ML20151R312
| ML20151R312 | |
| Person / Time | |
|---|---|
| Issue date: | 01/10/1986 |
| From: | Chan L, Ryder C NRC OFFICE OF NUCLEAR REGULATORY RESEARCH (RES) |
| To: | Silberberg M NRC OFFICE OF NUCLEAR REGULATORY RESEARCH (RES) |
| Shared Package | |
| ML20151R316 | List: |
| References | |
| NUDOCS 8602060021 | |
| Download: ML20151R312 (8) | |
Text
%
0
/ ma nerg[o UNITED STATES g
NUCLEAR REGULATORY COMMISSION n
y C
WASHINGTON, D. C. 20555 J Aft 1 0 W MEMORANDUM FOR:
M. Silb'erberg, Branch Chief Fuel Systems Research Branch Division of Acciden+ Evaluation FROM:
Lisa Chan Fuel Systems Research Branch Division of Accident Evaluation Christopher Ryder Fuel Systems Research Branch Division of Accident Evaluation SUP1ECT:
SUMMARY
OF THE MEETING ON THE EFFECT OF IONIZING RADIATION ON THE STAR!LTTY OF CESTUM TODIDE INTP0 DUCTION A series of scoping ~ experiments were conducted at the.Sandia National Laboratories (SNL) to study the effect of innizing radiation on the stabilitv of Csl in a reaction.
Results from the axpariments are reoorted in Enclosure 1 and indicated a significant decompositinn of Cs!. Casium was ratained in the steel pipe of the experimental apparatus and iodine was collectad downstream in the' condensate tank. Many interpretations of the exnarimental observations have been proposed in the technical community. To raview these and other proposals and to discuss further experimants to determine tha effect of radiation over a wida range of accident conditions, a meetino was hald at the Argonne National Laboratory (ANL) on December 12 and 13, 1985.
Three indeoandent consultants whose expertise lies in the area of radiation chemistry were invited to the meeting.
Thay were Dr. A. Allen, a retired scientist from the Brookhaven National Laboratory (and a scientist on the former study aroup of the American Physical Society), Dr. N. Bidler from Savannah River Lahnratorv, and Dr. M. Sauer from ANL.
Dr. S. Gordon from ANL was also invited, but was unable to attend the meeting.
Peoresentatives from the varinus research laboratories were also present.
A list of tha meetino attendees is attached in Enclosure 2.
The meeting was structured in the form of oresentatinrs followed by discussions and recommendations. The first day of the meetina consisted of opening remarks by M. Silberberg, presentations describing the Sandia work by Elrick and Powers, presentations on Electric Power Research Institute (FPPI) and Atomic Energy of Canada Limited (AECL)--Whiteshell comments and interoretations on the Sandia work by Rit7 man and Wren resoectively, and finally round tabla discussions. The second day consisted of discussions about follow-un research.
The meeting agenda and presantations are Enclosura 3.
8602060021 060110 PDR TOPRP EXISANL D
r s
M. Silberberg JAN 10 m RADIATION EXPERIMENTS WITH Csl Separate effect experiments were condJcted previously at SNL to investigate the transport behavior of Csl in the presence of structural materials.
Results from these experiments indicated little interaction between Csl and structural surfaces such as stainless steel and Inconel in steam for temperatures up to 1300 deg K.
However, the interaction was not studied in a radiation field which is expected in the reactor coolant system during a severe accident. To determine the effect of radiation on Csl transport, a series of scoping experiments were conducted with the reaction zone of the apparatus exposed to a Co-60 source. The design of and the conditions in the apparatus were kept the same as those used in the previous experiments with CsI and stainless steel so that a direct comparison of the experimental results could be made. A total of three experiments were conducted in the radiation field.
In these experiments, a mixture of CsI, steam and H,3 (from the steam oxidation of stainless steel during the experiment) were passed through a 1-inch diameter pipe where deposition coupons (stainless steel and silver plated copper) with different pretreatments were placed along the pipe. The flow exiting the pipe was collected in a condensate tank. Sequential samples of the condensate were taken for post-test analysis.
Except for the presence of a radiation field, the only notable difference between the present tests and the previous Csl tests was in the time of test duration. The run duration for the radiation tests was consistently kept at 9 hours1.041667e-4 days <br />0.0025 hours <br />1.488095e-5 weeks <br />3.4245e-6 months <br /> while that for the' previous tests was 3 hours3.472222e-5 days <br />8.333333e-4 hours <br />4.960317e-6 weeks <br />1.1415e-6 months <br />.
IfCslsurvivesthereactionzone,withorwithoutradjation,andentersthe condensate tank as such, the molar concentration of Cs and I in the condensate should be the same. Condensate samples from the experiments without radiation.showed the presence of eoual amounts of cesium and iodine indicating that there was no interaction between Csl and stainless steel.
However, condensate from the radiation experiments indicated that iodine was present at quantities (ppm) at least 5 times higher than the amount of cesium; as explained below, the unaccounted cesium was found to be retained within the
' piping.
Examination of the deposition coupons from the experiments without radiation showed two oxide layers on their surfaces. The outer layer was found to contain Fe 0 while the inner layer consisted mainly of an iron-chromium 34 oxide. No cesium or iodine was found in either layers. Analysis of some of the deposition coupons from the radiation experiments showed three layers.
The outer two layers had similar compositions as those found on coupons in the absence of radiation. The third layer was observed only on the radiated coupons.
Cesium was found to be deposited in the middle layer in the radiation experiments.
From these results, it seems that gamma radiation caused the Csl to decompose'and the cesium to deposit on the stainless steel surface.
Volatile iodide was believed to have formed and moved downstream to the condensate tank.
The discussion on this part of the presentation focussed on three issues, the hydrogen to steam ratio in the radiation experiments, the acidity of the condensate tank, and the form of the iodine measured in the tank.
i M. Silberberg JAN 10 g present in the experimental system was a result of steam oxidation of stainless steel.
It is important to know the amount of hydroger generated since the presence of hydrogen can alter the chemistry of the stainless steel surface and possibly that of the gas phase. The quantity of hydrogen generated in the experiments was measured and was believed to be at a low enough level which would not alter the chemistry of the system.
The pH of the condensate samples from the previous Csl experiments and the present radiation experiments was approximately 2, and 3 to 4, respectively.
If all of the iodines entered the tank as HI, the acidity of the condensate would be about pH 4.
This still cannot explain the lower pH values of some of the condensate s'amples. The acidity of the condensate from the previous experiments may have been caused by-the asbestos gaskets in the apparatus.
These~ gaskets were removed in the latter experiments. A more probable cause of the low pH in the condensate is residue from the acid that was used to clean the apparatus after each experiment.
A detailed analysis of the forms of iodine in the condensate tank was not done. This analysis may have suggested the forms of volatile iodine that existed in the system after the reaction zone. At present, only iodide (I')
was measured.
Significant amounts of other forms of iodine are expected in the condensate tank after long durations and when the condensate is acidic.
Concern was also expressed that a direct comparison between results from
~
experiments with and without radiation cannot be made because of differences in the apparatus of each set of experiments. Although the design and the size of the apparatus were the same, the concern was not diminished.
Ritzman presented a list of prototypic severe accident conditions and compared them to those used in the Sandia experiments. He concluded that some of the conditions used in the Sandia experiments are near the low ends of the ranges of conditions he cited, in particular, the bulk gas velocity, the hydrogen to steam ratio, and the radiation level. He noted that the gas composition in the radiation experiments was also far from prototypical because the presence of other fission products and aerosols were not considered.
Ritzman also presented experimental evidence from EPRI-sponsored in-pile experiments and results from examination of TMI samples which implied the existence of Csl under accident conditions. The in-pile source term experiments at TREAT indicated the presence of crystalline materials in the aerosol sampling system. These crystals were found to contain cesium and iodine in a 1:1 atom ratio. Data on the fission product distribution from TMI-2 also showed the same fractional inventory of cesium and iodine in the remaining core debris, vessel internals, RCS piping, and coolant and possibly in the reactor and auxiliary buildings. Beahm noted that the TMI-? samples analyzed may not be representative of the regions 'in the reactor and plant where the samples were taken and that the average o'f results from a few samples taken from a certain region does not represent data from the overall region.
Cubicciotti was unable to attend the meeting, but provided written comments (Enclosure 4) noting that the test conditions in the Sandia experiments were not well controlled and that the observation from the radiation experiment 3 was the result of an experimental artifact. He also cited the TMI data as
i M. Silberberg support for the existence of CsI.
His comment that radiation could have increased the temperature of the coupon which would have increased the reaction rate and decomposition of CsI, was shown to be invalid.
Wren-estimated that a temperature increase of only 25 deg V could be expected on the. coupon surface. This is insignificant compared to the temperature of the experiment (1300 deg K).
EXPERIMENTAL ANALYSIS Sandia attempted to identify the mechanism which resulted in the decomposition of Csl in a radiation field and in the presence of structural materials by means of a thermodynamic analysis based on the hypothesis of gas phase ion interacticn with CsI. When the experimental system was exposed to gamma rays, the structure lying directly in the path of the ray was the steel piping containing the gaseous mixture of CsI, steam, H, and a carrier gas.
This 3
resulted.in the interaction of radiation with the steel piping and the generation of secondary electrons.
It.is believed that these electrons will subsequently interact with the vapor mixture, particularly steam, to create ionic species such as H and additional electrons. These ions and electrons willthenreactwithCs}0causing decomposition. Assuming that ion formation is faster than its destruction, calculations carried out for H,/P 0 ratios up 9
to 10 and total pgessure up to 100 atmospheres showed the domihant ionic species to be H,0 and I, and the dominant cesium species to be Cs0H (and cesium at high Ho concentrations). Based on the calculations, Sandia postulated the overalt reaction which caused the decomposition of Csl to be:
2H O + Csl H
+ Cs0H + f 2
3 The effect of gamma radiation in this case is to convert CsI to Cs09 and iodide. Since Cs0H was shown previously (in separate experiments) to react with stainless steel in the presence of steam (NUREG/CR-3197), Sandia believed thtt the observation of. cesium retention on the steel surface in the radiation experiments was caused by the gas phase formation of Cs0H from Csl followed by surface reaction of Cs0H with stainless steel.
The deposition of Cs0H on steel surfaces freed the iodine from further reforming Csl outside the radiated reaction zone.
The dominant iodine species was predicted to be I.'
Other iodine species such as HI, I and I were calculated to be order of 2
magnitudes lower under the conditions of the experiments.
Since the Sandia results were published, alternative explanations were proposed.
Wren presented substantial analytical arguments against the Sandia explanation. Assuming that all of the energy degosited on the steel was transferred to the steam and used to produce H 0.
The concentration of H 0 was estimated to be 5 orders of magnitude loweh than the concentration of s!
d in the experimental system, indicating that the reaction between the H 0 and Csl is insufficient to account for the observations.
Ifinsteadallofthe energy were used to ionize CsI, which will subsequently decompose, the amount of Csl affected was predicted to be no more than 0.01% of the initfil quantity of Csl in the system.
Independent calculations conducted by Allen (Enclosure 5),
and Alexander also indicated Csl conversion and ion production rate to be insufficient.
M. Silberberg JAN 101Mi Further calculations by Wren showed the rate of H 0+' destruction to be 3
generally faster than ion production and chemical reaction.
Kinetic calculations which consider gas phase interactions involving radicals (instead of ions) were also conducted.
Results indicated that radiation induced radical reactions occurred only for conditions of low Csl concentrations (about 1 ppm) and high radiation fields (about 30 krad/sec).
These conditions did not. exist in the Sandia radiation experiments, but are expected to be present in the RCS under severe accident conditions.
Two other ways for Csl to decompose are by oxidation and by thermal degradation. These explanations are unlikely. Oxidation could have occurred only with a large excess of oxygen in the system, an excess which is not evident. Thermal degradation could have occurred only if a large error in the temperature measurements occurred, an error which is not evident.
In conclusion, Wren suggested an alternative explanation involving the metal surface of the piping.
He hypothesized that gamma radiation may alter the oxide surface of stainless steel which becomes reactive with Csl.
The invited consultants, Allen, Bidler*, and Sauer supported the idea of surface reaction instead of gas phase ionic interaction as the explanation for the radiation' induced Csl decomposition.
Allen proposed a similar explanation.
He believed that gamma radiation causes the excitation or the formation of reactive sites on the stainless steel surface.
Csl molecules transported to the solid surface would absorb some of this excitation energy and subsecuently hydrolyze to form Cs0H and iodine. Cs0H then reacts with the structural elements in the oxide layers while iodine diffuses back to the gas phase to form HI. Alexander commented that such reactive sites or imperfections on the stainless steel surface can be annealed easily at the high temperatures of the Sandia experiments.
Sauer agreed with and extended the idea of a surface reaction for the disappearance of cesium (Enclosure 6).
First, he calculated a G-value of about 1000 to 10000 atoms /100 eV of energy absorbed in the gas. Pecause this value is unrealistically high, a surface reaction must be considered. Second, he calculated a G-value of 82 atoms /100 eV of energy absorbed in the oxide layer. Because this value is also high, he suggested that energy absorbed beneath the oxide layer may be important.
RECOMMENDATIONS-Recommendations for further experiments were made by the meeting attendees.
It was suggested that a material balance for the Sandia experiments should be carried out, but. prior to such analysis, analysis of the remaining condensate samples should be completed to provide the necessary data.
The reference Csl experiments in the absence of a' radiation fibld should be rerun with the present apparatus so that a direct comparison between results from experiments with and without radiation can be made.
Powers proposed an experiment (s) which not only confirms the present experimental observation under radiated conditions but also satisfies the need for a reference test.
This is accomplished by' repeating one or more of the previous experiments at the same
- Written comments will appear at a later date.
M. Silberberg JAN 101986 experimental conditions but with the radiation source on and off in an.
alternating pattern. Results from radiation-on and radiatier-off conditions would be compared to determine the effect of radiation, fonducting the test (s) in an alternating pattern for several times would provide information on the stainless steel surface. When confirmation is obtained, additional tests or detailed surface analysis need to be planned to determine the surface mechanism responsible for the observation.
As discussed before, analytical results obtained by Wren indicated that a gas phase interaction involving radicals and Csl could occur under the conditions of high radiation level (about 30 krad/sec) and low Csl concentrations (about 1 ppm). These conditions were not satisfied in the Sandia radiation experiments and consequently, this rechanism is not the cause of observed Csl dissociation.
However, the conditions of high radiation levels and low Csl concentrations may exist in the reactor coolant system during a severe accident.
Gas phase interaction resulting in Csl decomposition, therefore, may play an important role in actual reactor accidents.
Powers will conduct analyses to determine the extent of such interaction under various accident conditions.
Several attendees commented that independent confirmation experiments should be carried out at other facilities.
EPRI responded that such tests are under consideration and their candidate contractor is AECL at Whiteshell. A description was subsequently presented on the proposed EPRI-Whiteshell tests.
The experimental apparatus is similar to the one at SNL. An enclosed Co-60 camera will be built which will provide a maximum of 350 rad./sec. at the pipe centerline. A total of four tests are planned in Phase 1 of the program.
The first two tests are essentially repetitionsTof the Sandia experiments. The third test will use carbon steel deposition coupons instead of 304 stainless steel and test 4 will repeat test 1 except at a higher steam flow rate (a factor of 10). The proposed schedule for Phase 1 is approximately six months after program initiation.
EPRI recommended that further experiments should be conducted at relevant accident conditions, namely, higher Csl concentrations and higher radiation levels. Disagreements were voiced on the value of a prototypic radiation level. However, there was a strong consensus that the Cs! concentration cited by EPRI is high and that lower concentrations than the one used in the current experiments should be employed in future experiments.
Hobbins recommended that a Cs/I and H,,/H 0 ratios of 10 and a total pressure 7
of 100-atmospheres should be used in future experiments.
It is argued that high pressure tests are difficult to conduct because of design difficulties.
A suggestion was made that low pressure tests should be conducted first to understand the mechanism and if confirmation is necessary, then an optimum number of high pressure tests could be carried out.
The higher Cs/I ratio suggested is recognized to be important since surface saturation of the reaction sites may play an important role in'Cs! decomposition.
Another recommendation was that the experiments be done with a real time detector, such as laser fluorescence, to obserle both the reactants entering the radiation field and the products leaving the radiation field.
M. Silberberg JAN 10 MMi Powers recommended that an assessment should be made on the impact of this issue, namely, the chemical. form of indine in the reactor coolant system, nn the final iodine source terms in sovere accidents.
The NRC has initiatad an intensive analytical effort at ORNL to estimate the containnent iodine source term assuming that Csl and/or volatile fodines are formed in the RCS for several accident sequences and plants la PWR, a BWP, and possibly an ice-
' condenser plant). The PWR calculations are scheduled to be completed in February and the BWR calculations, in Pay.
Lisa Chan Fuel Systems Research Pranch Division of Accident Evaluation Christopher Ryder Fuel Systems Research Branch Division of Accident Evaluation
Enclosures:
0 '#"'
6 as stated
.