4 ms·

I'm very much conservative about this paper because while the graph in Fig. 3a says $T_c^{zero}$ being 110 K, the "zero" resistance is actually only 1e-5 Ω (!)

I'm very much conservative about this paper because while the graph in Fig. 3a says $T_c^{zero}$ being 110 K, the "zero" resistance is actually only 1e-5 Ω (!) and even if the sample is possibly superconducting its critical temperature would be much lower than 110 K anyway. I'm aware of the difficulty in obtaining larger samples, but the label in that graph is really misleading regardless.

nukros 3y ago
The random jumps in the resistivity around 250 K suggests that there is very bad contact resistance or a possible alternate current path. The drop in resistance doesn’t look like a typical phase transition (it is very broad in temperature). My guess would be that the voltage leads are not very well connected to the region of the sample that the current is flowing through. This could actually give a drop in the measured voltage (and therefore “resistance”) even though the actual resistance is increasing as the sample is cooled. It would be interesting if they also included two-point resistance across the current leads.
- segfaultbuserr 3y ago
  One of the authors commented in their previous video pre-announcement that they suspect it was a probe cleanliness problem, which they would investigate later.
  - throwawaylinux 3y ago
    Is this kind of thing typical for experiments working to confirm such significant results, or is there a mad rush to publish? Do you really get a lot of prestige from confirming somebody else's result? I would have thought the risk to reputation from sloppy work would far outweigh it (see: the cold fusion shenanigans).
    - ianai 3y ago
      Probably a lab that confirms any piece of it will get a significant share of attention and acclaim. I heard the research group got scooped by a former researcher. Former researcher outed them before they were ready for all the relevant questions. Not a particularly good thing to do for a career to out colleagues like that.
      - throwawaylinux 3y ago
        > Probably a lab that confirms any piece of it will get a significant share of attention and acclaim. Why is that though? They're taking a discovery that somebody else came up with, and (mostly) follow a recipe they're given. I mean if they were synthesizing a theorized substance or significantly improving its production, or measuring an effect that was previously undetectable by known instruments and experiments then those things would deserve acclaim on their own. Not saying they aren't good scientists and labs working on this or reproducing is worthless, the reward just doesn't seem big enough not to be meticulous about it. If you're right then you'll be one of the dozens of labs that reproduced it and all credit goes to original discoverers. If you're wrong you'll be the ones who bungled the experiment and share just about equal blame.
        jacquesm 3y ago
        In science the fights are so fierce because the stakes are so small...
        CamperBob2 3y ago
        Typically a valid comment in academia at large, but not when high-temperature superconductors are the subject of the fight. If these claims pan out, these guys may be the next Andy Viterbi or Ray Stata.
        soerxpso 3y ago
        It demonstrates efficiency and ability to do something complicated and challenging under time pressure. Maybe it shouldn't be considered a big academic achievement, but it's definitely an impressive practical achievement. I don't think that being wrong looks bad. There are a lot of understandable ways to be wrong, and it's not like any of these labs are being dishonest about their level of confidence (you can't really be "wrong" if you're not overcommitted to a stance).
junofan 3y ago
Spec sheet says 10 uohm is as low as their probe goes.
pythonguython 3y ago
I believe certain it is near impossible to actually measure truly zero resistance. Instead, a sudden drop near Tc, combined with the meissner effect is what confirms superconductivity.
segfaultbuserr 3y ago
The resistance curve stopped at 1e-5 Ω because it has nearly reached the resolution limit of their instrument. Using 4-wire Kelvin sensing, the measurement current was 1 mA, the voltage read-out across the material was around 1 to 10 nanovolts - this is around the scale that some of the world's best general-purpose voltmeters can measure down to. The authors assumed that the negligible voltage drop, combined with the abrupt drop and flattening of the resistance vs temperature curve is already a convincing-enough indication of superconductivity. Hopefully more rigorous testing will be done by one of the teams later.
- lifthrasiir 3y ago
  Which is indeed their argument. But in that case they haven't proved that $T_c^{zero}$ = 110 K, they only proved that $T_c^{zero}$ < 110 K. (And I should also note that $T_c^{zero}$ doesn't imply the actual critical temperature $T_c$, while the title strongly suggests so.)
- amelius 3y ago
  Curious, why didn't they use a larger current for the measurement?
  - segfaultbuserr 3y ago
    Who knows... Too afraid to damage their only sample?
kurthr 3y ago
Well, it's at a 1mA test current (I'm assuming a 4 point probe to mitigate contact resistance) so 10u Ohm would give a 10nV DC signal. That's a pretty small voltage signal and comparable to a fairly standard chopper stabilized instrumentation amp rms noise at room temp near DC. I'm not sure whether higher currents are all that reasonable. Cooling the precision amp is an option to reduce the noise figure, but at that point you'll want optical isolation and battery supply too. https://www.analog.com/media/en/training-seminars/tutorials/MT-055.pdf https://www.analog.com/media/en/training-seminars/tutorials/...
- guraf 3y ago
  [flagged]
  - kurthr 3y ago
    Yeah, I ran low temp STM/BEEM systems in UHV for 4 years at JPL's MDL. I assume you know what that means. Even today low frequency nV signals are still hard. Did you know you can do zero-bias tunneling sensing?
    - robbiep 3y ago
      Glorious to watch
- throwaway849755 3y ago
  A comment above says they are using https://www.qdusa.com/products/ppms.html https://www.qdusa.com/products/ppms.html with the DC 4-wire option. All my instincts would prefer they were using the AC 4-wire option, given any reason to question results, and/or issues with contact effects, etc.
  - kurthr 3y ago
    Thanks! That's great information. We had to make our own, but it's good to see that there are decent off the shelf systems. I agree that contact resistance on these tiny samples has got to be a huge issue (can they use mercury?). However, what I've seen suggests that even the low frequency AC measurements are more complex due to capacitive coupling effects. However, narrow bandwidth lock=in should improve noise... the equipment still just claims 10u Ohm.

nukros 3y ago

The random jumps in the resistivity around 250 K suggests that there is very bad contact resistance or a possible alternate current path. The drop in resistance doesn’t look like a typical phase transition (it is very broad in temperature). My guess would be that the voltage leads are not very well connected to the region of the sample that the current is flowing through. This could actually give a drop in the measured voltage (and therefore “resistance”) even though the actual resistance is increasing as the sample is cooled. It would be interesting if they also included two-point resistance across the current leads.

segfaultbuserr 3y ago
One of the authors commented in their previous video pre-announcement that they suspect it was a probe cleanliness problem, which they would investigate later.
- throwawaylinux 3y ago
  Is this kind of thing typical for experiments working to confirm such significant results, or is there a mad rush to publish? Do you really get a lot of prestige from confirming somebody else's result? I would have thought the risk to reputation from sloppy work would far outweigh it (see: the cold fusion shenanigans).
  - ianai 3y ago
    Probably a lab that confirms any piece of it will get a significant share of attention and acclaim. I heard the research group got scooped by a former researcher. Former researcher outed them before they were ready for all the relevant questions. Not a particularly good thing to do for a career to out colleagues like that.
    - throwawaylinux 3y ago
      > Probably a lab that confirms any piece of it will get a significant share of attention and acclaim. Why is that though? They're taking a discovery that somebody else came up with, and (mostly) follow a recipe they're given. I mean if they were synthesizing a theorized substance or significantly improving its production, or measuring an effect that was previously undetectable by known instruments and experiments then those things would deserve acclaim on their own. Not saying they aren't good scientists and labs working on this or reproducing is worthless, the reward just doesn't seem big enough not to be meticulous about it. If you're right then you'll be one of the dozens of labs that reproduced it and all credit goes to original discoverers. If you're wrong you'll be the ones who bungled the experiment and share just about equal blame.
      - jacquesm 3y ago
        In science the fights are so fierce because the stakes are so small...
        CamperBob2 3y ago
        Typically a valid comment in academia at large, but not when high-temperature superconductors are the subject of the fight. If these claims pan out, these guys may be the next Andy Viterbi or Ray Stata.
      - soerxpso 3y ago
        It demonstrates efficiency and ability to do something complicated and challenging under time pressure. Maybe it shouldn't be considered a big academic achievement, but it's definitely an impressive practical achievement. I don't think that being wrong looks bad. There are a lot of understandable ways to be wrong, and it's not like any of these labs are being dishonest about their level of confidence (you can't really be "wrong" if you're not overcommitted to a stance).

junofan 3y ago

Spec sheet says 10 uohm is as low as their probe goes.

pythonguython 3y ago

I believe certain it is near impossible to actually measure truly zero resistance. Instead, a sudden drop near Tc, combined with the meissner effect is what confirms superconductivity.

segfaultbuserr 3y ago

The resistance curve stopped at 1e-5 Ω because it has nearly reached the resolution limit of their instrument. Using 4-wire Kelvin sensing, the measurement current was 1 mA, the voltage read-out across the material was around 1 to 10 nanovolts - this is around the scale that some of the world's best general-purpose voltmeters can measure down to. The authors assumed that the negligible voltage drop, combined with the abrupt drop and flattening of the resistance vs temperature curve is already a convincing-enough indication of superconductivity. Hopefully more rigorous testing will be done by one of the teams later.

lifthrasiir 3y ago
Which is indeed their argument. But in that case they haven't proved that $T_c^{zero}$ = 110 K, they only proved that $T_c^{zero}$ < 110 K. (And I should also note that $T_c^{zero}$ doesn't imply the actual critical temperature $T_c$, while the title strongly suggests so.)
amelius 3y ago
Curious, why didn't they use a larger current for the measurement?
- segfaultbuserr 3y ago
  Who knows... Too afraid to damage their only sample?

kurthr 3y ago

Well, it's at a 1mA test current (I'm assuming a 4 point probe to mitigate contact resistance) so 10u Ohm would give a 10nV DC signal. That's a pretty small voltage signal and comparable to a fairly standard chopper stabilized instrumentation amp rms noise at room temp near DC. I'm not sure whether higher currents are all that reasonable. Cooling the precision amp is an option to reduce the noise figure, but at that point you'll want optical isolation and battery supply too. https://www.analog.com/media/en/training-seminars/tutorials/MT-055.pdf https://www.analog.com/media/en/training-seminars/tutorials/...

guraf 3y ago
[flagged]
- kurthr 3y ago
  Yeah, I ran low temp STM/BEEM systems in UHV for 4 years at JPL's MDL. I assume you know what that means. Even today low frequency nV signals are still hard. Did you know you can do zero-bias tunneling sensing?
  - robbiep 3y ago
    Glorious to watch
throwaway849755 3y ago
A comment above says they are using https://www.qdusa.com/products/ppms.html https://www.qdusa.com/products/ppms.html with the DC 4-wire option. All my instincts would prefer they were using the AC 4-wire option, given any reason to question results, and/or issues with contact effects, etc.
- kurthr 3y ago
  Thanks! That's great information. We had to make our own, but it's good to see that there are decent off the shelf systems. I agree that contact resistance on these tiny samples has got to be a huge issue (can they use mercury?). However, what I've seen suggests that even the low frequency AC measurements are more complex due to capacitive coupling effects. However, narrow bandwidth lock=in should improve noise... the equipment still just claims 10u Ohm.