A Solid-State Maxwell Demon

[ZL]

We will get to the paper of Germano soon Nonlinear. All this preliminary discussion serves the purpose to ease the understanding of Germano's fallacies.

Continuing from my last post, let's take a look now at the two answers on the Quora page that are at least principally correct. One of those that was given first in 2012 by Abde Ali Kagalwalla, PhD Student, UCLA says:

... this potential cannot be measured directly by connecting a voltmeter across the diode. The reason for this is that as soon a connection is made between the diode terminal and a conductor (metal), a Schottky diode is created at the metal-p/n junction. The two Schottky diodes, created at the p-terminal metal interface and n-terminal metal interface effectively reverse the effect of the built-in potential. As a result, you would see zero potential difference across the diode.

I suppose that by "diode terminals" he means the two ends of the diode semiconductor, which are outside the depletion region (and not the metal legs that are attached to these surfaces). This is only principally correct, because the metal terminals that are attached to the semiconductors of most commercial diodes form ohmic contacts with the semiconductors, not schottky diodes.

Next to the junction surface in the semiconductor of a schottky diode there is a depletion region, similar to the one in a common diode. The width of this region (and the built-in potential) is large enough to block the current flow in the reverse direction (with small leakage), and thus act as a diode and rectify current. If the doping concentration is further increased, then the depletion region is getting narrower, the built in potential smaller, and the reverse leakage current larger, until it becomes useless as a diode. By proper doping techniques ohmic contacts are formed at the terminals of common p-n diodes.

However, the answer is principally correct in the sense that even the ohmic contacts between metal and semiconductor have a contact potential, similar to those between two different metals with different work functions. These two contact potentials between the semiconductor and metals coupled in series form a potential difference that exactly opposes the voltage across the p-n depletion region, therefore no net current will be driven by these voltages. This is why we can't measure the built-in voltage by simply making ohmic contacts with the instrument probes when touching the semiconductor surface. But this doesn't mean that we can't measure the built in voltage by other methods, and prove their existence.

This case is analogous to the situation when you want to measure the contact potential difference of a thermocouple without temperature gradient on it. There is a voltage across its terminals, but you can't measure it with a simple voltmeter, and it cant drive a current in a closed loop if you short-circuit its terminals. The reason for this is that when you connect metals to the thermocouple additional contact potential differences will be created at those contacts where two different metals touch. The total sum of all such potential differences in a closed circuit will be zero, they will not drive a net current, and thus a contact potential difference will be unmeasurable by this simple method.

The next answer on the Quora page that is worth discussing was given by Daniel Fernandes, EE PhD Student at Stanford University in 2014, which was upvoted by another notable person, Idham Hafizh, PhD candidate in Electronics, Politecnico Di Milano. It was posted a year later than the top wrong answer of Rehan; apparently Daniel didn't fall for the deception. Actually this answer doesn't originate from Daniel, but instead he simply posted a quote from a textbook:

The contact potential across W is a built-in potential barrier, in that it is necessary to the maintenance of equilibrium at the junction; it does not imply any external potential. Indeed, the contact potential cannot be measured by placing a voltmeter across the devices, because new contact potentials are formed at each probe, just canceling V0. By definition V0 is a equilibrium quantity, and no net current can result from it.

Solid State Electronic Devices by Streetman and Banerjee, Sixth Edition.

This is actually the best answer on that Quora page, and it supposed to be voted up by consensus to the top. But apparently, the distorted, corrupted, and illusory 'democracy' (voting) doesn't serve the real truth and the interests of the readers, but rather the corrupting corporate interests.

This quote from the textbook is a bit vague (perhaps deliberately so), in order to strike a compromise between giving a totally wrong explanation that supports the disinfo agenda, and a completely correct description with detailed explanations that would imply the possibility to violate the 2nd law of thermodynamics. At one point in the book the author acknowledges that there is a real potential difference between the p and n regions outside the depletion region, and thus also between the two ends of the semiconductor:

The electric field appears in some region W about the junction, and there is an equilibrium potential difference V0 across W. In the electrostatic potential diagram of Fig. 5–11b, there is a gradient in potential in the direction opposite to E, in accordance with the fundamental relation 4 E(x) = -dV(x)/dx. We assume the electric field is zero in the neutral regions outside W. Thus there is a constant potential Vn in the neutral n material, a constant Vp in the neutral p material, and a potential difference V0 = Vn-Vp between the two. The region W is called the transition region, and the potential difference V0 is called the contact potential.

But, in the quote posted earlier he is trying to obfuscate the implications of this fact by immediately adding that "The contact potential... does not imply any external potential." This statement is vague, and it can be interpreted at least in two different ways. One possible interpretation is that by "external potential" he means the external voltage across the metal terminals of a commercial diode. In this case his statement is correct, because there is indeed no net voltage across these terminals due to the opposing contact potentials between the semiconductor and metal terminals that cancel the built-in voltage.

The other possible interpretation is that the "external potential" here means the potential difference between the two external surfaces of the diode semiconductor's ends (the semiconductor surfaces farthest from the p-n junction) when there are no metal terminals attached. In that case this statement is false, because there is a voltage across these external extremities of the semiconductor surfaces, and also an external electric field in the vacuum (or air, or other non-conductive material that surrounds the semiconductor). The author himself has acknowledged this fact in the latest quote.

The other vague statement of the author that more forcefully dissuades people from asking inconvenient questions about this matter is: "By definition V0 is a equilibrium quantity, and no net current can result from it." Again, one possible interpretation of "no net current" here is to assume that this is a DC current, which is expected to flow in a closed circuit composed of metal conductors, and thus including the pesky contact potential differences that cancel the current. In this case the statement is correct. Most readers interpret it this way, and stop investigating the phenomenon any further as is appears uninteresting and fully explained.

The other possible interpretation is that "no net current" means no instantaneous transient current inside the semiconductor either. With this interpretation the statement is incorrect, because the diffusion current (driven by thermal energy) that builds up the built-in voltage can drive a net current within the semiconductor, whenever the balance between the drift current and the diffusion current is disrupted. This net current will flow until the balance is established again. This is a transient process, but nevertheless a real process, and if it is periodically repeated, it can drive a continuous net current even if it is not a smooth DC. Such a device was proposed by Dr. Sheehan.

Even though the authors of the quoted textbook Ben G. Streetman and Sanjay Kumar Banerjee did their best to dissuade readers from exploring this phenomenon to violate the laws of thermodynamics, this is still one of the clearest and best official explanations. In other official literature there are much more aggressive attempts to conceal the possibility to use this as a Maxwell demon. In my next post I will expose a few samples of such aggressive propaganda.

[ZL]

Continued from my last post.
Now let's take a final look at the chronological order of the answers on that Quora page. The first was posted in sep. 2012 by Abde referring to schottky diodes, which was principally correct. The establishment didn't like this partial truth, therefore they have posted the completely wrong debunking answer of Rehan on Apr 7, 2013, and promoted it to the top as the best explanation, chosen by artificial consensus. Then on Jul 28, 2014 Daniel Fernandes has posted a quote from a textbook that provided the best answer, and it is more authoritative than that of Rehan. This spoiled the initial plan of disinformation, demonstrating that there are still people out there who can think independently and see through the hoax. They needed a plan B to debunk the truth again, thus came Farhad and posted his nonsense on Jan 26, 2016. This pretty much looks like a poker or chess game, where the truth and freedom is at stake.

We have discussed earlier the hair rising redefinition of the word "voltage" in the answer of Rehan. This is a very aggressive attack on scientific truth. It is such a radical assertion that it must have come from higher 'authorities' of disinformation, and in that case it must be present and advertised at other places as well. Googling the expression from Rehan's post : 'A voltage refers to a difference in fermi level between two points on the device' we get a Wikipedia hit about Voltage. Here is a relevant quote from that page that pushes the same aggressive agenda:

The electric field is not the only factor determining charge flow in a material, and different materials naturally develop electric potential differences at equilibrium (Galvani potentials). The electric potential of a material is not even a well defined quantity, since it varies on the subatomic scale. A more convenient definition of 'voltage' can be found instead in the concept of Fermi level. In this case the voltage between two bodies is the thermodynamic work required to move a unit of charge between them. This definition is practical since a real voltmeter actually measures this work, not a difference in electric potential.

https://en.wikipedia.org/wiki/Voltage

It is trivially obvious that the electric field is not the only force that can move the electric charges and create current. Simple mechanical force, like the one driving the belt of a Van de Graaff generator can move electric charges and generate current. In our case of Sheehan's diode the thermal energy can move free charge carriers via the process of diffusion and create the diffusion current. This doesn't provide a legitimate reason to change the definition of voltage. The fact that the electric potential can vary on a subatomic scale, does not mean that the original definition is useless, and therefore it has to be thrown out through the window. If the awkward newly proposed redefinition of voltage supposed to represent the "thermodynamic work required to move a unit of charge" between two points, then it is not electric in nature anymore. But the unit of Volt is purely electric in nature, and therefore voltage supposed to remain such as well, unless of course the 'authorities' want to confuse people by mixing apples with oranges. They are trying to mix up the work done by thermal movement (apples), with the work done by purely electric forces or electric field (oranges) under the term "thermodynamic work". If a mathematician would want to redefine the arithmetic operation of addition as 2 nails + 2 hammers = 4 nails, then people would be rightfully shocked and we would hear plenty of protests. Why is this not happening in the case of new voltage redefinition attempt?

Here is another example of obfuscation in the book "Fundamentals of Solid-State Electronics" by Sah, Chih-Tang:

The fundamental esoteric point is that it is the gradient of the quasi-Fermi potential that drives the diffusion plus the drift current as shown by (331.1) to (331.4A) while the gradient of the electrostatic potential produces only a electric or electrostatic field that drives only a drift current. This drift current is completely canceled by the diffusion current at equilibrium and hence the potential that drives the drift current cannot be measured.

(The emphasis is not mine but present in the book as well). The fact that the drift current cancels the diffusion current in an equilibrium doesn't mean that the potential difference and electric field that drives the drift current doesn't exist for an outside observer, and can not be measured. It can be measured, and it has been measured, just not in the simplistic way of touching the two probes of a common voltmeter to the terminals of the diode. The effort to mix apples with oranges and combine the "diffusion plus drift currents" under the banner of quasy-Fermi potential is present here as well. Why not treat the diffusion current as a separate phenomenon from the drift current as they really are? The diffusion current doesn't need the existence of a built-in voltage (or drift current) because it is driven by the thermal energy combined with the gradient of free charge carrier concentrations (diffusion). The drift current is driven by an electric field. They are completely separate phenomena happening at the same place at the same time.

Did you notice the expression "The fundamental esoteric point" (underscored in the book for emphasis)? Esoteric point? How does esotery become part of science, especially when dealing with relatively simple phenomena that can be verified by accurate measurements? This in itself is a tell-tale sign of shady intentions.

I am sure that if we would continue to search diligently we could find more attempts in literature to hide the truth about the real nature of diffusion current in a diode, and the fact that it can be used as a Maxwell demon. But this much will suffice for now, and in my next post I will start discussing Germano's paper.

[ZL]

The proper discussion of Germano's paper requires some illustrations and mathematical equations, which can't be easily implemented in a post on this forum. Therefore, there is no way around writing a paper about it in pdf format. This will take more time than just cobbling up a forum post, and it might require a couple of days.

In mean time if anybody already knows what the errors in Germano's paper are based on our explanations so far, then you are welcome to post your explanations.

[not_a_mib]

There may be an easier way to measure the depletion layer potential that does not involve any expensive atomically-tiny pointy objects.

Suppose we have two plates, one of P-type silicon, another of N-type, one micron apart forming a capacitor, and a separate PN diode made of the same materials.  Initially touch the P and N parts of the diode to the matching plates.   The junction should charge up the capacitor to around 0.6 volts.  Now disconnect the diode, then pull the plates apart to one centimeter.  Faster than you can say V=Q/C, the plates should now have 6000 volts across them, enough to make a tiny spark.  For practical measurement, one could connect a charge amplifier between the plates, then measure the charge transferred as they separate.  https://en.wikipedia.org/wiki/Charge_amplifier

This same technique should work to measure almost any contact potential between conductors.

[ZL]

Great ideas not_a_mib!

The basic principle of your first suggestion to increase the potential difference by separating the plates of charged capacitor was actually mentioned in Germano's debunking paper "A Note on Solid-State Maxwell Demon" and he also referred to two other papers that have discussed this method. One of them is "The discovery of the electric current" by Piero Cotti and here is a relevant quote from it:

After 1792, Volta searched for the contact voltage between different metals. In 1795, he succeeded in building a sensitive static electrometer, in which he made use of a variable capacity (Fig. 1). Thus, he had a very sophisticated measuring device at his disposal, with which voltages of around one volt could be measured. In 1799, he carried out the first measurements of contact voltage between metals, a phenomenon which is only measurable when the contact is broken (Fig. 2).

https://kundoc.com/pdf-the-discovery-of-the-electric-current-.html
(it is wise to download everything connected to this research as long as you can...)

But here is the "magic" claim that Germano uses as a weapon "...contact voltage between metals, a phenomenon which is only measurable when the contact is broken". This has been weaponized by Germano and others in his mentioned paper as:

Far from the depletion region there is no free charge accumulation. A simple laboratory experiment with Cu and Zn plates and a gold-leaf electroscope can confirm such a behavior [6,7]. Only when the two metals are removed apart the charges, initially localized within the depletion layer, are free to spread across the surfaces of the metallic plates [6,7,8], satisfying electrostatic equi-potentiality, see Fig. 4.

It is trivially simple to realize why the actually existing free charge accumulation at the ends of the diode can't be measured with a simple gold-leaf electroscope. Its sensitivity is not good enough for this. It can't measure a voltage as low as 1V. If it could do this, then Volta would not have needed to design a special variable capacity electroscope for the purpose, but he could have measured it with a common metal leaf electroscope. Separating the two dissimilar metals from each other is necessary only in order to amplify the voltage and make it measurable with a metal leaf electroscope. But the contact voltage is there even when the two metals are in contact. This voltage is measurable with today's very sensitive high input impedance electronic electroscopes even without separating the two metals (by electrostatic induction; without touching the terminals).

Harper has invested a lot of effort and science into measuring the contact potential accurately and reliably using the principle of variable capacitance (to amplify the voltage). There is quite a bit of unexpected difficulty in doing this in practice. The difficulty is that while the metals are being separated, there can be several points of contact to be broken, and much of the charge can get discharged in this process (flow back to the other metal) while the plates are in very close proximity, but not touching anymore. Here is the paper of Harper:

The Volta Effect as a Cause of Static Electrification by Harper
https://www.jstor.org/stable/98725?seq=1#page_scan_tab_contents
(if you sign up for a free account at jstor, you can download this paper for free)

Even though your version of this measurement still requires separating charged plates to amplify voltage, it can still debunk Germano's argument. If it were true that there is no free charge accumulation far from the depletion region (and thus no voltage across the diode semiconductor terminals) as Germano claims, then the two attached semiconductor capacitor plates could not pick up any charges at all. Then one could not measure any voltage across the separated plates either.

The only difficulty is the fabrication of the external semiconductor plates that have the exact same material properties and doping concentrations as those of the diode terminals. Another difficulty is to do the measurement in conditions where the contamination of the plates can be prevented, otherwise the doping can be altered. The best option would be to make the plates and the diode from scratch, from the same material using the same doping. I am not sure though if this can be practically accomplished at home without access to clean room etc. If you don't want to actually measure the charge and voltage, but you are satisfied with simply detecting if there is any charge at all, then I suppose one could attempt such an experiment at home as well.

But even if you would detect a charge by this method, Germano and his debunker fellows could still argue that the charges did not originate from the diode p-n depletion region, but from the contacts between the diode terminals and capacitor plates when they were separated. This way their claim that the charges appear only when the plates are separated could remain vindicated.

However, if you make use of his analogy between the diode's built-in voltage, and two dissimilar metal-metal contact potential difference, then you are in much better position to disprove the disputed claim. If instead of a diode, you use a thermocouple (or just connect two dissimilar metals with high work function differences), and instead of the semiconductor plates, you use capacitor plates made of the same two metals as those of the thermocouple, then that can definitely disprove Germano's argument. It can do that, because there can't be a contact potential difference between the capacitor plate and thermocouple terminal if they are made of the same metal. Therefore, if there is any charge and voltage on the capacitor plates after the separation/amplification, then that could have originated only from the contact potential of the junction (which has not been separated). These charges must have been present at the terminals of the thermocouple far away from the junction, which debunks Germano's claim that there could not be any free charge accumulation outside the depletion region, while the dissimilar metals are in contact. Therefore this is an experiment that is worth performing at home, to put this argument to rest once and for all.

The other option you mentioned to use modern semiconductors and opamps to built sensitive electroscopes that can measure even 1V with extremely low leakage current is another, perhaps even better option to experimentally disprove the disputed argument. For those with deep pockets this should be a piece of cake by using one of the high end electroscopes of Tektronix, like the "Keithley Electrometers for Ultra-High Resistance/Ultra-Low Current Measurements" which has got these characteristics:

The 5½-digit Model 6514 and Model 6517B Electrometers offer 1fA sensitivity, >200TΩ input impedance on voltage measurements, and charge measurements down to 10fC. The 6½-digit Model 6430 Sub-Femtoamp Remote SourceMeter SMU Instrument can measure current with 1aA sensitivity. Its low noise and drift performance make it ideal for research on single electron devices, highly resistive nanowires and nanotubes, polymers, and electrochemical applications.

https://www.tek.com/keithley-low-level-sensitive-and-specialty-instruments/keithley-high-resistance-low-current-electrom
(or similar like PASCO Model ES-9078).


I suppose the readers of this forum can't afford such instrument, but with some extra elbow grease they can design and build their own electrometers using schematics like those on the wiki page. Or perhaps even a primitive circuit like the one used by Bill Beaty could be used to at least detect (even if its measurement accuracy would not be good enough) the existence of free charge at the terminals of the thermocouple. As long as the whole device is kept at the same temperature it should provide an acceptable practical proof.

Ridiculously Sensitive Electric Charge Detector
http://amasci.com/emotor/chargdet.html

Regarding the paper I am working on that explains the errors in Germano's arguments in great detail; it takes longer to prepare than I thought, but it is coming. The reason for the delay is this claim of his:

...This behavior does not match what happens in laboratory experiments and in the real world.
It is already well known that this is not what really happens (see the Volta effect [8]). When two metals with different work functions (and similarly, when an n- and a p-semiconductor) are joined, the charge drift is only local and the charge displacement remains localized within the thin depletion layer, in equilibrium. Far from the depletion region there is no free charge accumulation. A simple laboratory experiment with Cu and Zn plates and a gold-leaf electroscope can confirm such a behavior [6,7]. Only when the two metals are removed apart the charges, initially localized within the depletion layer, are free to spread across the surfaces of the metallic plates [6,7,8], satisfying electrostatic equi-potentiality, see Fig. 4.

He talks about this nonsense "...It is already well known that this is not what really happens..." as if it would be a well established non-controversial fact that everybody supposed to know (and accept), and which has been already proven in the papers he referred to here. Well, to give it justice, I have had to find and read these reference papers, and see what kind of proof they can offer to support this false claim. Then since I was forced to put so much time and effort into this, then let me make the disproof even more accurate and convincing by arranging some FEM simulations of a diode's electrostatic field that surrounds it, and include such images. He hasn't got a single argument that can stand a scientific scrutiny, based on well established laws of electrostatics and semiconductor physics. Hopefully it will be ready in a few more days.