Absorption spectrum of uranium metal vapor. The absorption spectrum of uranium metal vapor is very complex, with over 300,000 lines at visible wavelengths. However, many of these absorption lines are very sharp, with sufficient displacement between a 23aU absorption line and the MSU absorption line for the corresponding transition, and without overlap of the 23 line with the «’U line for a different transition, to permit selective excitation of the 235U atoms. However, choice of the wavelength most suitable for a practical process is made difficult by the large number of possibilities. Janes et al. [J2] discuss some of the alternatives.

History. In the United States, laser isotope separation (LIS) with uranium metal vapor has been investigated experimentally by the Lawrence Livermore Laboratory (LLL) of the U. S. DOE at Livermore, California, and by Jersey-Nuclear-Avco Isotopes, Inc. (JNAI), a joint venture of Exxon Nuclear Company of Bellevue, Washington, and Avco-Everett (Massachusetts) Research Laboratory, which holds a number of patents on this method, of which the most significant are those of Levy and Janes [Jl, L2].

Workers at LLL [T3, D2] have reported production of milligram quantities of uranium enriched to 2.5 percent 235U by this method. In the LLL work, the source of uranium metal vapor was a uranium-rhenium alloy, chosen to reduce attack by the hot metal on the containing crucible. Deflection of 235U ions was by an electric field.

In the JNAI work, solid uranium metal is vaporized by an electron beam impinging on its surface, and deflection of 23SU ions is by either space charge expansion, a magnetic field, an electric field, or a combination. The JNAI process, as described in patents [Jl, L2] and a 1977 article [J2], has evolved through several stages. The next section describing the uranium metal LIS process follows the 1977 article.

Process description. Figure 14.41 is a schematic assembly drawing of one module of the JNAI uranium metal laser isotope separator. Figure 14.42 is a transverse section of this module. In Fig. 14.41, separation takes place inside a vacuum chamber about 1 m long. The uranium vapor source at the bottom consists of a charge of uranium metal, held in a water-cooled crucible, whose top surface is heated to 3000 К by a sheet of high-energy electrons curved and focused in a line on the uranium by a magnetic field of 100 to 200 gauss. Uranium vapor atoms diverge radially upward from the heated line source and travel in straight lines because of the absence of collisions in the high vacuum. These atoms flow upward between longitudinal, cooled,

product collector plates, oriented so that the atoms move parallel to them and do not impinge. The space between the plates is illuminated by light from a system of lasers, to be described later, which ionize most of the 235U atoms selectively, while leaving most of the 238U atoms un-ionized. The 235U ions, being electrically charged, can be deflected from the outward flowing uranium vapor and caused to impinge on and adhere to the product collector plates. Three possible methods for deflecting the 23SU ions are (1) expansion with energetic electrons released when the uranium is ionized, (2) motion in circular orbits around longitudinal magnetic field lines, or (3) deflecting by electric fields produced by giving adjacent collector plates alternative positive and negative charges. Un-ionized 238U atoms move outward beyond the product collector plates and condense on the upper cooled tails collection surface.

For maximum capacity, the uranium vapor density should be as high as possible. An upper limit is around 1013 atoms/cm3, because at higher density collisions between 235U ions and 238U atoms, or charge exchange between them, would occur too frequently, resulting in too high deflection of 238U. Assuming a plate height of around 4 cm, a flow area 4 cm wide by 100 cm long, and a uranium vapor thermal velocity of 40,000 cm/s, the uranium feed rate per module would be

— °-ш 8 a«*>

which represents a maximum daily 235U production rate of

(0.063 g uranium/sX0.00711 g 235U/g uranium)(86,400 s/day) = 39 g 235U/day (14.356) per module 1 m long.

The lasers used to ionize the 235U should be pulsed sufficiently often to irradiate all 23SU atoms passing between the plates. With a plate height of 4 cm and a vapor velocity of 40,000
cm/s, this requires a pulse repetition rate of 10,000 Hz. This, and other requirements to be described below, require development of lasers more advanced than any now available.

The light path through the module is limited to around 1 m to prevent the uranium metal vapor from itself becoming a laser, with consequent loss of selectivity. This length limitation prevents full utilization of laser photons in a single module and makes desirable connecting several physically separate modules optically in series as suggested by the second chamber shown in Fig. 14.41.

Laser requirements. In order to utilize photons efficiently, absorption by M5U should be selective. A 235U absorption line should be found that (1) occurs at a frequency at which 238U does not absorb, and (2) has a high absorption cross section, to reduce the light path needed for efficient use of photons. Because the isotope shift between 23SU and 238U absorption frequencies is of the order of 1 in 50,000, the first requirement calls for use of a very narrow 235U absorption line. Because the absorption lines for transitions in which uranium is ionized are very broad, it is necessary to ionize the 235U atoms in two or more steps, in which the first step is selective excitation of 235U to an energy level below the uranium ionization potential of

6.18 eV. This would be followed by less selective absorption of one or more additional photons of sufficient energy to ionize the excited 23SU atoms but of too little energy to ionize the unexcited 238U atoms. One of the JNAI patents [L2] suggests use of a narrow-frequency laser supplying visible light at 502.74 nm to excite 23SU, followed by ultraviolet light at 262.5 nm to carry the excited atoms over the 6.18 eV ionization level. At 502.74 nm, the 235U absorption line, of half-width around 0.001 nm, is displaced 0.01 nm from the 23®U absorption line, so that the required selectivity is obtained.

The energy E imparted to the 23SU atom by absorption of a photon of wavelength X is evaluated from Planck’s law,

(14.357)

h is Planck’s constant, 6.62559 X 10"M (J • s). c is the velocity of light, 2.997925 X 108 m/s. The energy in electron volts V is

where e is the electron charge, 1.60210 X 10 19 C, so that

he (6.62559 X Q-UX2.99192S X 108) _ 1,23981 X 10~6 eX ~ 1.60210 X КГ19 X (m) “ X (m)

Hence the energy given the 23SU atom by successive absorption of photons of wavelength 502.74 and 262.5 nm is

X (nm) E (eV)

502.74 2.466

262.5 4.723

7.189

Since the total 7.189 eV absorbed by 235U exceeds its ionization potential of 6.18 eV, this two-step photon absorption process imparts sufficient energy to ionize 235U. But since 238U absorbs only the 262.5-nm photon, 238U receives only 4.723 eV and is not ionized. Other possible combinations of two or more photons are described by Janes et al. [J2].

Even though the foregoing photon absorption process selectively ionizes 235U, charge exchange between 23SU ions and neutral 238U atoms and atomic collisions deflect enough 238U atoms to the collector plates to limit the heads enrichment factor to around 10. For example, a product content of 6 percent 23SU is the highest value that has been obtained from natural uranium. At the same time, however, very complete stripping of 23SU from tails is claimed. Three advantages cited for this kind of separation performance are as follows:

1. A single stage of separation suffices to produce uranium of high enough enrichment for light-water reactors.

2. More complete stripping is achieved than is economical in gaseous diffusion or the gas centrifuge.

3. This LIS process can be used to produce uranium containing 2 to 3 percent 235U from tails from these other processes.

The lasers for the process just described have three requirements more exacting than any yet developed:

1. They must deliver pulses with a frequency of 10,000 Hz.

2. To be economical they must last for a year or more to deliver over З X 10n pulses before replacement.

3. They must deliver far more energy per pulse than any high-frequency lasers now available.

Development problems. Despite the promise apparent in this laser enrichment process, it has a number of development problems. As just stated, lasers with higher repetition rate, longer life, and higher energy must be developed. Optical windows that do not lose transparency or mechanical integrity from deposition of uranium or intense illumination must be developed. Materials problems associated with handling corrosive uranium metal at high temperatures must be solved. Perhaps most important of all, convenient means must be developed for charging uranium to the high-vacuum, high-temperature system and for collecting and removing the separated uranium product and tails fractions. This LIS process appears to have one of the
disadvantages of the Y-I2 electromagnetic process, of having feed material deposit all over the vacuum chamber, necessitating troublesome interruptions for disassembly and clean-out.

Economic estimates. Despite these problems, JNAI was sufficiently optimistic about the ultimate economics to go ahead with pilot-plant construction. At this stage of development, however, economics are very uncertain. This is illustrated by Table 14.27, which compares estimates of process characteristics and costs made by JNAI and a Japanese group. The specific power estimate of Janes et al. is around that predicted for the gas centrifuge. The estimate of Ozaki et al., although 10 times higher, is lower than that of gaseous diffusion (Sec. 4.7). The unit investment costs predicted by both groups, although very different, are much lower than for gaseous diffusion or the gas centrifuge and are the principal reason for the interest being shown in this process.

Two features that make separative work cost estimates very uncertain are uncertainty about laser energy efficiencies and ignorance of operating and maintenance costs, which can be obtained only by completing the development and making life tests on plant equipment.