ISSN 1066-3622, Radiochemistry, 2008, Vol. 50, No. 6, pp. 650–654. © Pleiades Publishing, Inc., 2008. Original Russian Text © S.A. Bogatov, A.A. Borovoi, A.S. Lagunenko, E M. Pazukhin, V.F. Strizhov, V.A. Khvoshchinskii, 2008, published in Radiokhimiya, 2008, Vol. 50, No. 6, pp. 565–568.

650

Formation and Spread of Chernobyl Lavas S. A. Bogatova, A. A. Borovoia, A. S. Lagunenkob, E. M. Pazukhinb,

V. F. Strizhovc, and V. A. Khvoshchinskiia

a Russian Research Centre Kurchatov Institute, Moscow, Russia b Institute of Safety Problems of Nuclear Power Plants, National Academy of Sciences of Ukraine,

Chernobyl, Kiev oblast, Ukraine c Nuclear Safety Institute, Russian Academy of Sciences, Moscow, Russia

Received May 8, 2008

Abstract—The results of implementation of the ISTC Project “Development of a Model of the Behavior of Nuclear Fuel in the Active Stage of the Chernobyl Accident” are summarized. The project work was jointly undertaken by the Russian Research Centre Kurchatov Institute (RNTs KI) and Nuclear Safety Institute, Russian Academy of Sciences (IBRAE RAN). The aim of the project was to systematize a huge amount of data on lava-like fuel-containing materials (LFCM), collected over the 20 years of research work at the Shelter, into database and to develop a model of LFCM formation and spread during early post-accident days. The model will be helpful in producing recommendations on accident prevention and promote development of the best technologies for lava removal, thereby reducing the financial expenses and dose burden. Also, with the model developed, the results of the virtually unique “experiment” with the nuclear fuel of the Unit 4 reactor, organized on a huge scale, can be used for finding solutions to the general nuclear safety problems.

We summarize here the results of implementation of the ISTC Project “Development of a Model of the Behavior of Nuclear Fuel during the Active Stage of the Chernobyl Accident” by the Russian Research Centre Kurchatov Institute (RNTs KI) jointly with the Nuclear Safety Institute, Russian Academy of Sciences (IBRAE RAN). The aim of the project was to systema-tize a huge amount of data on lava-like fuel-containing materials (LFCM), collected over the 20 years of re-search work at the Shelter, into a database and to de-velop a model of LFCM formation and spread during early post-accident days. This task is being accom-plished, above all, as part of practical activities aimed to eliminate of the consequences of the Chernobyl ac-cident. At the present time, conversion of the Shelter to the environmentally safe condition, assisted finan-cially, technically, and organizationally by the entire world community, is under way in Chernobyl. Stabili-zation of building structures has already been achieved. Today’s agenda includes construction of a new safe Arch confinement to enclose the existing Shelter. The next, most complicated, task will consist in removal from the Shelter of the nuclear fuel and radioactive materials to be eventually disposed of. As

known, much of this fuel (~100 t) is part of the lava formed during the accident [1]. In this context, the lava formation and spread model developed can be helpful in determining more precisely the amount, location, and properties of the lava in inaccessible rooms, as well as in producing recommendations on accident prevention, and will promote development of the best lava removal technologies, thereby reducing the finan-cial expenses and dose burden.

PACS numbers: 28.41.Te, 89.60.Gg DOI: 10.1134/S1066362208050131

Second, the Project goals are topical in view of the fact that, with the lava formation and spread model, the results of the virtually unique “experiment” with the nuclear fuel of the Unit 4 reactor, organized on a huge scale, can be used for finding solutions to the general nuclear safety problems, associated with formation of corium.

After the explosions in Unit 4 of Chernobyl NPP, the condition of the rooms in which the lava formation began had already little in common with the pre-accident situation. The same was true of the structures in these rooms. Therefore, the first task in development of a lava formation model was to reconstruct the con-dition of the destroyed unit after the explosions (the

RADIOCHEMISTRY Vol. 50 No. 6 2008

FORMATION AND SPREAD OF CHERNOBYL LAVAS 651

conditionally chosen time was half an hour after the beginning of the accident). In undertaking these ef-forts, we did not seek to reconstruct the entire se-quence of events that occurred during the accident.

Three data files were used: initial data on the struc-tures and materials of Unit 4; results of inspections of the post-accident condition of the reactor cavity and subreator rooms; and results of examination of the ge-ometry and physicochemical properties of the LFCM accumulations in the Shelter.

We carried out verification, analysis, and systemati-zation of a huge amount of experimental data yielded by the research activities undertaken in Shelter in 1986–2005 by RNTs KI, Radium Institute, IBRAE RAN, Chernobyl NPP, Shelter ISTC, and many other organizations involved in elimination of the conse-quences of the Chernobyl accident (see, e.g., [1–8]). We used the data contained in articles, reports, inspec-tion acts, as well as construction drawings, etc. Also, photo and video materials were attracted. The collec-tion of information was complete in 2005. For the re-sults, including details on the composition and amount of the materials that became part of the lava (Table 1), see [9, 10].

In simulation of the lava formation processes we took into account three heat energy sources: residual heat liberation from fuel of Unit 4 of Chernobyl NPP; heat energy from graphite burning; and heat energy from the steam–zirconium reaction, of which the first is of deciding importance (Fig. 1).

In formation of lava, many materials that become its part or contacted its major streams acted as peculiar

Table 1. Materials available in the reactor cavity (room no. 504/2) and in the subreactor room no. 305/2 at the beginning of stage 2 of the accident

Material Available in room nos. 504/2a and 305/2, t Became part of LFCM, t Fuel (U) 120 90 Steel 1300b <20c Serpentinite mixture 580 160 Subreactor plate concrete – 130 Structural concrete that got into the cavity 950 480 Cavity filling materials 300 280 Zirconium ? 45 Graphite 750 Negligible amount

a Within the reactor space. b Disregarding scheme S materials and unfused communication lines at the reactor bottom. c 330 t melted and spread over the subreactor rooms.

thermometers, which enabled reconstruction of the temperature ranges of the processes involved.

Molten metal. The major source of this metal was steel from the destroyed sector of OR (reactor base) scheme [1]. The temperature at which the metal melted and spread must have exceeded 1500oС. There were metal thermometers in the lava itself. All types of LFCMs comprise metal spheres (globules), which also evidences that the lower limit of the lava formation temperature exceeded 1500oС.

Fig. 1. Contribution to the total thermal power spent for lava formation from various heat energy sources (for one of the lava formation scenarios): (1) integrated power spent for the lava formation, (2) contribution from residual heat energy liberated by fuel, and (3) contribution from chemi-cal reactions.

RADIOCHEMISTRY Vol. 50 No. 6 2008

BOGATOV et al. 652

Fig. 2. Tentative liquidus projection for the UO2–SiO2–ZrO2 ternary system: (F, T) primary crystallization field of (F) solid solutions based on U and Zr oxides with a fluorite structure and (T) tetragonal ZrO2, and (S) crystallization field of SiO2 (cristobalite). The cross-hatched zone corre-sponds to the possible composition of LFCM.

Chernobylite. Chernobylite is a crystalline zirco-nium silicate with 10–12 wt% impurity of uranium, formed during the active stage of the Chernobyl acci-dent and occurring in all LFCMs. It cannot be formed at temperatures below 1600oС.

Molten uranium dioxide globules. They are re-vealed by microscopic examinations of LFCM. As shown by backscattered electron microprobe analysis of the lava, the globules incorporate 5–6 wt % impurity Zr which isomorphically substitutes U. The phase dia-gram of the ZrO2–UO2 system shows that the oxide with this composition melts at 2500–2600oC. The de-veloped lava formation scenario took specifically this temperature range as the maximum possible one, valid for a short period. The effective temperature and the time of exposure to it for the nuclear fuel were addi-tionally estimated from the diffusion coefficients of Cs and Sr [1].

The sequence of stages in formation of the Cherno-byl lavas can be very briefly outlined as follows (see, e.g., [3, 10]). Fast increase in the reactor power level and disturbance of the cooling regime caused a sharp increase in the temperature of the fuel, which resulted in its dispersing and breakage of the shells of fuel rods and fuel channels. The explosions that destroyed the core enabled interaction of the fuel fragments with the structural materials, above all, Zr, and later with the serpentinite that filled the OR scheme and the inter-compensator gap, sand, concrete, etc. High-temper-

ature interaction of the fuel with Zr yielded the ura-nium–zirconium eutectic. The contact of the latter with SiO2 (the main component of the lavas) resulted in for-mation of the UO2–SiO2–ZrO2 ternary system. The minimal temperature of the liquidus surface for this system is equal to the melting point of the ternary eutectic, ca. 1500oС1 (Fig. 2) [10].

Presumably, the major part of the lava, the so-called brown and black ceramics, was formed by spe-cifically this scheme.

Along with U, Zr, Si, and O (elemental analysis of the LFCM revealed nearly 20 elements), a significant amount of other elements became part of LFCM upon interaction of the irradiated fuel with structural mate-rials.

The conditions in selected sectors of room no. 305/2 could favor lava formation by different routes as well. Indeed, as shown above, the temperature could lie within 1500–2600oС. Naturally, we can presume the existence of several lava formation sites; each of them had its characteristic temperatures and formed lavas with a composition specific for this site. In se-lected lava formation sites, the conditions could favor interaction of the core fragments according to other phase diagrams. Examples can be found in direct melt-ing of UO2 pellets (at temperatures of 2800°С and above) or UO2–SiO2 interaction according to the dia-gram in Fig. 3 [11]. This diagram suggests that, at 2100–2300°С, the U content in the resulting system can significantly exceed its average content in the lava (see below).

Our discussion should take into account the fact that, in the subreactor room, there exist not only neat lava but also mixtures of the lava with the unfused core fragments (UO2) [1]. This is essential for nuclear safety purposes, since calculations show that, in many cases, the lava + core fragments + water combination is more hazardous than lava combined with water [12].

An important result of the project implementation in 2006 was mapping of the accumulations of LFCM in room no. 305/2 [10]. To this end, room no. 305/2 was conditionally divided into sectors by drawing characteristic cross sections. For each of them we iden-

1 Considering the impurities in the lavas, such as Al, Mg, Ca, Mn, Ti, Na, and K, the minimal temperature of liquidus of a complex multicomponent system such as ТСМ can be even lower. This is suggested by the experimental viscosity of LFCM within the glass transition point area and the experimental melting temperature of natural lavas, analogs of the Chernobyl lavas.

E ~

F

RADIOCHEMISTRY Vol. 50 No. 6 2008

FORMATION AND SPREAD OF CHERNOBYL LAVAS 653

Table 2. Estimated volume and amount of fuel in the big horizontal stream (according to the KI–IBRAE model and expert estimates [1])

Room no. Estimated in terms of the KI–IBRAE model Expert estimates [2] LFCM volume, m3 fuel amount, t LFCM volume, m3 fuel amount, t

304/3 60 6.4 50–70 6 ± 2

301/5 24 2.5 8–30 2.0 ± 1.5 303/3 4.5 0.47 2–7 0.5 ± 0.3 301/6 16 1.7 8–30 2.0 ± 1.5 Total 104 11 105 ± 35a 11 ± 4a

a Taking into account the lava in the “Elephant Leg.”

tified the LFCM spread area based on the entire body of the available verified information.

The developed lava formation model suggests that, deep in the subreactor room no. 305/2, there can be accumulations of the fuel containing a significant amount of uranium. At the present time, the combined results of thermal and neutron measurements suggest the existence of two such hypothetical accumulations [12]. An increase in criticality for these accumulations is associated with the entry and outflow of water from their location sites. Unfortunately, the existing knowl-edge does not allow precise estimation of the hazard they pose. This necessitates keeping alert against pos-sible risks, in particular, preventive introduction of neutron absorbers into the accumulation. After the Arch would enclose the Shelter, the amount of water getting into the Shelter would sharply decrease, so that the changes of the neutron fields, associated with these accumulations, could become unnoticeable. This does not mean, however, that one can disregard the potential hazard they pose. Thus, all the appropriate measures for preventing critical incidents should be provided for the activities undertaken inside the Shelter.

The project work included estimation of the effi-ciency of countermeasures taken during the active stage of the accident, e.g., supplying water to the de-stroyed reactor, throwing various materials from heli-copters into the Unit 4 debris, placing a cooled plate under the unit, preventing the development of the “Chinese syndrome,” etc. Based on the current knowl-edge and using the KI–IBRAE model, we attempted to reconstruct the situation and elucidate how the coun-termeasures affected in reality the processes that oc-curred in the destroyed unit [13].

The lava spread model took into account the LFCM composition and thermal characteristics (residual heat

energy liberation, initial temperatures of the lava masses, temperature dependence of the lava viscosity, etc.). Special attention was given to the routes by which the lava spread and to its interaction with the building structures.

In lava spread simulation, we verified the model parameters for the so-called “horizontal stream,” which is fairly adequately understood. We identified the lava spread area in those rooms (sectors of rooms) for which experimental data are lacking and also esti-mated the amounts of lava in the rooms and in the en-tire horizontal stream. The calculations in terms of the KI–IBRAE model confirm that rooms at the +9.000 m mark are filled with solidified LFCM, as indicated in Table 2.

Table 2 suggests that the results of the model calcu-lations agree well with the previous estimates. The model also allows excluding from consideration those scenarios in which the lava of the horizontal stream has an appreciably smaller volume and contains an appreciably smaller amount of fuel compared to the data from Table 2: In that case, cooling would render

Fig. 3. Phase diagram of the UO2–SiO2 system.

RADIOCHEMISTRY Vol. 50 No. 6 2008

BOGATOV et al. 654

impossible its spread over rooms at the +9000 m mark and for a distance of ~25 m.

REFERENCES

1. Analysis of the Current and Forecasts of the Future Safety Status of the Shelter, Report of the Shelter Interdisciplinary Scientific and Technical Center, Boro-voi, A.A., Executor-in-Chief, Chernobyl, 2001, archive no. 3836.

2. Anderson, E.B., Bogatov, S.A., Borovoi, A.A., et al., Lava-Like Fuel-Containing Materials of the Shelter, Preprint of the Shelter Interdisciplinary Scientific and Technical Center, Chernobyl, 1993, no. 93-17.

3. Pazukhin, E.M., Radiokhimiya, 1994, vol. 36, no. 2, pp. 97–142.

4. Wells of the Shelter. A Summary of Data (album), Report of the Nuclear and Radiation Safety Division, Shelter Interdisciplinary Scientific and Technical Center, Chernobyl, June 9, 1998, no. 09/05-66.

5. Borovoi, A.A., Lagunenko, A.S., and Pazukhin, E.M., At. Energ., 1998, vol. 84, no. 4, pp. 356–362.

6. Borovoi, A.A., Lagunenko, A.S., and Pazukhin, E.M., Radiokhimiya, 1999, vol. 41, no. 2, pp. 187–192.

7. Borovoi, A.A., Lagunenko, A.S., and Pazukhin, E.M., Probl. Chernobylya, 2000, no. 6, pp. 13–16.

8. Borovoi, A. and Gavrilov, S., Development of the Database “Nuclear Fuel and Radioactive Waste in the Shelter of Chernobyl NPP,” Preprint of the Nuclear Safety Inst., Russian Acad. Sci., Moscow, 2001, no. 2001–01.

9. Bogatov, S., Borovoi, A., Gavrilov, S., et al., Half an Hour after the Beginning of the Accident, Moscow: OKPRINT, 2006.

10. Bogatov, S.A., Borovoi, A.A., Gavrilov, S.L., et al., The Database on the Pre- and Post-Accidental Location and Condition of the Nuclear Fuel of Unit 4 of the Chernobyl NPP, Preprint of the Russian Research Centre Kurchatov Inst., February 1, 2007, no. 130–11/2.

11. Bechta, S., Phase Diagrams for Multicomponent Systems Containing Corium and Products of Its Interaction with NPP Materials (CORPHAD.2), 11th CEG–SAM Meet., Dresden, March 6–9, 2007.

12. Borovoi, A.A., Materials on Nuclear Safety of the Shelter, Preprint of the Russian Research Centre Kurchatov Inst., April 6, 2007, no. 130–11/3.

13. Borovoi, A.A., Pazukhin, E.M., and Strizhov, V.F., The Efficiency of Measures on Elimination of the Con-sequences of the Chernobyl Accident (Active Stage of the Accident), Preprint of the Russian Research Centre Kurchatov Institute, Moscow, 2007, no. 6471/11.