Is 85Kr from nuclear fuel rods really a significant risk to human health?

[Enthalpy]

Krypton-85 ⁸⁵Kr is a beta emitter with 687keV maximum energy and 10.7 year half-life. It's produced by uranium fission with 0.27% yield (137Cs and 90Sr: 6%). Here is a report, a bit old and in French, but figures read the same in English anyway:
http://www.irsn.fr/EN/Research/publications-documentation/radionuclides-sheets/Documents/Kr85SAN.pdf

When reprocessing used nuclear fuel rods at La Hague, solid fission products like ¹³⁷Cs and ⁹⁰Sr are separated for storage, but ⁸⁵Kr is just diluted and emitted in the atmosphere. For instance in 1999, La Hague emitted 2.9*10¹⁷ Bq of ⁸⁵Kr, more than the radioactivity of ¹³⁷Cs and ⁹⁰Sr released by the Chernobyl disaster.

Iodine, caesium, strontium radioisotopes fall on the soil more quickly and locally if emitted, and accumulate in the food chain. Krypton, as a noble gas, dilutes in the whole northern hemisphere's atmosphere, where in 2001 it added 1.2 Bq/m³, with 3/4 of it coming from La Hague - it must be worse by now.

How bad is such an added radioactivity, much smaller than natural sources anyway? It irradiates a human body by 4nSv/yr, or 500nSv/yr at the skin as beta rays don't go deep. No experimental data tells the effect of such a low dose; scientific near-consensus is to extrapolate proportionally from higher doses, taking 1% more risk of a fatal cancer per 0.2Sv exposure. Over 30 years, this adds only 6*10^-9 risks of casualty - but for each of the 6 billion people in the hemisphere, meaning an estimated 36 casualties.

Note the 1% per 0.2 Sv is a whole-body average, but external beta rays damage essentially the skin. It seems that 500 nSv/yr at the skin is a bit worse than the deep 4 nSv/yr taken here.

Some people challenge that the risk is proportional to the dose; they want to see some dose threshold below which the risk vanishes. Among them is Areva, the operator of La Hague.

I consider that since ⁸⁵Kr arrives from the power plants confined in fuel rods, keeping all ⁸⁵Kr confined in some storage can't be that difficult. And to avoid three dozens possible deaths, it should be done, without arguing about risk models.

 

[nikkkom]

I consider that since ⁸⁵Kr arrives from the power plants confined in fuel rods, keeping all ⁸⁵Kr confined in some storage can't be that difficult. And to avoid three dozens possible deaths, it should be done, without arguing about risk models.

The cheapest solution probably is to reprocess fuel only after several decades of dry cask storage. 50 years gives x32 reduction in Kr-85, 100 years - x1000.

 

[daveb]

Just out of curiosity, where did you get the info on the added dose from this? Normally, krypton (as well as any other noble gas) doesn't stay in the lungs, nor does it get absorbed easily, and would only expose someone if it decays while in the lungs. This is why Xe-133 (and previously, krypton) was used as a radiopharmaceuthical to image lung function.

 

[Enthalpy]

The conversion to the added dose uses the conversion factor given in the linked Pdf.

I also imagined effects at the lungs, but the Pdf logically states that external irradiation is by far more important. That's because the range of betas in the surrounding air vastly exceeds the radius of our lungs, so more betas are susceptible to reach our skin than our lungs.

Which connects to one uncertainty: betas stop mainly in our skin, but I've taken the dose referred to the whole body mass (4 nSv/yr) and used it to evaluate fatalities (by the linear extrapolation). The higher dose (500 nSv/yr) at the more resistant skin may be more threatening.

The other way, in case our lungs are much more sensitive to betas than the average of the body, they might be the main risk, more so than the skin, despite getting a smaller dose.

----------

Storage is the solution for ⁸⁵Kr with 10.7 yr half-life. As the amount makes few m³ at 1atm 300K, a separate storage is perfectly possible as well and permits to reprocess the rest of the spent fuel.

 

[nikkkom]

Storage is the solution for ⁸⁵Kr with 10.7 yr half-life. As the amount makes few m³ at 1atm 300K, a separate storage is perfectly possible as well and permits to reprocess the rest of the spent fuel.

Calculate the heat output and gamma dose of said few m^3. Millions of rads/h. Deeply unfunny stuff. The storage won't be easy and cheap.

Now imagine an accident when the storage vessel overheats and/or cracks... >>>8(

If you let Kr-85 decay while it resides inside ceramic fuel pellets inside spent fuel rods inside steel container inside dry casks, you don't need to design and build a new kind of storage for a new kind of highly radioactive volatile substance. You only need to expand the dry cask storage - the thing you already have designed and built.

Dry cask can not overheat. (Well, it can, but it'll take a deliberate action to make it happen). Even if dry cask is cracked (rather unlikely event), the release will be limited since the fuel ceramic is not a gas.

 

[Enthalpy]

Here an elementary example of a Krypton-85 storage, to give a general sense of its difficulty. I didn't and won't check how to catch ⁸⁵Kr from the fuel rods; boiling nitrogen?

A tube of ID=0.17m and L=6m (sketch below, click to enlarge) stores the ⁸⁵Kr extracted during one mean week at La Hague, at absolute 0.95b if it were to reach +100°C. Bent in a U shape and immersed, it holds the gas even if a seal leaks.

Aluminium AA5083 is one interesting choice. 20mm walls and two thick welded plugs let it sink head up.

Betas of mean 251keV produce 223W heat directly in the wall, very easy to cool. The dose at the walls is several magnitudes below a fission reactor, and betas damage alloys little.

⁸⁵Kr produces a 514keV gamma in 0.43% of the disintegrations. Bremsstrahlung in aluminium adds little. 20mm walls shield the gammas already by a factor >60, so a human standing along the tube in air would get 0.1Sv in >20min.

The produced Rubidium-85 melts readily and the liquid might dissolve aluminium. Water in the tube would catch Rb but the alkali corrodes aluminium. A mild acid instead would still produce hydrogen, not so nice. I imagine a metal with varied valences solves that, like Iron in FePO4 becoming RbFePO4, or Prussian blue, or potassium ferricyanide - ask a chemist. Vacuum doesn't buckle the tube.

A pool of 12m*12m (sketch below, click to enlarge) would store the ⁸⁵Kr extracted during 15.4yr; this is enough if older, emptier tubes are re-filled with newer ⁸⁵Kr, but I'd prefer a bigger pool and fewer operations. Some 3m water shield the gammas and cool the 180kW by natural convection of few dm³/s. An exchanger at one side suffices, easing crane operations. A direct exchanger with some 15m³/s air would be challenging, but heat pipes between the pool and the blown fins make it easy.

Marc Schaefer, aka Enthalpy

 

[Dmytry]

Regarding the linearity of low dose effect, the rationale is that any curve is linear when you zoom in deeply enough at a non-special point on the curve. If you consider the dose-response curve for DNA damage in general, the extra irradiation=0 is a totally unremarkable point on the curve, corresponding to lifetime cancer rate of about 40%.
One plausible kind of threshold effect is an activation of costly defence mechanism at higher dose; if there's such effects, the slope of the dose response curve determined at high doses would be lower than the slope at low doses, that's it, the LNT estimates would under-estimate the impact.

I propose an outside the box solution. Require Areva to fund free skin cancer screening for the impact of saving equivalent number of lives; perhaps with 5x margin to be sure. That may be much more cost effective - if indeed enforced - than storage of Kr-85. Or not. In any case the important thing is to decide to actually do something about the extra deaths vs the current status quo where it is all-ok. (One issue though is that it would put biasing pressure on the science, due to significant commercial interest in promotion of threshold model, and would damage integrity of science as you can't simply promote the threshold model, you have to get rid of fairly basic mathematics and logic for it to be plausible and uncontested. The controversy over low dose effects may be as bad as global warming controversy even though the effect being denied would be much smaller)

 

[Joffan]

Actually the consensus, at least as represented by BEIR VII, is that LNT does not hold at low doses and signficantly overstates the health risks.

In order to continue using LNT in some form, however, BEIR introduces a scaling called "DDREF" - Dose and Dose-Rate Effectiveness Factor. This scales down the LNT calculation to some lower value, which since DDREF is ont a fixed number is a matter of experimentation. However, this breaks the model concept, since rates are no longer linear with actual observed impact at high doses (and introduces a discontinuity somewhere, too).

ICRP observed, http://www.icrp.org/publication.asp?id=ICRP Publication 99 :
"Unless the existence of a threshold is assumed to be virtually certain, the effect of introducing the uncertain possibility of a threshold is equivalent to that of an uncertain increase in the value of DDREF, i.e. merely a variation on the result obtained by ignoring the possibility of a threshold."
- that is, you can ignore the threshold by manipulating DDREF to explain away experimental low-dose results.

 

[Enthalpy]

BEIR VII is publicly available there:

dels-old.nas.edu/dels/rpt_briefs/beir_vii_final.pdf

from which I cite the first paragraph, just under the title:

"A comprehensive review of available bio-logical and biophysical data supports a “linear-no-threshold” (LNT) risk model—that the risk of cancer proceeds in a linear fashion at lower doses without a threshold and that the smallest dose has the potential to cause a small increase in risk to humans."

 

[Enthalpy]

Regarding the linearity of low dose effect, the rationale is that any curve is linear when you zoom in deeply enough at a non-special point on the curve.

I certainly agree that any small dose adds to the existing background, so if the response curve has any reasonable property, the added risk must be proportional to the small added dose.

But strong experimental data exists for higher doses only, which aren't a small addition to natural background. The effect of small added doses would be extremely difficult to observe.

We ignore neither how many natural cancers result from the natural background of ionizing radiation. If natural radiation causes 1/10 or 1/100 of natural cancers, then the added dose is 10 or 100 times less dangerous as well.

Hence proponents of a threshold model seek examples of places with a higher natural radiation.

 

[Enthalpy]

Reprocessing plants for nuclear fuel, like La Hague, emit also tritium, which they dilute in the Ocean.

Here's data (for 1999 and in French, sorry, but figures read the same in English):
http://www.irsn.fr/EN/Research/publications-documentation/radionuclides-sheets/Documents/Tritium_H3_v1.pdf

The amounts are small, and I couldn't consistently estimate the risks for humans; it seems that significant risks would only result from a non-uniform dilution, for instance if the food chain absorbs tritium before it's fully diluted - and that's hard to model.

----------------------

Anyway, if tritium had to be stored, it looks rather easy.

Fine powder of dessicated alumina is sold like plaster, to be mixed with water and moulded in form by the user to make ceramic objects - I suppose the reaction produces Al₂O₃°3H₂O, maybe with a hydroxide fraction. That's a stable, inert compound.

Make dilute superheavy water with the tritium. Let this water react with the dessicated alumina powder to obtain a ceramic, maybe the size of a tennis ball. Then, use normal water to add a 20mm cap of hydrated alumina around the ball, to absorb any Bremsstrahlung from the weak betas.

La Hague emitted 1.3*10¹⁶ Bq/yr of tritium in 1999; over 17.7yr this cumulates only 640g of pure tritium. If ten Al₂O₃ incorporate one tritium atom (for mechanical stability and against proliferation) it takes 340kg of ceramic balls which cumulate only 210W (5.7keV mean electron energy), emit no betas and very little gammas absorbed by shallow water.

 

[Dmytry]

Well, the 'large' doses (many times natural background) are still a small addition to the total DNA damage rate and cancer rates. When you are looking at just the 'radiation' DNA damage, it may seem like you are extrapolating from 0.5 sievert to 0.001 sievert, which is rather huge range, but when you're looking at the DNA damage, you're still on a small piece of a curve. There's nothing special about nuclear radiation; skin in particular takes a lot of damage from UV, and that's what the radiation - even fairly 'big' (compared to background) doses - adds a little to.
The issue is that to detect 1% cancer rise over 40% background with confidence of 95% , well, you need the sample size of about 10 000. For the 0.1% difference from 0.01 Sievert, the sample size increases to 1 million.
This whole 'where is the evidence of cancer rise at low doses?' where the rise of 0.1% is expected, is just not even worth speaking about between anyone who's not numerically illiterate. The best that could be done is the measurement that dose of 0.1 Sievert raises the cancer rate by 1% (from BEIR).

re: tritium, that's one isotope which could conceivably be bio-concentrated as it's atomic weight is trice the atomic weight of hydrogen and it doesn't work like hydrogen in many ways. Fortunately it seems nothing concentrates it.

 

[Joffan]

Enthalpy, I know what the words says, but the formulas used tell a different story. It's not accurate to regard LNT/DDREF as being consistent with LNT.

 

[Caniche]

Sheep ,cows and grass are very efficient at concentrating strontium ,especially in springtme.

 

[Dmytry]

re:LNT/DDREF : some very simple mathematics excludes possibility of any statistically significant, direct empirical data at below 0.1 Sievert. At above 0.1 Sievert there is clinically observable biological response other than cancer, as well as activation of well documented defensive responses that try to minimize the total amount of sustained DNA damage by halting the cell reproduction, so indeed at above 0.1 Sievert (in short timespan) the response does begin to get non-linear. There aren't any observed biological mechanisms that would make the DNA damage response slope between 0.1 and 0.5 Sievert be higher than between 0 and 0.1 Sievert, and there's a plenty of mechanisms that rather decrease the DNA damage (cell division arrest, activation of DNA repair, programmed cell death of cells most affected, etc). Bottom line is that at higher doses, the organism starts doing something about the radiation.
And of course the most important thing here is - the lifetime cancer rate is not around the zero, but around 40%; there is a lot of other DNA damage; the ionizing radiation, even at doses of 0.5 Sievert, is still not a major contributor.

 

[nikkkom]

Reprocessing plants for nuclear fuel, like La Hague, emit also tritium, which they dilute in the Ocean.

Here's data (for 1999 and in French, sorry, but figures read the same in English):
http://www.irsn.fr/EN/Research/publications-documentation/radionuclides-sheets/Documents/Tritium_H3_v1.pdf

The amounts are small, and I couldn't consistently estimate the risks for humans; it seems that significant risks would only result from a non-uniform dilution, for instance if the food chain absorbs tritium before it's fully diluted - and that's hard to model.
...
...
La Hague emitted 1.3*10¹⁶ Bq/yr of tritium in 1999; over 17.7yr this cumulates only 640g of pure tritium. If ten Al₂O₃ incorporate one tritium atom (for mechanical stability and against proliferation) it takes 340kg of ceramic balls which cumulate only 210W (5.7keV mean electron energy), emit no betas and very little gammas absorbed by shallow water.

Becqerels are not born equal. Tritium's betas are VERY weak: 18,5 kEv maximum. That's 0.0186 MeV.

Then, use normal water to add a 20mm cap of hydrated alumina around the ball, to absorb any Bremsstrahlung from the weak betas.

20mm?? Bremsstrahlung of 18.5 kEv beta? That can generate only some miniscule amount of soft X-rays. They are efficiently absorbed by about a millimeter of aluminum.

Tritium also does not bioaccumulate.

 

[Joffan]

Dmytry, the evidence points in the direction that health damage below exposures of 0.1Sv is either zero or disportionately low. That's why DDREF was introduced - the evidence of harm at low doses was below anything consistent with LNT and the model required a reducing factor to bring it back into the realms of statistical feasibility. Somehow no-one had the insight (or courage?) to point out that this breaks the "L" in LNT. A threshold fits the data just as well, with more compatibility with approaches to other risks, but appears to be regarded as heretical.

Read the abstract for ICRP 99, last sentence of first paragraph: "Unless the existence of a threshold is assumed to be virtually certain, the effect of introducing the uncertain possibility of a threshold is equivalent to that of an uncertain increase in the value of DDREF, i.e. merely a variation on the result obtained by ignoring the possibility of a threshold."

Your hypothesizing about increased damage at low levels would be interesting speculation, but the whole body of data doesn't support it.

 

[Caniche]

Maybe someone should study the leukemia trends in Japan?