
19
By way of contrast, the maximum efficiency of a discharge pumped laser is
limited by the energy E* required to form the lowest energy, rare-gas
metastable species. Thus the quantum efficiency of a discharge pumped, KrF
laser is pV/K \ = 50 per cent. The maximum achievable efficiency has been
estimated to be -v 25 per cent.
An important feature of E-beam, controlled discharge, KrF lasers is the
role played by the halogen-bearing species in discharge stabilisation. The
fast reaction of F~ with Kr* enhances discharge stability further by keeping
the metastable density relatively low. No such stabilisation occurs in
discharge pumped, Xe2 lasers. E-beam controlled discharge lasers have an
additional advantage over pure E-beam pumped systems in that the bulk of their
energy comes from the discharge rather than through a thin foil.
Early transverse discharge (avalanche mode) KrF lasers were low in both
efficiency and power, but current devices are increasingly attractive. A
feature of these lasers is the fact that the formation of the excited
molecules is not dominated by three-body collisions as in E-beam pumped
devices but by efficient two-body harpooning reactions, i.e. reactions which
occur by way of an ionic-covalent curve crossing. Greene and Brau [1978] have
developed a theoretical model of UV pre-ionised, transversely discharged KrF
and ArF lasers which account for the time dependence of the discharge, the
kinetics of the atomic/molecular species and free electrons, and the growth of
laser emission.
Discharge type excimer lasers are usually pumped by a Blumlein circuit or
more often by an L-C inversion circuit similar to those used for pumping
nitrogen and C02 lasers. Although C02 lasers do not give as high a peak
output as nitrogen lasers, they are easier to construct and have larger
discharge volumes: they also require auxiliary pre-ionisation to obtain a
uniform discharge. One important advantage of discharge pumped lasers is that
they may enable high average powers at high repetition rates to be obtained at
lower circuit costs than do pure E-beam devices.
Table 3 presents data on typical, rare-gas halide lasers over this period
of development (1976-1980) from which it can be seen that overall efficiencies
are of the order of 1 per cent and thus much less than predicted achievable
values. Also, there is an inverse relationship between output energy and
repetition rate, and average powers have risen very slowly from less than 1
watt to a few tens of watts. The KrF laser of Fahlen [1980] decreased to -\. 60
20
per cent of its initial output after a few hours running. Clearly,
development of these lasers is difficult and many problems have still to be
overcome before they can operate reliably at the high powers and repetition
rates required for commercial separation.
5.3 Developments in Halogen Lasers
E-beam excitation of argon/I^, Br or C12 mixtures, whose emission spectra
are similar to those of the rare-gas halides, has been demonstrated by Chen
and Payne [1976] and McCusker et al. [1976]. Although the lower laser level
in Ar/U or Br^ mixtures is bound, rapid depopulation takes place by
collisional ly induced pre-dissociation. For Ar/CL mixtures, direct
depopulation through dissociation is also possible. Lasing has been induced
also in mixtures of Ar/Xe/CF3I but is better in Ar/HI and Kr/HI [Ewing and
Brau 1975b]. Peak outputs are of the order of a few MW at efficiencies of -v 2
per cent. Br^ and I~ excimer lasers are similar but their power outputs are
not high, probably because of self-absorption and bottle-necking. Woodworth
and Rice [1978] recently described an F? laser operating at 157 nm with a peak
power of 7 MW, an intrinsic efficiency of 3.5 per cent, and using a high
pressure mixture of He and a few hundreds of pascals of F?. All these devices
operate at low repetition rates (< 10 Hz). However, a fast discharge F2 laser
with similar characteristics to the above E-beam devices and said to be
capable of operating at high repetition rates has been described by Hohla and
Pummer [1979].
Overall, halogen-type excimer lasers have lower outputs, narrower
bandwidths, restricted wavelengths and lower repetition rates than rare-gas
halide lasers, and they suffer from corrosion problems. They do not appear to
be as significant for the LIS of UFg as do the rare-gas halide lasers.
5.4 Tuning UV Radiation to UFf Transitions^ ?b
Rare-gas halide lasers, halogen lasers and the comparatively recent mixed
halogen lasers [Diegelmann et al. 1979] all use diatomic species for their
active medium and are line-tunable only. Apart from fortunate coincidences,
they cannot be exactly tuned to UFg transitions (absorption lines). They can,
however, be used as pumps for continuously tunable dye lasers. Dyes have been
'eveloped which, when pumped with a KrF laser, will lase down to 311 nm [Zapke
e: al. 1980] v/hile an XeCl laser has been used to pump dyes covering the range
fro 340 to 710 nm [Uchino et al. 1979] with a maximum efficiency of 40 per
21
cent in the visible. However, the stability of such dyes under high
repetition rate, high power operating conditions must be taken into account.
Dye stability is a problem in the atomic LIS process (which uses visible
radiation) where it has been overcome by the use of expensive recirculation,
purification and dye make-up equipment. It is likely to be a much more
serious problem at the shorter wavelength being considered for potential UF
processes. A further problem with dye lasers is super-radiance at high powers
leading to loss of coherence.
Wavelength shifting can be achieved also by stimulated Raman scattering.
Djeu and Burnham [1977] observed intense electronic Raman emission from Ba
vapour pumped by an XeF laser, the optical conversion efficiency being greater
than 90 per cent. Loree et al. [1977] observed a conversion efficiency of ?>.
50 per cent in the first Stokes shift of H-, D and CH, when pumped by ArF and
KrF. Both of these are examples of line shifting; however, in Loree's study,
the first anti-Stokes line was also generated at an exceptionally high
efficiency (approximately half the first Stokes line). This result can be
linked with the recent work of Bischel et al. [1979] who used efficient
photolytic pumping methods to overcome E-beam absorption problems in KrF, and
generated 483 nm radiation with a tunable bandwidth of 40 nm. This raises the
possibility that exacl matching to a UF transition can be achieved by
continuous tuning of the pumping laser (without the use of dyes) and up-
converting its output at high efficiency (?*. 25 per cent) up to the required
value. High power pumps are required for the generation of high power Raman
emission, 200 J from XeF and 300 J from KrF [Avouris 1980]
Finally, as noted above, diatomic excimer lasers are not continuously
tunable; however, triatomic excimer lasers are. The first such laser was
described by Tittel [1980a] who obtained lasing from Xe^Cl centred at 518 nm:
he predicted that it could be tuned across its entire 450-550 nm fluorescent
band. This has been followed by the further announcement [Tittel 1980b] that
lasing emission has been obtained from Kr^F at 436 nm, at a bandwidth of 25
nm, and with a probable tuning range of 380-480 nm. Since the techniques
developed for rare-gas halide lasers should be transferable to this new class
of laser, it is likely that they may provide a single continuously tunable
source of UV radiation covering the wavelength of interest for the LIS of UF .
It remains to be seen whether they can be developed to the necessary levels of
power, repetition rate and reliability.
22
6. LASER IRRADIATION OF UFCo
6.1 Matrix Isolated UFg0
Matrix isolation is a well known technique for studying photochemical
reactions in an inert or reactive solid matrix which is cooled to liquid
helium temperatures and in which molecules can undergo reactions uncomplicated
by thermal effects. Given the spectral simplification that accompanies
cooling to such temperatures, it is not surprising that this technique has
attracted attention for use in laser induced dissociation/isotopic separation
reactions.
Ambartzumian et al. [1976] appear to have been the first to attempt such
studies with infrared radiation. They used pulsed CCL-TEA lasers, at powers
of 5-30 MW cnf2, and claimed isotopic enrichment in an SFg:Ar mixture (1:500
to 1:2000) which was held on a Csl substrate mounted on an He cryostat. Their
results have, however, been questioned by Davies et al. [1978] who were unable
to confirm them. Bernstein [1977] has reported the beginnings of a program of
uranium isotope separation based on the use of U(BH^)^, but has so far
published no irradiation results.
An unexpected development in this area is the observation that matrix
isolated materials can be photochemically affected at infrared laser powerp
densities of only a few mW cm . At the 1978 Atlanta Conference, Livermore
workers disclosed that they had irradiated thev_ band of UFfi in a matrix of
SiH^ at 10 pm, and that new absorption bands had appeared which they
attributed to UF& or UF4 [Catalono et al. 1978]. These results have been
questioned by a Los Alamos worker, P. Robinson [private communication 1979],
who interpreted them in terms of laser heating. However, the work of
Poliakoff et al. [1978] on Fe(CO)4 in a CH4 matrix seems to demonstrate
conclusively that isotopic selectivity is possible, and that reactions can bep
induced by low power lasers (< 5 mW cm" ) provided there is good coincidence
between laser output and the reactant absorption band.
The Livermore work does not lend itself very well to enrichment since the
vg band was too wide (t 16 cm" ) for adequate spectral resolution. However,
since the absorption band of the molecule depends on the matrix material,
suitable UFg-matrix combinations may possibly be found, giving bandwidths < 5
cm" , and significant isotcpe selectivity could be achieved by using side band
excitation. Nonetheless, while these studies on solid UF,. reactions areD
23
scientifically interesting, it is difficult to see how the process could be
scaled up to commercial requirements.
6.2 Gas Phase UF?
6.2.1 Single infrared photon excitation
Gurs et al. [1976] advanced the idea that isotope selective excitation of
UFg, with a single vibrational quantum, could lead to isotope selective
photochemistry. Eerkens [1976] claimed to have achieved a separation factor
of 1.1 by this means but his claims have not been substantiated. Gurs1 idea
was based on the use of the CO laser and the 3 v3 band of UF,, and Eerkens1 on
the CO laser and the v~ + v. + Vg combination band. (Undoubtedly similar
claims will be made for the \;3 band and the 16 ym laser.) In both cases, the
isotope separation is to be achieved by selective reaction of laser excited
UFg with some co-reactant.
Gurs1 and Eerkens1 claims assumed that differences in the reaction rates
could be described by
k? - exp(-E/kT)
k*? exp(-E-hv/kT)
where k*, k? are the rate for excited and unexcited UF E is the bimolecular
activation energy _and, v, the photon energy. The ratio of the reaction rates
for the U and U isotopic species is therefore
k*/k? = exp(hv/kT)
Unfortunately, while the ratio can be quite significant at or below room
temperature, even for 16 yrn photons, the theoretical basis of this approach is
wrong. During thermal activation to the transition state at E-hv, the initial
laser discrimination will always be lost (since deactivation is always more
efficient than activation) and subsequent reactions will be isotopically non-
selective. Alternatively, the quantum yield will be prohibitively small. The
special case which occurs when E is comparable to h , is also difficult to
realise in practice. Generally, laser enrichment is practical only when the
energy of excitation is comparable to the reaction activation energy and, for
vibration excitation with an infrared laser, only multiphoton excitation can
24
achieve this condition.
6.2.2 Single UV/VIS photon excitation
The intense (B-X) absorption band centred at -v 260 nm (see Figure 5)
would be ideal for isotope separation purposes if the spectrum in this region
could be isotopically resolved under adiabatic expansion conditions. However,
although no such experimental investigation of this region has been published,
the work of Kroger et al. [1978] makes it clear that no isotopically resolved
spectrum can exist in this absorption band. Using a polarised, 210 nm laser
beam, they observed that the F atoms in the photo-dissociated, photo-fragment
spectrum were asymmetrically distributed in that plane which is at right,
angles to the laser beam; it appears that the photo-activated UFg was
dissociating within this rotational period of the UFg molecule. This
observation eliminates the possibility that rotational, and therefore
isotopic, structure will occur in this spectral region. The observation by
Andreoni et al. [1978] that no fluorescence occurs from UFg irradiated at 270
nm further corroborates this conclusion.
A different situation prevails in the relatively low intensity (A-X)
absorption band covering the spectral region 340 to 420 nm. As discussed in
Section 2.2, there is strong evidence to suggest that under appropriate
conditions, probably adiabatic cooling, isotopically resolved spectra will be
found in this region.
The fluorescence measurements of Andreoni et al. [1977] suggest that the
crossover point for intersystem crossing from a dissociative electronic state
which correlates directly with ground state UFg and an F atom, lies at about
27 900 cm"1. Horsley et al. [1980] have irradiated UFg at wavelengths above
and below this energy level. Irradiation of pure UFg and 1:5, UFg/N2 mixtures
at 50 to 100C at argon ion wavelengths of 363.8, 35.1 and 350.8 nm gave
quantum yields of 0.5 to 1, depending on the temperature. (The quantum yield
was here defined as the number of molecules removed from the system per
photon.) Using both infrared absorption and laser snow detection methods,
they further showed that the rate of depletion of UFg in a static cell was of
the order of 2 x 10"1 Pa s"1 IT1. This rate increased to 2 Pa s"1 W"1
(corresponding to quantum yield of 2) when hydrogen was introduced to
scavenge the F atoms produced in the dissociation. Horsley et al. considered
the reaction sequence to be
25
UFg + hv -?- UF5 + F
F + H2 -?? HF+ + H
H + UFC ?* UFr + Ho b
The possibility that other reaction pathways are feasible, e.g. a
bimolecular reaction of H? with electronically or vibrationally excited UF,.,
indicates the need for careful control of irradiation wavelength and power in
order to avoid loss of selectivity by scrambling reactions. The conclusion
reached by Horsley et al., that at most only l/10th of the absorbing molecules
will undergo photo-deprivation, is probably the result of a calibration error.
Although Br2 (used in the calibration procedure) has an unbound upper state
and can be totally dissociated, UFg with bound upper and lower states cannot
be totally dissociated because these levels cannot be repopulated in a
molecular beam.
A further method by which UFg can be separated is by enhanced reaction of
the selected, excited isotopic species with other atomic species. Horsley's
group also considered this probability from a theoretical standpoint and
concluded that Br and I, but not H, atoms could be useful for selective
reaction with vibrationally excited species. However, this process appears to
be of no practical consequence since, in addition to the provision of
activated UFg, it is also necessary to generate Br (or I) atoms in situ and
these, as well as reacting as above, can also collisionally deactivate
vibrationally excited molecules through V-T processes.
6.2.3 Simultaneous infrared and UV/VIS photon excitation
This approach, first suggested by Robieux and Auclair [1965] and analysed
by Struve [1972], achieves isotope selectivity by single quantum vibrational
excitation of one species only, while a single UV/VIS quantum, selectively
absorbed by the vibrationally excited species, provides the energy necessary
to initiate photochemical reactions. This approach, in conjunction with
adiabatic cooling, was used at LASL (Laser Focus, 1976) to achieve some degree
of uranium enrichment. The research program concerned with scientific
feasibility has now been successfully completed, and the method is being
developed at Oak Ridge under the sponsorship of Department of Energy (DOE)
[Haberman 1980]. Cost estimates [Robinson et al. 1976] compared very
favourably with those of centrifuge enrichment (Table 4).
26
The main difficulty, but one in which continuous progress is being made
[Mace 1981], is the provision of high powered, high repetition rate laser
radiation sources which will operate for sufficiently long periods between
overhaul/replacement. Birely et al. [1976] estimate laser fluence
requirements of x 10 kW/cm2 and -^ 2 kW/cnr for the infrared and UV/VIS lasers
respectively. Also, the infrared laser wavelength must be within ? 0.02 cm"1
of the v- absorption peak of UFfi, and this additional requirement may
eliminate many 16 ym sources otherwise suitable for room temperature
irradiation of UF...o
6.2.4 Multi-photon infrared excitation
Isotope selective infrared laser induced multiphoton dissociation of SFg
was reported in 1975 [Ambartzumian et al. 1975; Lyman et al. 1975] usingp
pulsed C02 lasers in the energy fluence range of 1 to 100 J cm . Since then,
infrared multiphoton induced dissociation has been observed in many molecules,
such as OsO^, MoF^ and Nf^. However, isotope selectivity, when measured, was
found to be small for the heavy elements. Although the separation of the3? "34
S, S isotopes was practically 100 per cent in SFg where the isotopic
absorption bands are fully resolved, the enrichment fell markedly to a few per
cent for the relatively light SeFg [Tiee and Whittig 1978c] and MoFg [Freund
and Lyman 1978] molecules where the isotopic spectral shift is small (1.6 cm
for the vg vibration in SeFg) and considerably less than the absorption band
profile width.
In UFg and similar molecules, the anharmonicity problem associated with
vn -?? vn excitation to the quasi-continuum is overcome by using high powered
lasers to induce power broadening. For UFfi, the isotopic spectral shift is ?*.
? 1 ?0.2 c,n /amu, and the energy fluence necessary for dissociation (-v 1 J cm)
is comparable to that for SFg. Since fluences of this magnitude introduce
power broadening of the order of the isotopic shift, there is little value in
adiabatically cooling UFg to provide the resolution necessary for isotopic
selectivity.
To overcome this loss of selectivity caused by power broadening, it has
been suggested that isotope selectivity could be maintained by infrared
irradiation at power fluences where only a few photons (1-3) are absorbed and
the spectrum of the infrared excited UFg is shifted to a lower frequency where
the unexcited isotopic species becomes transparent. The excited species is
then irradiated by an infrared laser of different frequency to the
27
dissociation limit. This process was first described by Ambartzumian et al.
[1978] for the separation of the osmium isotopes in OsO^.
Such a process has been patented for UFg by Kaldor and Rabinowitz [1976]
who used a 16 ym laser to activate the molecules and a C00 laser to dissociate
them; no technical details, in particular separation factors,were disclosed.
Tiee and Whittig [1978d] described a similar system in which a CF^ laser was
used to excite UFg so that the 8.6 ym transition of unexcited UFg was
broadened and shifted to allow excited UFg molecules to absorb 9.6 pm C0?
laser radiation. Enhancement factors of 1 to 100 in the dissociation were
obtained but no mention was made of enrichment. Alimpiev et al. [1979] have
reported the dissociation of UFg at room temperature by both multiphoton CF.
laser radiation and the combined action of CF. and C0? lasers. As expected
under these conditions, a mass spectrometric investigation of the residual UF
gas failed to show any isotopic selectivity.
7. CONCLUSIONS
The only UFg-based process confirmed to give some degree of enrichment is
based on sequential infrared plus VIS photon excitation. Processes based on
single or multiphoton infrared irradiation with 16 ym or CO^ lasers are
unlikely to achieve significant separation, although multiphoton excitation
will readily cause photo-dissociation. The demonstration of UV/VIS
fluorescence in UFg indicates the presence of an excited state sufficiently
long-lived for LIS purposes. However, until the adiabatically cooled spectrum
of UFg in the region of 400 nm is known, no firm conclusions can be drawn as
to the feasibility of a separation process based on a single visible photon.
Adequate spectral resolution for either the infrared plus VIS process or
the conceptual VIS process requires the co-expansion of UFfi with a monatomic
or diatomic gas; for all practical purposes it cannot be attained by
expansion of UFg alone. Co-expansion and the correct use of shock waves in
the expanding gas have the beneficial side effect that the process can be
carried out at relatively high pressures where pumping costs are much lower
and the gases can be protected from background impurity contamination. High
supersaturation can be achieved in this process with dT > 45 K and probably as
high as 100 K. The engineering of an adiabatic expansion system is a major
exercise, with the literature suggesting a development period of two to three
years for the Smalley system.
28
Although the CF. laser is likely to accomplish the infrared activation of
UFg no one yet has fine tuned it to the appropriate UFg lines. Recent
developments indicate that the Raman shifted CO^ laser is a more likely source
of tunable, high powered 16 ym radiation.
Excimer lasers based on the rare-gas halides are the most promising
UV/VIS lasers. After three years of development they are only now approaching
the minimum requirements with respect to power, efficiency and particularly
repetition rate. The attainment of an adequate service life for these devices
has posed problems which are very difficult but appear, nevertheless, to be
capable of solution. This area is clearly one calling for significant and
innovative ideas.
We lack practical data on
. the problems of constructing and operating a complex system which
utilises adiabatic expansion, high resolution infrared lasers and
high powered UV/VIS lasers;
. the photochemical efficiency; and
. the collection efficiency.
Hence, at this stage, we cannot make a critical assessment of the existing
cost estimates of a UFg based LIS process.
Major overseas laboratories have devoted a great deal of time and effort
to the problem of evolving a commercial LIS process based on UFfi. Reliable
estimates of the cost of enrichment by such a process will become available
only when a pilot plant based on such a process is constructed and operated,
and results made available.
8. REFERENCES
Alimpiev, 0., Babichev, A.P., Baranov, G.S., Karlov, N.V., Karcherskii, A.I.,
Kulikov, S.Y., Martsynkyan, V.L., NaBiev, S.S., Nikiforor, S.M.,
Prokhorov, A.M., Sartakov, E.G., Skvortsova, E.P. and Khokhlov, E.M.
[1979] - Sov. 0. Quant. Electron., 9:1263.
29
Ambartzumian, R.V., Govokhov, Y.A., Letokhov, V.S. and Makarov, G.N. [1975] -
JETP Lett., 21:171.
Ambartzumian, R.V., Govokhov, Y.A., Makarov, G.N., Puretsky, A.A. and
Furzikov, N.P. [1976] - JETP Lett., 24:257.
Ambartzumian, R.V., Letokhov, V.S., Makarov, G.N. and Puretsky, A.A. [1978] -
Opt. Comm., 25:69.
Andreoni, A., Cubeddu, R., DeSilvestri, S., and Zaraga, F. [1977] - Chem.
Phys. Lett., 48:431. [1978] - Optical Sciences, Vol.9, High Power
Lasers and Applications (Komfra, K.L. and Walther, H. Eds).
Andrecu, D. [1977] - Opt. Comm., 23:37.
Baranov, U., Vasilev, B.I., Velikhov, E.P., Gorokhov, Y.A., Grasyuk, A.Z.,
Dyad'kin, A.P., Kazakov, S.A., Letokhov, V.S., Pismennyi, V.D. and
Starodubtsev, A.I. [1978] - Sov. J. Quant. Electron., 8:544.
Barch, W.E. [1975] - Opt. Comm., 15:358.
Bans, P. [1980] - J. Phys. Chem., 54:1797.
Bar-Ziv, E., Freiberg, M. and Weiss, S. [1972] - Spectrochim. Acta, 28A:2025.
Basov, N.F., Belenov, E.M., Isakov, V.A., Leonov, Y.S., Markin, E.P.,
Oraevskii, A.N., Romanenko, V.I. and Ferapontov, N.B. [1975] - JETP
Lett., 22:102.
Benetti, P., Cubedda, R., Sacchi, C.A., Svetto, 0. and Zaraga, F. [1976] -
Chem. Phys. Lett., 40:240.
Bernstein, S. [1977] - US Department of Energy Report 100-41181-1.
Birely,J.A., Cartwright, D.C. and Marinuzzi, J.G. [1976] - Proc. Soc. Photo
Instrum. Eng., 76:124.
Bhaumik, M.L., Bradford, R.S. and Ault, E.R. [1977] - Appl. Phys. Lett.,
28:23.
30
Bischel, U.K., Kligler, D.J., Pummer, H. and Rhodes, C.K. [1978] - J. Chem.
Phys., 69:4652.
Bischel, W.K., Kligler, D.J. and Rhodes, C.K. [1979] - CLEA Conf. on Laser
Engineering and Applications, 23-28 August, Washington, paper 9.6.
Buchwald, M.I., Jones, C.R., Fettermen, H.R. and Schlossberg, H.R. [1976] -
Appl. Phys. Lett., 29:300.
Burnham, R. [1979] - CLEA Conf. on Laser Engineering and Applications, 23-28
August, Washington, paper 9.5.
Burnham, R. and Djeu, N. [1976] - Appl. Phys. Lett., 29:707.
Byer, R.L. [1976] - IEEE, J. Quant. Electron., 12:732.
Byer, R.L., Herbst, R.L. and Fleming, R.N. [1974] - Appl. Phys. Lett.,
25:520.
Catalano, E., Barletta, R. and Pearsin, R. [1978] - IQEA Conf. Atlanta, paper
V9.
Chang, T.Y. and Magee, J.D. [1976] - Appl. Phys. Lett., 28:526.
Chen, C.H. and Payne, M.6. [1976] - Appl. Phys. Lett., 28:219.
Christensen, C.P. [1977] - Appl. Phys., Lett., 30:483.
Davies, B., Poliakoff, M., Smith, K.P. and Turner, J.J. [1978] - Chem. Phys.
Lett., 58:23.
Djeu, N. and Burnham, R. [1977] - Appl. Phys. Lett., 30:473.
DePoorter, G.L. and Rofor-De-Poorter, C.K. [1978] - Spect. Lett., 8:521.
Diegelmann, M., Hohla, K. and Kompa, K.L. [1979] - Opt. Comm., 29:334.
Eerkens, J.W. [1976] - Appl. Phys., 10:15.
31
Ewubgm J.J. and Brau, C.A. [1975a] - Phys. Rev., A12:129.
Ewing, J.J. and Brau, C.A. [1975b] - Appl. Phys. Lett., 27:350.
Fahlen* T.S. [1978] - J. Appl. Phys., 49:455.
Fahlen, T.S. [1980] - Laser Focus, June, p.34.
Feigelson, R.S., Kway, W.L. and Byer, R. [1980] - Laser Focus, October, p.22.
Fischer, R. [1980] - Laser Focus, July, p.77.
Freund, S.M. and Lyman, J.L. [1978] - Chem. Phys. Lett., 55:435.
Frey, R., Pradere, F., Lukasik, J. and Ducuing, J. [1977] - Opt. Comm.,
22:355.
Giordmaine, J.A. and Miller, F. [1965] - Phys. Rev. Lett., 14:973.
Green, F.T. and Milne, T.A. [1970] - Phys. Abs., 73:70089.
Greene, A.E. and Brau, C.A. [1978] - IEEE J. Quant. Electron., 14:351.
Grzybowski, J.M. and Andrews, L. [1978] - J. Chem. Phys., 68:4540.
Gurs, K. [1976] - Br. Nucl. Energ. Soc. Conf., 15-18 March, London.
Hagena, S. and Henkes, W. [1960] - Naturforschung, 15:851.
Hinkley, E.D., Harman, T.C. and Freed, C. [1978] - <A.ppl. Phys. Lett., 18:49.
Hohla, K. and Pummer, H. [1979] - CLEA Conf. on Laser Engineering and
Applications, 23-28 August, Washington, paper 9.9.
Horsley, J.A., Rabinowitz, P., Stein, A., Cox, D.M., Brickman, R.O. and
Kaldor, A. [1980] - IEEE J. Quant. Electron., 16:412.
Hughes, W.M., Shannon, J. and Hunter, R. [1974] - Appl. Phys. Lett., 24:488.
32
Hurst, H.J. and Wilson, P.M. [1972] - Spectra Lett., 5:275.
Hyer, R.C., Campillo, R. and Shapiro, A. [1979] - Laser Focus, August, p.22.
Jacobs, R.R., Prosnitz, D., Bischel, W.K. and Rhodes, C.K. [1976] - Appl.
Phys. Lett., 29:710.
Janner, K. [1976] - German Patent No.265950, Filed 1976, Disclosed 1978.
Jensen, R.J., Marinuzzi, J.G., Robinson, C.P. and Rockwood, S.L. [1976] -
Laser Focus, May, p.51.
Jetter, H. [1975] - AED-CONF-75-531-001.
Jones, C.R. and Buchwald, M.I. [1978] - 10th Int. Quantum Electronics Conf.,
29 May - 1 June, Atlanta, paper K7.
Kaldor, A. and Rabinowitz, P. [1976] - Belgium Patent No.846225.
Kim. K.C. and Person, W.B. [1978a] - J. Chem. Phys., 69:1963.
Kim, K.C. and Person, W.B. [1978b] - Chem. Phys. Lett., 54:253.
Kim, K.C., Holland, R.F. and Filip, H. [1978c] - Appl. Spectros., 32:287.
Konev, Y.U. and Lipatov, N.I. [1977] - Sov. J. Quant. Electron., 7:1017.
Kroger, M., Riley, S.J. and Kivei, G.A. [1978] - J. Chem. Phys., 68:4195.
Kudryartsev, Y.A. and Kuzmina, N.P. [1977] - Appl. Phys., 13:107.
Kung, R.T.V. and Itzkhan, I. [1976] - Appl. Phys. Lett., 29:780.
Kuntsevich, B.F., Stepanov, B.I., Trushin, S.A. and Churakov, V.V. [1978] -
Sov. J. Quant. Electron., 8:313.
Kynazev, I.N., Letokhov, V.S. and Lobko, V.V. [1979] - Opt. Comm., 29:33.
Lewis, W.B., Asprey, L.B., Jones, L.A., McDowell, R.S., Rabideau, S.W. and
Zeltmann, A.H. [1976] - J. Chem. Phys., 65:2707.
33
Lipkin, D. and Weismann, S.I. [1942] - UCRL Report A-520.
Loree, T.R., Sze, R.C. and Barker, D.L. [1977] - Appl. Phys. Lett., 31:37.
Lyman, J.L., Jensen, R.J., Rink, J., Robinson, C.P. and Rockwood, S.D. [1975]
- Appl. Phys. Lett. , 27:37.
Mace, P.N. [1981] - Los Alamos Report LA-UR-81-2541.
Mail, R.A., Markovitch, F.J. and Carr, R.H. [1977] - CR-2-718.
McCusker, M.V., Hill, R.M., Huestis, D.L., Lorents, D.C. and Nakane, H.H.
[1976] - Opt. Comm., 18:119.
McKee, T.J., Stoicheff, B.P. and Wallace, S.C. [1977] - Appl. Phys. Lett.,
30:278.
Miller, J.C., Allison, S.W. and Andrews, L. [1979] - J. Chem. Phys., 70:3524.
Osgood, R.M. [1976] - Appl. Phys. Lett., 28:342.
Piltch, M.S., Rink, J. and TalIman, C. [1975] - Opt. Comm., 15:112.
Poliakoff, M., Breedon, N., Lavies, B., McNeish, A. and Turner, J.J. [1978] -
Chem. Phys. Lett., 56:274.
Pummer, H., Bischel, U.K. and Rhodes, C.K. [1978] - J. Appl. Phys., 49:976.
Rabinowitz, P., Stein, A., Brickman, R. and Kaldor, A. [1980] - Laser Focus,
January, p.20.
Rabinowitz, P., Stein, A. and Kaldor, A. [1978] - Opt. Comm., 27:381.
Robieux, J. and Auclair, J.M. [1965] - French Patent No.1391738.
Robinson, C.P., Jensen, R.J., Cottor, T.P., Boger, K. and Greiner, N.R. [1976]
- Aust. Pat. No.19855-76.
Rutt, N. [1980] - Laser Focus, July, p.78.
34
Shapiro, A.M. [1953] - Dynamics and Thermodynamics of Compressible Fluid Flow.
Ronald Press, New York.
Sheremetev, G.D. [1957] - Opt. Spectrosk., 2:99.
Smalley, R.E., Ramakrishna, B.L., Levy, D.H. and Wharton, L. [1974] - 0.
Chem. Phys., 51:4363.
Stein, A., Rabinowitz, P, and Kaldor, A. [1978] - Opt. Lett., 3:97.
Suzuki, K., Saite, S., Otam, M. and Fukioka, T. [1980] - J. Appl. Phys.,
51:4009.
Sze, R.C. [1979] - J. Appl. Phys., 50:4596.
Telle, J.M. [1978] - CLEA Conf., Washington, paper K14.
Tiee, J.J. and Wittig, C. [1978a] - J. Appl. Phys., 49:41.
Tiee, J.J. and Wittig, C. [1978b] - IEEE Conf., Atlanta, paper K10.
Tiee, J.J. and Wittig, C. [1978c] - Appl. Phys. Lett., 32:236.
Tiee, J.J. and Wittig, C. [1978d] - Opt. Comm., 27:377.
Tittel, F.K. [1980a] - Laser Focus, April, p.34.
Tittel, F.K. [1980b] - Laser Focus, September, p.14.
Uchino, 0., Kyusu, U. and Fukuoka, H. [1979] - Appl. Phys., 19:235.
Wampler, F.B., Oldenborg, R.C. and Rice, W.W. [1978a] - Chem. Phys. Lett.
54:555.
Wampler, F.B., Oldenborg, R.C. and Rice, W.W. [1978b] - Chem. Phys. Lett.,
54:557.
Wampler, F.B., Oldenborg, R.C. and Rice, W.W. [1978c] - Chem. Phys. Lett.,
54:560.
35
Wampler, F.B., Rice, W.W., Oldenborg, R.C., Akerman, M., Magnuson, D.W.5
Smith, D.F. and Werner, 6.K. [1979] - Opt. Lett., 4:143.
Wang, C.P. [1976] - Appl. Phys. Lett., 29:103.
Wang, C.P. [1978a] - Sov. J. Quant. Electron., 8:1005.
Wang, C.P. [1978b] - Appl. Phys. Lett., 32:360.
Wang, C.P. and Gibb, O.L. [1979] - IEEE J. Quantum Electron., 15:318.
Wegener, P.P. and Wu, 8.J.C. [1977] - Adv. Colloid Interface Sci., 7:325.
Wenzel, R.G. and Arnold, G.P. [1976] - Appl. Phys. Lett., 15:1322.
Wexler, B.L. and Wagnant, R.W. [1979] - Appl. Phys. Lett., 34:674.
Woodworth, J.R. and Rice, J.K. [1978] - J. Chem. Phys., 69:2500.
Wu, B.J.C., Wegener, P.P. and Stein, G.D. [1978] - J. Chem. Phys., 68:308.
Zapke, W. [1980] - Laser Focus, January, p.79.
37
TABLE 1
MAJOR ABSORPTION BANDS OF INFRARED SPECTRUM OF UF,
Assignment
Vi + Vo
V2 + V3
v3 + v4
V.
vd
Frequency Relative Absorption
(cm )* Cross Section
1291 (7.7)
1157 (8.6)
827 (12)
625 (16)
186 (54)
2.0 x 10"3
2.3 x 10"3-d
9.5 x 10 4
1
5 x 10"2
* Values in parenthesis are the corresponding wavelengths (in
TABLE 2
GAS DYNAMIC SUPERSONIC NOZZLE PARAMETERS
Mach No.
Specific
Heat
Y =
Spec
Heat
Y =
Ratio
1.67
ific
Ratio
1.40
T(K)
P0/P
n /n0
A/At
VPo
T(K)
P0/P
n0/n
A/At
VPo
2
129
1
1
8
4
?
'
5
3
167
1
1
8
4
?
?
7
4
3
75
32
8
3
2.3
107
37
13
4
3
4
47
100
16
6
4
71
154
36
11
i
5
32
263
29
10
7
50
530
91
25
16
6
23
625
38
16
11
37
1600
190
53
53
7
17
1250
71
24
17
28
4200
385
120
64
Note TQ = 285 K
38
TABLE 3
PERFORMANCE DATA ON SOME EXCIMER LASERS DEVELOPED
IN PERIOD 1976-1980
Laser
KrF
(248 nm)
XeF
(350 nm)
ArF
(193 nm)
XeCl
Output Specific
(mJ) Output
(mJ/L)
3 0.3
130 2
50 0.5
10
100
-
2000
10 1
65 1
0.4 0.1
3
10
12
60 1 '
6.8 0.2
-
9
350
Overall Repetition
Efficiency Rate
(%) (Hz)
0.3 20
1.3
1
400
1 2000
0.5 1000
1
1 20
0.65
0.16 200
500
400
2000
0.65
0.06 5
130
400
-
Average
Power
(W)
0.06
-
-
3.3
-
200
0.2
-
0.055
-
3.3
24
-
0.004
10
1.6
-
Source
Wang [1976]
Burnham and
Djeu [1976]
McKee et al. [1977]
Wang [1978a]
Fahlen [1978]
Fahlen [1980]
Lumonics Data
Sheet [1980]*
Wang [1976]
Burnham and
Djeu [1976]
Christenson [1977]
Wang [1978b]
See Ref.26 of
Wang [1978a]
Wang and Gib [1979]
Burnham and
Djeu [1976]
Kudryantsev and
Kuzmina [1977]
Lambda Physik
Data Sheet [1980]*
Wang [1978a]
Sze [1979]
*Taken from data sheets of commercial manufacturers
39
TABLE 4
COST STRUCTURE OF 9 x 106 SWU/y CAPACITY
ENRICHMENT PLANTS (1976)
Enrichment
Process
Diffusion
Centrifuge
Laser
Capital
Costs
M
3100
2040
130
Power
Required
MW
2430
243
24
Power
Costs*
M
426
43
4
Other
Costs
M
36
194
16
*Power costs 0.02/kWh
41
0.8 -
Q-cra 0.4
238U V
3-absorption
T =300 K
0.2-
0
FIGURE
610 620 630
ENERGY (cm-1)
INFRARED ABSORPTION SPECTRUM OF
AT ROOM TEMPERATURE (after Jetter 1975)
10
Q_o
Q_
o
i?
LJ
<
CH
-310
30 50 70 90 110 130 150 170
TEMPERATURE (T?K)
FIGURE 2. POPULATION OF TEN LOWEST VIBRATIONAL STATES
vs TEMPERATURE (K) (after Jensen et al. 1976)
42
0
expansion
cooled
Room
temperature\
FREQUENCY
(a)
ELJ
12
10
8
6
4
2
0
V3 SF6(inN2)
- in Supersonic
- Flow (Mach 5)
7=Hf? Resolution -
1
0.8
0.6
0.4
0.2
0
940 945 950
FREQUENCY (cm-i)
(b)
955
SF(in N)
300 K -
945 950
FREQUENCY (cm-i)
(c)
FIGURE 3a. ABSORPTION SPECTRUM OF UF 6 COOLED BY EXPANSION
THROUGH A NOZZLE
FIGURE 3b. ABSORPTION SPECTRUM OF SF? AT 55 AND 300 K
AND 3c. (after Jensen et al. 1976)
43
10
0.1
- (Q)
80 100 120 140
T(K)
FIGURE 4a. ONSET OF SFe CONDENSATION vs TEMPERATURE
10"-*
150 200
T (K)
250 300
FIGURE 4b. EFFECT OF NOZZLE DESIGN ON CONDENSATION
PRESSURES (pk) (after Wu et al. 1978)
44
10-22
2000 3000 4000
WAVELENGTH (A
FIGURE 5. UV/VIS ABSORPTION SPECTRUM OF UF6
(after De Poorter and Rofor-de-Poorter 1978)