馃攳
Former fusion scientist on why we won't have fusion power by 2040 - YouTube
Channel: unknown
[1]
there's a famous saying that nuclear
[2]
fusion is 30 years away and always will
[5]
be
[6]
but not anymore it seems
[8]
the united kingdom's atomic energy
[10]
authority is promising an operational
[12]
reactor in 19 years
[14]
several startups are targeting 2030 or
[16]
sooner
[18]
lockheed martin is promising to have a
[20]
working prototype two years ago
[22]
just as soon as their time machine
[24]
division is ready with their project
[26]
in this video i will try to explain why
[28]
fusion energy will unfortunately
[30]
probably not be used to power any part
[32]
of our electrical grid by 2040 all
[35]
sources are in the description
[37]
it takes a lot of time and planning to
[39]
build a major infrastructure project
[41]
even when it's a long proven technology
[43]
like an airport or a rail line if you
[46]
had the money in hand today you'll be
[48]
lucky to build an airport in nine years
[51]
building a major nuclear site with the
[53]
handling of radioactive waste would make
[56]
things many times harder for an
[58]
experimental and totally unproven
[59]
nuclear technology like fusion the
[61]
challenge of ironing out all the
[63]
problems which arise in a couple of
[64]
decades is nearly insurmountable
[68]
suppose you can overcome all the boring
[70]
logistical challenges you will discover
[72]
that actually carrying out fusion
[73]
reactions is both surprisingly easy and
[75]
surprisingly hard
[77]
fusion is a nuclear reaction whereby
[79]
lighter atoms combine into heavier ones
[82]
as long as the atomic weight of the
[84]
final product is less than that of iron
[86]
or nickel energy will be given off two
[88]
oxygen atoms could fuse into a sulfur
[90]
atom and release energy for example the
[92]
challenge is that nuclei are positively
[94]
charged and hence repel each other the
[96]
larger and more charged the nucleus the
[98]
stronger the repulsive force
[101]
conversely hydrogen has the weakest
[103]
charge and its isotopes are the easiest
[105]
to force together the reaction of
[107]
deuterium one proton and one neutron and
[110]
tritium one proton and two neutrons is
[112]
the one which happens most readily of
[114]
all and therefore the one all the
[115]
experiments are targeting two deuterium
[118]
atoms can also fuse very easily the
[120]
repulsive force could be overcome if one
[122]
of the nuclei is held fixed and the
[124]
other is accelerated through a voltage
[126]
of hundred 000 volts
[128]
this is surprisingly easy you see the
[130]
reason youtube is called youtube is that
[132]
when i was a kid we used to have tvs and
[134]
computer screens based on a type of
[136]
particle accelerator called a cathode
[138]
ray tube one of which i bought for 50
[140]
bucks crazy i know but for just 50
[143]
dollars my tv could have accelerated
[145]
hydrogen isotopes through 20 000 volts
[148]
or a fifth of the way to reliably
[150]
achieving fusion
[151]
for a larger but not exactly
[153]
astronomical amount of money i could
[155]
have built a dedicated souped-up version
[157]
of this type of device called a fuser
[159]
this is just what a kid called taylor
[161]
wilson did when he was 14 years old his
[164]
fuser accelerated deuterium so that a
[166]
small fraction underwent fusion and
[168]
released energy the problem is that no
[170]
matter how big or small or how it's
[172]
configured the energy produced by a
[174]
fuser could never be turned into more
[176]
useful electricity than it consumes
[179]
to explain why try the following for me
[181]
rub your hands together to generate heat
[183]
by friction due to fundamental physics
[186]
like the second law of thermodynamics it
[188]
is not practically possible to turn that
[190]
small amount of heat into electricity
[192]
that heat will escape into the
[194]
atmosphere and despite humanity's best
[196]
attempts to blanket the earth and
[198]
greenhouse gases it will eventually be
[200]
emitted as infrared radiation into space
[203]
simply enabling fusion reactions is not
[205]
good enough to generate electricity the
[207]
reason i'm stressing this point so much
[210]
is that many of the startups say they
[211]
will demonstrate or do fusion on their
[214]
roadmaps to attract investment this is
[216]
what the kids are calling a nothing
[218]
burger there is no substance there it's
[220]
impressive for a 14 year old but for a
[222]
multi-million dollar company this is
[224]
trivial and just as with taylor wilson's
[226]
fuser demonstrating a small amount of
[228]
fusion reactions is in no way a
[230]
guarantee that the method could ever
[232]
even theoretically be scaled up to
[234]
produce electricity to have any hope of
[236]
generating useful power we would need to
[238]
do what is called thermonuclear fusion
[241]
in basic terms you must first heat your
[243]
fusion fuel to enormous temperatures of
[245]
tens of millions of degrees and keep it
[247]
confined in a state of plasma for a
[249]
given amount of time which is the part
[251]
that is surprisingly hard
[252]
the initial heating requires a large
[254]
input of energy before any fusion
[256]
reactions will take place so the
[258]
subsequent energy given off must be
[260]
large enough to make up for this
[261]
investment confining the plasma is of
[263]
paramount importance and the subject of
[265]
most current research because bad
[267]
confinement means that the heat will
[269]
leak out and more energy must be put in
[272]
the simplest way to think about it is to
[274]
imagine you are running a business
[275]
selling apples you'll have direct costs
[278]
paying the farmers transporting the
[280]
goods and you'll have the overheads
[281]
things like keeping the lights on
[283]
literally above your head
[285]
in that case the profit is a monetary
[287]
one and eventually a fusion plant would
[289]
have to make a dollar profit but for now
[291]
let's forget the costs and money and
[293]
only look at making a profit in terms of
[295]
electrical energy
[296]
proven technologies like coal power
[298]
plants do have energy overheads such as
[300]
the power needed to keep the lights on
[301]
or run the hand dryers in the toilets
[304]
then there are direct energy costs like
[306]
lifting the coal into the furnace the
[308]
electrical energy generated from each
[310]
lump of coal more than covers the energy
[313]
needed to mine it and move it the whole
[315]
plant easily overcomes the overheads and
[317]
makes a huge net gain in electricity for
[320]
any kind of fusion plant this is much
[322]
harder because both these types of
[323]
energy costs are enormous to run any
[326]
practical fusion plant we must pump
[328]
electricity into a number of hungry
[330]
systems things like powerful computer
[332]
control systems magnetic coils and so on
[335]
all fusion reactors would have to
[336]
maintain a very high level of vacuum for
[338]
instance which would not only require
[340]
turbine blades to be spun very rapidly
[342]
but also to be cooled to cryogenic
[344]
temperatures
[345]
there is also a direct energy cost for
[348]
every gram of fuel which must be heated
[350]
up which increases the worse the
[352]
confinement is
[354]
one big thing that fusion startups are
[356]
promising is to show a fusion energy
[358]
gain more energy out of your device than
[360]
you put in let me show you why this is
[362]
also not good enough with some very
[364]
rough numbers i will use an example of a
[366]
spherical tokamak which is the
[368]
configuration of two of the bold claims
[370]
i mentioned earlier suppose we capture
[372]
100 units of fusion energy from our
[374]
reactor as heat those pesky
[376]
thermodynamics from before mean i'd be
[378]
lucky to turn this into 50 units of
[380]
electrical energy 20 units go to the
[382]
overheads the magnetic coils pumps
[384]
control systems the other 30 are used
[387]
for heating and current drive in the
[389]
plasma
[390]
suppose we only use a technique called
[392]
electron cyclotron resonance heating
[394]
where energy is delivered to the
[395]
electrons in the plasma by microwaves
[397]
the microwaves will be generated by
[399]
giatrons at say 50 efficiency for a
[402]
total of 15 units of energy
[404]
the microwaves are transmitted through
[406]
wave guides into the reactor some are
[408]
lost on the way some are absorbed by the
[410]
walls
[411]
some are scattered or absorbed by edge
[414]
plasma that will never get hot enough to
[415]
undergo fusion reactions ultimately only
[418]
a third of the microwaves five units of
[421]
energy are absorbed in the core where
[422]
they will enable fusion reactions
[425]
so now we have to have a 20-fold gain on
[427]
our five units of energy to get back to
[429]
the 100 units of thermal energy we
[431]
started with realistically we need an
[433]
even higher gain still the highest
[435]
sustained energy output from a magnetic
[438]
device was 30 percent of the energy
[440]
input at the joint european taurus and
[442]
more recently at the national ignition
[444]
facility the fusion output slightly
[446]
exceeded the laser energy input
[448]
both of these are incredible feats which
[450]
took an international community many
[452]
years to achieve to now go up by several
[455]
orders of magnitude again in the next 9
[457]
years or even 19 years is at best overly
[460]
optimistic
[461]
by the way the problem with the nif is
[463]
that it does one shot every few hours
[466]
even if eventually electrical energy
[467]
could be generated in absolute terms
[470]
this is much less energy generation than
[472]
the environmentally friendly approach of
[474]
letting a single small goat graze on
[477]
grass and walk a little bit on a
[478]
treadmill
[480]
now you might think that i just pulled
[481]
some numbers out of my behind so let's
[483]
look at a peer-reviewed article from
[485]
2016 by a team at the princeton plasma
[488]
physics laboratory ppl has a proven
[491]
track record in fusion research and
[493]
experience running their own spherical
[495]
tokamak in their article they go over
[497]
all the aspects of a possible prototype
[500]
plant
[500]
in particular something they call the
[502]
engineering gain the true output of
[504]
usable electrical power for a given
[506]
input
[507]
the article painstakingly goes through
[509]
realistic best-case estimates for the
[511]
necessary power to all the relevant
[513]
systems like the pumps magnetic coils
[515]
and so on
[516]
you can check out this article in the
[518]
engineering gain formula but for now let
[520]
me explain one of the figures on the
[522]
x-axis is the major radius of the
[524]
spherical tokamak in meters a measure of
[527]
its size the larger it is the easier it
[530]
is proportionately to confine the plasma
[532]
and achieve gain the engineering gain is
[534]
on the left y-axis with three curves
[537]
corresponding to different values of the
[539]
thermal efficiency this is that
[540]
efficiency i mentioned of turning
[542]
thermal energy into electricity we are
[545]
shown curves for
[546]
0.59 or 59 in green 45 in red and 30 in
[551]
blue but then we have a fourth dashed
[553]
curve with a scale on the right y axis
[556]
which is the fusion gain with a range up
[558]
to 50. with the most optimistic 59
[561]
thermal efficiency to get above an
[563]
engineering gain of 1 and therefore be
[565]
making electricity we might choose a
[567]
major radius of 1.8 meters for our
[570]
tokamak this gets a comfortable
[572]
engineering gain with a corresponding
[573]
fusion gain of 19 at 45 thermal
[577]
efficiency we would need to choose a
[578]
major radius above 2.1 meters the
[581]
engineering gain is about the same as
[583]
before but the fusion gain is something
[584]
like 37 so just as in my example we need
[588]
a 20 30 or even 100 fold fusion gain
[591]
over the input energy to have a
[592]
meaningful gain in electricity it's not
[594]
enough to just show again to be making
[596]
electricity
[598]
if you thought things were complicated
[599]
so far you're not wrong this is a common
[601]
bamboozling tactic that's frequently
[603]
used in fusion let's suppose that we run
[606]
a very unethical privately funded
[608]
startup with our own spherical tokamak
[610]
we have achieved a fusion gain of two
[612]
over the five units of microwave energy
[614]
delivered to the plasma this is already
[616]
far better than what has been shown so
[618]
far but we find some roadblock wherein
[620]
we know that we won't be able to go any
[622]
higher being unscrupulous in our press
[625]
release we would completely neglect to
[627]
say that we began with 50 units of
[629]
electrical energy we would just say that
[631]
we took five units and turn them into
[633]
ten success
[635]
the press and the venture capitalist
[636]
would just see it as number go up
[639]
even if the venture capitalists know
[640]
that your reactor will never produce
[642]
electricity all they need to do is use
[644]
the hype to sell their shares for higher
[646]
than they bought them
[648]
i mentioned that the easiest fusion
[650]
reaction to make work and hence the one
[651]
that everyone eventually wants to use is
[653]
that between the two isotopes of
[655]
hydrogen deuterium and tritium deuterium
[657]
occurs in seawater but tritium is
[660]
radioactive and does not occur naturally
[662]
fortunately each deuterium tritium
[664]
reaction releases a neutron which can be
[666]
absorbed by lithium on the edge of the
[668]
reactor to create tritium through
[669]
another nuclear reaction one tritium
[672]
atom creates one neutron the neutron is
[674]
absorbed to create one tritium atom this
[677]
process will also have inefficiencies as
[679]
the tritium decays radioactively and
[681]
some neutrons will be lost
[683]
fortunately there is a reaction where a
[685]
neutron can strike a beryllium atom and
[687]
release a second neutron thereby getting
[689]
two tritium atoms for the price of one
[692]
so now your commercial fusion reactor
[693]
needs to not only break even on the
[695]
energy but also on the tritium
[697]
production or breeding plucking stupid
[699]
numbers out of my behind again we start
[701]
off with 100 atoms of tritium 5 decay or
[704]
are lost 50 fused to produce neutrons of
[707]
which 5 are also lost 45 neutrons must
[710]
now produce 55 tritium atoms to get back
[713]
to the 100 we started with a reactor
[715]
could run indefinitely with a so-called
[717]
tritium breeding ratio of about 1.15 to
[720]
1.3 this would be done by carefully
[723]
engineering a thick blanket filled with
[725]
beryllium and lithium all around the
[727]
reaction chamber all good on paper but
[730]
in practice none of the startups i
[731]
mentioned have even released plans let
[733]
alone work through and tested the
[735]
engineering of these kinds of blankets
[737]
academics including those at the uk
[739]
atomic energy authority are doing some
[741]
of the research and design but again
[743]
none of them have been fully built let
[745]
alone properly tested inside a working
[747]
fusion reactor
[748]
to actually absorb those neutrons coming
[750]
from your fusion reactions which by the
[752]
way are carrying the majority of the
[753]
energy you must physically have roughly
[755]
a meter of dense material around your
[757]
fusion reactor because the neutrons are
[759]
not charged they are not slowed down
[761]
unless they interact with a nucleus
[762]
directly you must therefore put enough
[765]
atoms in their way to stop them when a
[767]
neutron is absorbed by any element other
[769]
than lithium or beryllium it will
[771]
typically turn the stable nucleus into a
[773]
heavier radioactive one if the
[775]
phosphorus in your dna absorbs a neutron
[778]
it will turn into an unstable isotope
[780]
which will then decay by means of beta
[782]
emission and become a sulfur atom i hope
[784]
i don't need to say this but both of
[785]
these things are very bad for your dna
[788]
if you don't physically have enough
[790]
material around your reactor for example
[792]
if you claim that your reactor could fit
[793]
in a shipping container you've got two
[795]
big problems you are losing many of the
[797]
neutrons that are carrying your fusion
[799]
energy so it will be that much harder to
[802]
have a gain in energy and you're making
[804]
the air all around highly radioactive it
[807]
is possible to do other fusion reactions
[809]
but as long as it involves deuterium
[811]
which is most of the other practical
[813]
reactions neutrons will be given off and
[815]
must be stopped to avoid making things
[817]
radioactive the proton boron reaction is
[820]
the only a neutronic one but it is
[823]
hundreds of times harder to achieve so a
[825]
high fusion gain will be that much
[827]
harder
[828]
thinking that the issues of neutrons and
[830]
tritium breeding can be solved as an
[831]
afterthought is like saying hydrogen and
[834]
oxygen can power rocket now go build me
[836]
a mission to mars it's just not that
[838]
easy
[839]
one of the main proposals inside the
[840]
article from princeton was the creation
[842]
of a fusion nuclear science facility
[845]
this would allow the scientific and
[847]
technological challenges of working with
[848]
neutrons and breeding tritium to be
[850]
addressed
[851]
getting such a facility just to start
[853]
experiments by 2040 would be a challenge
[855]
however
[858]
i've been very gloomy so far but for
[860]
what it's worth i think there is real
[862]
progress being made by some groups the
[865]
international collaboration of eta is
[867]
really making strides it is taking the
[869]
slow but steady approach
[871]
being built now and ready by 2025 is a
[874]
conventional tokamak which is a tried
[876]
and tested concept scaled up with things
[878]
like superconducting magnets the plan is
[881]
to steadily work up to a fusion gain of
[883]
10 and show continuous uninterrupted
[885]
operation for up to an hour very
[887]
importantly it has several test beds for
[889]
lithium blankets to actually see what
[891]
they would behave like in a real reactor
[894]
on the other side and i must admit i'm
[896]
biased towards them are mit and their
[898]
spin-off commonwealth fusion systems
[900]
both have that kind of lean rough and
[902]
ready startup mentality but it's backed
[904]
with a wealth of experience and real
[906]
groundbreaking tech
[908]
they have demonstrated new
[909]
high-temperature superconductors which
[911]
could really help push their fusion gain
[913]
up and they are designing their reactor
[915]
with technologies like tritium breeding
[916]
blankets from the ground up also while
[919]
they probably will miss some of their
[920]
milestones their website doesn't have an
[923]
end date that i can poke fun at
[925]
thanks for listening the sources are in
[927]
the description
[928]
if there is enough interest i will make
[930]
a follow-up video about the technical
[931]
aspects of fusion such as why a fuser
[933]
can never generate electricity how
[935]
different reactor concepts work and
[937]
details of the main challenges which i
[939]
didn't have time to mention
Most Recent Videos:
You can go back to the homepage right here: Homepage