4 ms·

I enjoy those sorts of stories, people doing research into something they find interesting opening up a field for others. If this material (hBN) could be made

I enjoy those sorts of stories, people doing research into something they find interesting opening up a field for others. If this material (hBN) could be made industrially at scale I believe it would completely change the capacitor market. Based on papers that describe the amount of charge (current) graphene can carry, especially when isolated with hBN, I imagine that a capacitor made with an hBN dielectric would completely blow away its closest competitor both in capacitance per cubic mm and in low ESR. Think dots the size of periods replacing components that are now the size of rice grains.

mikepurvis 7y ago
Asking from a place of total ignorance, but would this impact integrated circuit capacitors as well, or only discretes?
- marcosdumay 7y ago
  Usually, it's hard to mix materials in an integrated circuit. Adding graphene to silicon is the kind of thing that looks a lot like it can be done. But it wouldn't be cheap, so I doubt people will ever do that just for capacitors. (But I imagine everybody would jump at it if it means better transistors.)
  - mikepurvis 7y ago
    Makes sense, thanks for the response. IC resistors are implemented as switched capacitors, but I suppose that's only in amplifiers or other analogue ICs where such a thing is relevant, and there the volume is probably not significant enough to benefit from the size-down? Anyway, still very significant to get smaller and more efficient discrete capacitors.
ksec 7y ago
>Think dots the size of periods replacing components that are now the size of rice grains. That is possibly 100x difference in size? So if this was 1000x the cost per cubic mm, it would only be 10x more expensive per unit of same Capacitance.
- ChuckMcM 7y ago
  That would be one way to look at it, or 100x the amount of capacitance in the same volume. In the radio stuff I'm currently doing there are challenges with volume required by the current capacitor technology.
lightedman 7y ago
"If this material (hBN) could be made industrially at scale" It exists already as an abrasive, cubic and hexagonal crystal forms. It is what I use when I don't feel like using diamond for stone grinding and polishing.
- ChuckMcM 7y ago
  Cool, I didn't realize it was a commercial abrasive. It seems that as an abrasive it is shipped as a powder. The trick then would be to create sheets of it at workable sizes. The original article describes how the authors do that (growing larger crystals) that but they are using a pretty time intensive process.
DoctorOetker 7y ago
Regarding "made industrially at scale": I am visualizing a scaled up pizza knife (pizza wheel, ...), where the teeth are essentially the pointed weight in the picture in the article ;) Regarding capacitors from hBN? Where do you get the idea that hBN has a remarkably high dielectric constant? I couldn't find any such discussion in the article... The energy density for a dielectric material is epsilon_material E^2. The other comments blindly assume that what you say is true and wonder about the impacts? What I like about HN are the important subjects being discussed, what I dislike is how it shines through that the mostly software oriented audience became software oriented in the past because of some phenomenon I suspect but do not know the name for: I might call it natural science escapism, or perhaps rammification escapism. I don't know which are the major causes but I suspect its a combination of: 1) Software really is a poor man's science: many people have access to computers, but most don't have access to expensive tools like oscilloscopes, and expensive experience like rigorous physics education, ... To the extent this is the cause I certainly don't blame people for having been or being poor! I also believe many people as kids viewed programming as their future escape from poverty. While this may have been a valid view in the past I no longer believe it is. (when computers were still very expensive and people somehow found opportunity or access to one; or a little later during the transition from expensive to cheap computers, where most people did not have the foresight yet that programming would become an essential but common skill, like reading and writing, and those that did became the last wave of "whiz kids", since nowadays most disciplines can directly benefit from and assume basic skills in programming, essential programming skills are just not that special any more, but they are still essential, like reading and writing). 2) Another possibility is that programming is much closer to mathematics than it is to science: mathematics and programming are conceptually closer to deductive reasoning, while the natural sciences like physics and chemistry are conceptually closer to abductive reasoning. I say closer because mathematics also involves some abductive reasoning (think of Russel's paradox prompting a reformulation of mathematics or logic), but these are relatively rare events. The natural sciences involve the scientific method: a cycle of deductive and abductive reasoning (where the old model is used to deductively predict an outcome, which is used to notice a dedected anomaly, and the anomaly is used to abductively propose different modifications of the models postulates, which then ideally allow new deductive predictions to be made and experimentally verified). So fundamental sciences involve roughly 50% deductive and 50% abductive reasoning. A good universitary physics programme essentially exposes the students to a constant series of false bottoms where every historical discovery sweeps the rug from under the students' feet. This can subjectively be interpreted as painful or stimulating, those kids who find such learning (say at high school level) to be too painful tend to prefer programming or mathematics over the natural sciences, a form of escapism from abductive reasoning to deductive reasoning. Think "just give me the correct axioms / requirements / postulates ... first, so I never have to change them again!", if you identify with this you will prefer programming or math above natural sciences programmes aimed at fundamental research. I really wish there was some kind of second "hacker news" for the more natural science inclined people, who are not interested in yet another web programming library (I would frequent both depending on my mood). In point 2) when I distinguish between fields of study with more or less abductive reasoning, I do not mean abductive reasoning in the mind of the practicioner, but abductive reasoning on the shared model or field of study. Indeed every job or task a human performs entails roughly 50% deductive and 50% abductive reasoning, but the abductive reasoning is limited to refining the individual model or understanding of a problem, as opposed to refining the collective field of study or understanding of a model or problem. For example when a programmer is bugfixing his code, the programmer is not actually improving computer science, but rather his own misunderstanding of what the code actually does as opposed to what the programmer mistakenly thought he instructed the computer to do. When Einstein for example resolved the constant speed of light in different inertial frames by formulating relativity theory, he did not merely improve his own understanding of the universe, but our collective one!

mikepurvis 7y ago

Asking from a place of total ignorance, but would this impact integrated circuit capacitors as well, or only discretes?

marcosdumay 7y ago
Usually, it's hard to mix materials in an integrated circuit. Adding graphene to silicon is the kind of thing that looks a lot like it can be done. But it wouldn't be cheap, so I doubt people will ever do that just for capacitors. (But I imagine everybody would jump at it if it means better transistors.)
- mikepurvis 7y ago
  Makes sense, thanks for the response. IC resistors are implemented as switched capacitors, but I suppose that's only in amplifiers or other analogue ICs where such a thing is relevant, and there the volume is probably not significant enough to benefit from the size-down? Anyway, still very significant to get smaller and more efficient discrete capacitors.

ksec 7y ago

>Think dots the size of periods replacing components that are now the size of rice grains. That is possibly 100x difference in size? So if this was 1000x the cost per cubic mm, it would only be 10x more expensive per unit of same Capacitance.

ChuckMcM 7y ago
That would be one way to look at it, or 100x the amount of capacitance in the same volume. In the radio stuff I'm currently doing there are challenges with volume required by the current capacitor technology.

lightedman 7y ago

"If this material (hBN) could be made industrially at scale" It exists already as an abrasive, cubic and hexagonal crystal forms. It is what I use when I don't feel like using diamond for stone grinding and polishing.

ChuckMcM 7y ago
Cool, I didn't realize it was a commercial abrasive. It seems that as an abrasive it is shipped as a powder. The trick then would be to create sheets of it at workable sizes. The original article describes how the authors do that (growing larger crystals) that but they are using a pretty time intensive process.

DoctorOetker 7y ago

Regarding "made industrially at scale": I am visualizing a scaled up pizza knife (pizza wheel, ...), where the teeth are essentially the pointed weight in the picture in the article ;) Regarding capacitors from hBN? Where do you get the idea that hBN has a remarkably high dielectric constant? I couldn't find any such discussion in the article... The energy density for a dielectric material is epsilon_material E^2. The other comments blindly assume that what you say is true and wonder about the impacts? What I like about HN are the important subjects being discussed, what I dislike is how it shines through that the mostly software oriented audience became software oriented in the past because of some phenomenon I suspect but do not know the name for: I might call it natural science escapism, or perhaps rammification escapism. I don't know which are the major causes but I suspect its a combination of: 1) Software really is a poor man's science: many people have access to computers, but most don't have access to expensive tools like oscilloscopes, and expensive experience like rigorous physics education, ... To the extent this is the cause I certainly don't blame people for having been or being poor! I also believe many people as kids viewed programming as their future escape from poverty. While this may have been a valid view in the past I no longer believe it is. (when computers were still very expensive and people somehow found opportunity or access to one; or a little later during the transition from expensive to cheap computers, where most people did not have the foresight yet that programming would become an essential but common skill, like reading and writing, and those that did became the last wave of "whiz kids", since nowadays most disciplines can directly benefit from and assume basic skills in programming, essential programming skills are just not that special any more, but they are still essential, like reading and writing). 2) Another possibility is that programming is much closer to mathematics than it is to science: mathematics and programming are conceptually closer to deductive reasoning, while the natural sciences like physics and chemistry are conceptually closer to abductive reasoning. I say closer because mathematics also involves some abductive reasoning (think of Russel's paradox prompting a reformulation of mathematics or logic), but these are relatively rare events. The natural sciences involve the scientific method: a cycle of deductive and abductive reasoning (where the old model is used to deductively predict an outcome, which is used to notice a dedected anomaly, and the anomaly is used to abductively propose different modifications of the models postulates, which then ideally allow new deductive predictions to be made and experimentally verified). So fundamental sciences involve roughly 50% deductive and 50% abductive reasoning. A good universitary physics programme essentially exposes the students to a constant series of false bottoms where every historical discovery sweeps the rug from under the students' feet. This can subjectively be interpreted as painful or stimulating, those kids who find such learning (say at high school level) to be too painful tend to prefer programming or mathematics over the natural sciences, a form of escapism from abductive reasoning to deductive reasoning. Think "just give me the correct axioms / requirements / postulates ... first, so I never have to change them again!", if you identify with this you will prefer programming or math above natural sciences programmes aimed at fundamental research. I really wish there was some kind of second "hacker news" for the more natural science inclined people, who are not interested in yet another web programming library (I would frequent both depending on my mood). In point 2) when I distinguish between fields of study with more or less abductive reasoning, I do not mean abductive reasoning in the mind of the practicioner, but abductive reasoning on the shared model or field of study. Indeed every job or task a human performs entails roughly 50% deductive and 50% abductive reasoning, but the abductive reasoning is limited to refining the individual model or understanding of a problem, as opposed to refining the collective field of study or understanding of a model or problem. For example when a programmer is bugfixing his code, the programmer is not actually improving computer science, but rather his own misunderstanding of what the code actually does as opposed to what the programmer mistakenly thought he instructed the computer to do. When Einstein for example resolved the constant speed of light in different inertial frames by formulating relativity theory, he did not merely improve his own understanding of the universe, but our collective one!