Microsoft Bets That Fusion Power Is Closer Than Many Think

Re:Microsoft Bets That Fusion Power Is Closer Than Many Think by DennisP on Wednesday May 10, 2023 @01:08PM (#35890234)

Their combined D-D/D-He3 reaction will emit 6% of its energy as neutron radiation, compared to 80% for D-T. Mostly they'll be lower-energy neutrons, and it's only a 50MW reactor. They don't plan to make it bigger, just to make lots of them.

They're considering doing the D-D reaction in a separate reactor to produce the He3. The D-He3 reaction is purely aneutronic, and while some D-D will still happen, they can tune it so it doesn't happen much. That would mean very little neutron radiation at the power plants.

[ Parent | on HN ]

Re:Microsoft Bets That Fusion Power Is Closer Than Many Think by schiffern on Wednesday May 10, 2023 @01:31PM (#35890578)

>they can tune it so [D-D] doesn't happen much

Can someone run through this, briefly?

I tried to reproduce this from First Principles, but I can't really make the numbers close.

Let's generously assume perfect tritium scavenging, so the relevant tuning parameters are temperature[0] and D-to-He3 ratio. However just using those two parameters, I can't make aneutronic fusion without starving the system of deuterium and plunging the power density, which (by the reciprocal relation) explodes the reactor size (and cost) for a given power output.

Can anyone square this circle?

[0] https://commons.wikimedia.org/wiki/File:RxDT-DD-DHe3.jpg

[ Parent | on HN ]

Re:Microsoft Bets That Fusion Power Is Closer Than Many Think by htag on Wednesday May 10, 2023 @01:29PM (#35890546)

> The D-He3 reaction is purely aneutronic, and while some D-D will still happen, they can tune it so it doesn't happen much. That would mean very little neutron radiation at the power plants.

Yea. I'm just very curious as to what these numbers look like in practice. Neutrons are essentially waste emissions with their plan, but physics and engineering constrains will determine how much they can tune out D-D reactions.

[ Parent | on HN ]

Re:Microsoft Bets That Fusion Power Is Closer Than Many Think by schiffern on Wednesday May 10, 2023 @04:49PM (#35893302)

> very curious as to what these numbers look like in practice

At their stated "optimum" temperature of 200 million kelvin[0] or ~17 keV, the selectivity of the two cross-sections (D-H3/D-D) is about 0.7[1].

Let x = the fraction of D (we'll generously assume no T contamination, so the remainder 1-x is He3), plug in the reaction energies[2], and rearrange.

Graph is here: https://www.wolframalpha.com/input?i=plot+y+%3D+1.225+x%5E2%...

The top line is total non-neutron power, and the bottom line is neutron power. The horizontal axis is the fraction of D in the mix, and the vertical axis is proportional to power.

It might help compare the relative tradeoffs if you zoom in on the rising slope of the power curve, and normalize both curves to 1: https://www.wolframalpha.com/input?i=plot%3A+y%3D1%2F3.9503+...

[0] https://www.reuters.com/technology/microsoft-buy-power-nucle...

[1] https://commons.wikimedia.org/wiki/File:RxDT-DD-DHe3.jpg

[2] https://en.wikipedia.org/wiki/Nuclear_fusion#Criteria_and_ca...

[ Parent | on HN ]