杏吧原创

IMAGINE a container full of nuclear waste buried 1 kilometre below the
Earth鈥檚 surface鈥攁 subterranean oven packed with steel, clay, concrete and
radioactive material. If you thought such a place is of no interest to
microbiologists, you鈥檇 be wrong. After all, if bacteria once survived the
rigours of Mars, as some NASA scientists believe, why not the less harsh
conditions of a radioactive waste repository?

In fact, so long as they have water, microorganisms can adapt to almost any
environment on the Earth鈥檚 surface or beneath it. Bacteria thrive, for example,
around volcanic vents on the ocean floor, extracting energy from sulphur
compounds at temperatures claimed to be more than 200 掳C. Bacteria and fungi
were also found in the coolant of the Three Mile Island nuclear reactor in
Pennsylvania after it narrowly escaped meltdown in 1979. Next to these examples,
the environment of an underground waste store seems almost benign. Although a
repository for the most radioactive waste might exceed 100 掳C for decades,
it will be nothing like as hot as an undersea vent. And while the level of
radiation in this type of repository could hit 200 sieverts a year鈥�100 000
times normal background levels鈥攂acteria in a reactor鈥檚 cooling water would
be exposed to many times this dose.

Bacteria have a number of ways of getting into a repository. They can hitch a
lift in the waste, for example. Alternatively, we know from mining that even in
rock that contains no microbes to begin with, the act of excavating underground
structures can introduce a wide range of potential colonists.

Underground repositories will have to last for tens of thousands of years, if
not longer. To contain the waste, engineers plan a series of barriers that they
know will eventually break down. Their hope is that through careful design they
can delay any leaks until radioactivity in the waste has decayed to safe levels.
It is clear to most people that chemical, physical and geological factors must
be taken into account in those designs. Not so obvious, however, is the impact
of microbial action on the containment barriers. But as more underground
repositories are planned, and their likely safety comes under increasing
scrutiny, those effects are being investigated in detail.

Deep disposal is being considered by an increasing number of nations as a way
to dispose of civil and military nuclear waste. Some repositories are planned to
take high-level waste, others for less radioactive material (see
鈥淗ot stuff鈥�).
Britain, France, the US, Canada and other countries are all
investigating variations on the theme. And deep disposal of intermediate and
low-level waste is already taking place in countries such as Finland. In
Switzerland, the nuclear waste company, NAGRA, has put forward a plan for a deep
repository in the north of the country for high-level waste from the
reprocessing of spent fuel.

Confidence that a repository will keep its charge safe for millennia is built
on studies of 鈥渨orst case scenarios鈥濃�攖he application of Murphy鈥檚 law,
which assumes that anything that can go wrong will go wrong, at the worst
possible time and in the worst possible way. For bacteria, this means assuming
that a colonising population adapts quickly to the conditions in the repository
and grows at the maximum possible rate.

Living off iron

In order to grow, bacteria need certain elements鈥攎ainly carbon,
nitrogen, sulphur and phosphorus鈥攎ost of which are present in ground water
or in the nuclear waste itself. They also need energy, which is available from a
variety of biochemical reactions. In the presence of oxygen, for example,
Thiobacillus ferrooxidans thrives by capturing the energy released by the
oxidation of iron from its ferrous (Fe²⁺) to its ferric (Fe³⁺) state. On the
other hand, iron reducers such as Geobacter species and Shewenella
putrefaciens need no oxygen. They couple the oxidation of organic compounds
or hydrogen to the reduction of iron from its ferric to its ferrous state.

These factors are incorporated into our MGSE model鈥攆or microbe growth
in subsurface environments鈥攚hich estimates the growth rate of populations
of bacteria ideally suited to conditions underground. From an inventory of the
elements and energy in a set volume of the repository, the model predicts the
maximum growth rate. From this we can estimate the quantity of any microbial
by-products, and their likely effects.

We first applied this method to the NAGRA plan for storing waste from
reprocessed fuel. The plan calls for the high-level waste to be fixed in molten
glass and sealed in steel cylinders 25 centimetres thick. These would be buried
1200 metres underground, surrounded by a thick layer of compacted clay.

While the whole repository is designed to last for millennia, the lifetime of
the canisters is 1000 years. We looked at whether bacterial action would destroy
the canisters much sooner than this. Nutrients and energy would be in short
supply within the repository, but bacteria could still be active. Jim Philp and
colleagues at Napier University, Edinburgh, have shown that sulphate-reducing
bacteria such as Desulfotomaculum nigrificans can speed up corrosion of
steel in the absence of oxygen by a factor of up to 1000. We calculated,
however, that under the conditions in NAGRA鈥檚 proposed repository, microbially
induced corrosion would have only a small influence on the canister鈥檚
lifetime.

We also estimated that when the canister does eventually fail, each cubic
centimetre of clay will contain a maximum of 1 million organisms. Typical soil
contains about 10 000 times as many. We conclude that the rate of growth of
these organisms is too slow to make any impact on the containment barriers in
the repository. Simcha Stroes-Gascoyne of the company Atomic Energy of Canada,
has calculated a similar rate of growth for a proposed Canadian system for
disposing of spent fuel that has not been reprocessed.

These results suggest that microbial action will be of only minor
significance in high-level waste disposal sites. The same cannot, however, be
assumed for repositories built to house less radioactive wastes鈥攖he
intermediate and low-level categories. The materials involved range from metals
and ion-exchange resins to clothing and paper. The waste would be packed in
steel drums and grouted with cement. The drums would then be surrounded by
cement (see 鈥淪ealed and delivered鈥�).

The lower radioactivity within such a repository, together with its
hotchpotch of different materials, could well suit bacteria better than
conditions in a high-level waste store. This means that microorganisms could be
more of a threat to the containment barriers. Sulphate-reducing bacteria
will be joined by others that can digest paper and other organic materials.

But what damage could these microbial colonies do? One of the biggest
problems expected鈥攅specially in the presence of organic matter鈥攊s
that they will produce gases such as carbon dioxide, methane and hydrogen. Nirex
has developed a computer program called Gammon to quantify the gas produced both
by microbes and by ordinary chemical reactions such as metal corrosion. We
calculate that over a period of 1000 years the volume of gas produced in a
proposed Swiss repository for low and intermediate-level waste could be as much
as 1 million cubic metres. That鈥檚 1 cubic metre of gas for every cubic metre of
waste. This volume of gas could pose real problems. To start with, bacteria
might incorporate radioactive isotopes, such as carbon-14, in the carbon
dioxide and methane they expel, and this could leak out. Alternatively, if
enough gas builds up, it could force contaminated water along underground cracks
in the rock towards the surface.

Rubber boots

So far, we have assumed the worst that could happen. But what are the real
risks? To calculate the odds that bacteria will in fact undermine the integrity
of a repository, we need to refine our model. One way to do this is to study
natural geological sites that resemble repository conditions. But when we apply
our model to such sites, we find discrepancies.

In geological formations at Maqarin in Jordan, near its borders with Israel
and Syria, chemical conditions are similar to those we expect to find inside a
repository for low and intermediate-level waste. The groundwater at Maqarin is
very alkaline. With a pH of around 13 it is caustic enough to dissolve
rubber boots. Still, some microorganisms鈥攕ulphate-reducing bacteria such
as Desulfovibrio species, for instance鈥攈ave adapted to the
extreme alkalinity. According to our model, these conditions could support a
population of up to 100 million organisms per cubic centimetre. But their
populations are between 1000 and 10 000 times smaller than this. A group led by
Karsten Pedersen of the University of Gothenburg in Sweden is now carrying out
detailed microbiological work at Maqarin to improve our understanding of the
ecosystem.

Another way to refine the model is in the laboratory. But again, we find
discrepancies between the model鈥檚 predictions and experimental results. At the
British Geological Survey at Keyworth near Nottingham, we filled 鈥渟low cookers鈥�
with a coarse soup of low and intermediate-level waste, rock, steel, resins and
concrete. After four years, when we dismantled these vessels, we found both
aerobic and anaerobic bacteria at population densities of about 1 million per
cubic centimetre鈥攁gain much lower than predicted by the model. Put
together, these findings suggest that bacteria may be finding the nutrients in
the repository harder to utilise than the model allows for.

In caves and mines, bacteria form biofilms: large, mixed mats of cooperating
bacteria that can reach thicknesses of up to 4 centimetres (鈥淪lime city鈥�,
New 杏吧原创, 31 August, p 32). It may be that bacteria in a repository
are using some of their energy just to establish and maintain biofilms.

So it is not all bad news. There is also evidence, from observations in the
field and from laboratory experiments that the presence of underground
microorganisms could actually improve the containment. One theoretical
concern, for example, is that microorganisms will break down large, insoluble,
radioactive polymers into smaller, soluble molecules. This would increase the
likelihood of radioactivity getting into groundwater.

Glutinous bioflims

Evidence from natural systems and laboratory experiments, however, shows that
the more complex polymers used to solidify some types of waste鈥攕uch as
bitumen and resins鈥攁re broken down very slowly, and only in the absence of
more palatable materials.

And once they have been broken down, other bacteria find the smaller
molecules easier to consume than the large polymers. This in turn would have a
positive effect鈥攑roducing biomass that could actually immobilise
radioactive isotopes.

There are other benefits too. Bacteria that remove oxygen from the repository
create an environment that favours the formation of iron minerals that bind
radioactive elements in their crystal structures. Ferrous oxide, for example,
can capture uranium atoms. These minerals sometimes form within the
microorganisms themselves, and stay there until they decay to safe levels. Since
bacteria prefer to be stationary in glutinous biofilms rather than free-floating
and mobile, there is little danger that they will carry their radioactive burden
out of the repository.

Despite all the effort that has gone into understanding how microorganisms
will behave in nuclear waste repositories, their overall effect has yet to be
established. Until they are understood, and can be controlled, extreme caution
is called for. The authorities that license repositories know this only too
well, and so tend to apply Schweingruber鈥檚 corollary to Murphy鈥檚 law: 鈥淢urphy
was an optimist.鈥�

* * *

Sealed and delivered

THE design for a deep repository for high-level waste put forward by the
Swiss company NAGRA includes a series of barriers. First, the waste is sealed in
a glass matrix which is highly resistant to radiation damage, binds
radionuclides such as uranium and plutonium, and corrodes very slowly. The glass
matrix is sealed into steel cylinders which should isolate the waste for a
minimum of 1000 years. The canisters are designed so that their corrosion
products include iron oxyhydroxides which act as a chemical buffer that absorbs
radionuclides.

The canister is surrounded by a backfill of bentonite, a clay that slows
diffusion of dissolved radionuclides. The final barrier is geological: the
repository will be built deep in rock that will contain any escaping
radionuclides, that has little movement of groundwater, and that can offer
physical protection to the engineered barriers.

In deep repositories for low and intermediate-level waste, the consensus
seems to be that radionuclides should be physically restrained in steel drums
filled with a highly alkaline grout鈥攁 special type of cement. The high
pH is important because radionuclides tend to be less soluble in
alkaline conditions than in neutral or acidic media. The backfill used to
surround the steel drums will also be alkaline. Once again, the site of these
repositories should be chosen so that the radionuclides take tens of thousands
of years to reach drinking water aquifers, by which time the levels of
radioactivity should have decayed to safe levels.