News, Reviews & Discussion of EEStor Inc. & Quantumscape Corporation
What We Have Learned – The Topic.
This discussion is a continuation of the blog post of the same name. When I posted the first post on that thread I honestly believed that all the members of the forum (who have technical back grounds) understood and accepted the “brick model”. When I found that was not true I attempted to explain the model more clearly. I am afraid that I failed in that, but one thing led to another and I have kept on posting. Essentially I am reviewing points that were raised on the old site and which I think are pertinent to “How the EESU Works”. Just a few more posts and I will be done.
The usual rules apply: Please keep replies short, to the point and polite. ee-tom is banned.
I had started to review the possible polarisation mechanisms (for the EESU) so I am going to repost the two posts concerning that, then I will continue.
Regards,
Peter
Tags:
Dielectric Absorption:
This is the name given to the behaviour of a dielectric when space charge is present.
I Wrote:
If the region extended from plate to plate then it would be leakage current, not space charge.
There is a time constant associated with space charge and unfortunately, accurate models are not available for cases when the effect is significant. Times of significant fractions of hours to hours are reported.
An accurate measurement of leakage current cannot be made until the space charge has stabilised.
It is quite possible that the leakage current for a particular measurement is actually the space charge current.
I recall that rather than wait for this kind of measurement to stabilise I have said (quite honestly): “It is less than 1 micro amp”. A computer controlled measurement system might wait say 2 minutes and then do the measurement, which could be totally erroneous, nevertheless being the correct value of the current flowing at that time.
I have no information regarding dielectric absorption in the EESU. I just think it should be considered.
Regards,
Peter
Beware the Killer Flux:
Charge, Field and Flux.
The presence of charge results in field and that is the only manifestation. The “point” in “point charge” locates the field, whether it wobbles or not.
Because the field reduces as the inverse of distance from the “point” then the notion of “flux” is convenient. In matter this can only be an approximation, the so called “macroscopic approximation”.
Total Flux is actually field integrated over an enclosing area and this integral is equal to the sum of the charges inside divided by e0, this is the real Gauss’s Law, beware of imitations.
The real situation is properly described using quantum mechanics. However, the concept of local field is useful (more tractable numerically) in many cases.
To determine the local field of a polarised dielectric the following steps can be used:
1 Freeze all the charges in position.
2 Remove the molecule/unit cell/dipole of interest.
3 Calculate the field due to the polarisation of the dielectric at this point; let us say it is Ed.
4 Subtract this from the applied field and the sum of the field due to the charge on the plates. The result is El, the local field
El = Ea + P/e0 – Ed…………………………(P/e0 is the “killer flux”, P is plate polarisation).
If El is greater than the field the molecule/unit cell/dipole can stand then this fact has to be taken into account in a rational and scientific manner. If not, all is okey dokey.
Comment:
1) If the polarisation of all molecules is uniform then the Clausius-Mossotti result is obtained. For the case where the layers are not polarised uniformly I have written a program to calculate the field. I used this to calculate the field at the alumina/CMBT interface and posted it a few posts earlier on this thread.
2) Because of symmetry, when the removed molecule is replaced it's field will cancel when calculating the macroscopic field at that location.
The idea of flux is really only useful (in a so called “dielectric”) when the material is “homogenous and isotropic”. That is a phrase used by Dan; I quote it because he is usually right.
Regards,
Peter
Pn, Ev, and Af are looking at perfect crystals for maximum efficiency. Perfect crystals usually are done slowly under specific temperatures and under special conditions(pressure, composition). Are we someday looking at giant Weir crystals powering spaceships that depend on stress or breakage to shift into hyperdrive? Or maybe stressed crystals to release 'Death Rays'. I for one favor the hyperdrive, but LM may be looking at protecting our nation, as well as a profit.
DW has speeded up the crystal formation, and the Moore's Law pertaining to perfect crystal formation likely has begun. Will we someday see that giant spaceship crystal power one big alien fan that takes us to the distant planets. I won't be around, but it would be fun to see.......maybe even take a trip.
I refer to DW in my statement as Carl Nelson stands in the background. But, WOW does he stand TALL. May I be excused for leaving him out. My hat is off to both of these men for what I hope they are about to become....Nobel Prize winners, and great contributors to the advancement of our society.
Hello DowntoEEarth,
I have not seen any explanations by PN, Eev or AF for the polarisation achieved in the EESU so I can’t comment on their ideas.
Particles which are essentially single crystals do seemed to be called for though.
The promise that EESU development offers is truly mind boggling.
For example by reducing the number of recharge cycles to say, one hundred, the energy density could be increased by an unbelievable amount.
Hand held lasers of lethal power would be no problem. All American school teachers could be issued with these as a protection against nutters. This is not a solution I advocate but (unfortunately) I could see it happening.
Regards,
Peter
The blue line is measured, the red line is calculated. The purpose is to show that DF in ordinary dielectrics is “well understood”. In particular the DF is constant at low frequencies. So it is conceivable that the EESU DF measured at 100 Hz is much the same as the figure at 1 kHz.
By the way, if you look closely at the blue line (measured) you will see it has been drawn using a ruler and a French Curve.
Honestly, the only way to get good data these days is to measure it yourself!
Regards,
Peter
Hmm, still that's at AC. At DC the story is more complicated.
Hello JCatania,
Yes, but DF is specified at 1 kHz. The puzzle has many pieces, this may be one of them.
Regards,
Peter
The DF of Layer A is 0.24 at an implied 100Hz (assumes same frequency as for capacitance) , which is well above top end of the DF range of PeterP's graph above, though the frequency is well below the frequency range of the graph. Above 100KHz on the above graph the DF appears to be proportional to the frequency.
Layer D's DF is 4.5% at implied 100Hz. Same applies, though not quite so much.
The question then is whether the DF of EEStor's Layer A and Layer D vary linearly with frequency (going with the DF range from the above graph) or are almost independent of frequency (going with the frequency range). My money would be on linear dependency with frequency, but only because it would make all the calculations easier (i.e. constant ESR). In practice it is 50:50.
If the DF has indeed been measured at 1Khz (unstated) rather than 100Hz (assumed same as capacitance) then this makes little difference to the comments above, though if it is in the regime where ESR is constant rather than DF constant it makes a huge difference (factor of 10 or 20) to the expected DF at grid frequencies.
Regards,
TP
Hello Technopete,
You Wrote:
My money would be on linear dependency with frequency, but only because it would make all the calculations easier (i.e. constant ESR). In practice it is 50:50.
This is extremely unlikely. Nature does not care about your computational capability. Nor is gambling appropriate (Sorry, I forgot you are a Quantum Mechanic, it’s all rolling the dice for you guys. Go ahead and gamble. (J/K)).
Constant esr would most likely be due to the plate DC resistance. Make an assumption about thickness, material and work it out.
I have explained earlier what causes the low frequency DF to be constant in “conventional capacitors”. You do need to know the physics of the EESU before you place your bet.
Regards,
Peter
The Killer Flux Fallacy:
The argument:
Put 60 C/m^2 on the plates of a capacitor then a “gigantic field results”.
This is true. But charge is not “put” on the plates of a capacitor. The capacitor “charges up”.
The correct approach:
In order that the potential difference between the plates of a capacitor is 3500 volts when charged to 60 C/m^2 there must be a reverse field to cancel the “gigantic field”.
To generate the correct field the dielectric must be polarised to 60 C.m/m^3.
In order to evaluate the possibility of this, the charging process must be analysed in detail starting at zero volts with zero charge on the capacitor.
Starting with full charge is bad logic and bad physics.
It is excellent misdirection though.
Regards,
Peter
Ferroelectric Behaviour:
Whether a material is ferroelectric or not depends to a large extent on the geometry. For example it is well known that a spherical particle of BT will not be ferroelectric if its diameter is less than 680 nm. The text book capacitor has a large area and small thickness. An homogenous dielectric in this capacitor can easily be analysed assuming uniform polarisation (a good assumption).
If α is the polarisability of a molecule and there are N molecules/cubic metre then we can write:
β = N. α
Where β is a material property and is dimensionless.
A little algebra reveals that if k is the apparent permittivity:
β = 3.(k-1)/(k+2)
As k goes from 1 to infinity then β goes from 0 to 3.
However the molecule has no such restraint, β can be greater than 3. In this case the dielectric is ferroelectric, it polarises itself due to the feedback effect.
So to get high k you just find a ferroelectric material and do something to reduce β.
Pressure will do it, so will temperature, changing geometry will do it too. Of course dopants can be introduced to reduce β too.
So that is what Eestor do to get high k. The only problem is the control. Of course you could just make a large number of samples and sort them to get the k you want. Then stability becomes an issue. Apparently Eestor have solved that problem too. Remarkable.
The relationship given above is for a thin flat disk, the relationship for a sphere is different.
Regards,
Peter
PeterP said:
The Killer Flux Fallacy:
The argument:
Put 60 C/m^2 on the plates of a capacitor then a “gigantic field results”.
This is true. But charge is not “put” on the plates of a capacitor. The capacitor “charges up”.
The correct approach:
In order that the potential difference between the plates of a capacitor is 3500 volts when charged to 60 C/m^2 there must be a reverse field to cancel the “gigantic field”.
To generate the correct field the dielectric must be polarised to 60 C.m/m^3.
In order to evaluate the possibility of this, the charging process must be analysed in detail starting at zero volts with zero charge on the capacitor.
Starting with full charge is bad logic and bad physics.
It is excellent misdirection though.
Regards,
Peter
OK. So you've started with zero volts and zero charge, as you suggest. How does the slow rise from zero to 60C/m^2 work?
And what do you end up with when you slowly and methodically get up to 60C/m^2?
Regards,
TP
© 2014 Created by B.