Pinhole drive :D

[vlaiv]

I had a brief read on em drive that got me thinking ....

After much of back and forth with newtonian mechanics, firing balls into walls of space ship, shooting electrons into magnetic fields and letting their trajectory curve, then smashing them into walls - all of this in my head of course and all in vain, I came up with an idea that, as of this time, can't find fault yet, so decided to share it with more knowledgeable people. Someone could find and point out problem in my reasoning.

We will start with really simple setup - cube shaped spaceship. We attach torch on ceiling and shine light on the wall on the other side - bottom. So far nothing interesting happens - we fire off photons - they carry a bit of momentum with them "pushing" spaceship in opposite direction - but rather soonish they collide with the opposite wall transferring back momentum - spaceship is "pushed" back - total motion 0. So far so good.

Then we do a strange thing. We place a pinhole in light's path. What happens here? Well some of the photons happily go in a straight line thru the hole and hit opposite wall, but we won't look at those. We shall take a look at ones that due to diffraction spread after going thru the pinhole. Here is interesting thing - momentum when we fired photons was in downward direction - whole vector pointing down - but when they strike opposite wall - being "bent" by pinhole and spread into diffraction pattern - they no longer have momentum pointing straight down - but rather at an angle. So vertical component of their momentum is less than starting momentum. Horizontal momenta cancel out due to symmetry of diffraction pattern - and sum of vertical components of momentum vectors of all photons that were bent by pinhole and diffraction is less than total momentum of fired photons. Conclusion? Total horizontal momentum is 0, total vertical momentum is not 0 - hence we have motion

What do you make of this?

 

[vlaiv]

Just to add a bit further thought on this matter - universe is full electromagnetic radiation - and if diffraction effects can cause positive net momentum - this could be explanation for dark energy?

[D4N]

I think that the diffracted light imparts a force on the object it diffracts around thereby cancelling out the thrust.

[Paul M]

It doesn't fly.

I just tried it with a good torch, a shoe box and bathroom scales

I did notice the apparatus getting lighter as the batteries died...

[vlaiv]

I'm not sure that it does. This would imply change of frequency of a photon due to change in energy - some of it's energy must be transferred to pinhole plate, but I don't think we get this with diffraction - it's not like Compton scattering, I don't think that photon touches the pinhole plate at all.

I'm also trying to check path integral for single slit - I don't think there is reduction in frequency (which would mean change of energy of a photon) in calculations of possible paths - these are just virtual paths / all possible paths - not actual paths that photon takes, we don't really know what happens to the photon in flight, or is it in flight at all. If you think about it - from photon's point of view - it has been absorbed in the same instant it was emitted - if something travels at speed of light - time stands still. 

[vlaiv]

Scrap frequency shift from previous post. It does not need to change frequency - light bouncing of a mirror will transfer momentum but will not change energy.

But I still don't think there is momentum exchange between pinhole plate and photon.

[saac]

13 hours ago, vlaiv said:

if something travels at speed of light - time stands still. 

Surely only from the perspective of an observer in a different frame of reference.  

 

Jim

[vlaiv]

This is really starting to melt my brain ...

I further entertained the idea of momentum transfer between photon and pinhole plate - explanation why such a drive would not work.

Let's say there is momentum transfer, and indeed there is momentum transfer when mirror is hit by photon. So let's consider diffraction mirror - or ridged mirror.  We know that such a mirror will not bounce photon at the same angle as angle of incidence. Further more - to get probability distribution at detector we need to calculate all the paths from source to mirror and from mirror to detector. But in actual case of photon bouncing first of ridged mirror and then hitting detector - we can determine true trajectory of photon - by looking at momentum transferred to both mirror and detector - we certainly know true trajectory of photon in this case - how does that translate to dual slit case? If photon going thru dual slit also interacts with slit, and transfers momentum - there is "which path" information, also by looking at the detector we can also have which path information. But if there is which path information - QM states - no interference patterns can form, but interference is there.

So what would this mean? I conclude (maybe wrongly so) that either: there is no momentum transfer at slit, or there is momentum transfer and that implies that ensemble / hidden variable interpretation of QM is correct (there is a single path of photon, but which way in that exact instance is taken is governed by something we can't determine) - something that is currently not accepted as true interpretation of QM.

Or the other way to think about mirror is as follows: photon is emitted - momentum of source is changed, photon travels and it's wave function is evolving until it hits mirror - at that point there is wave function collapse - momentum is transferred to mirror, photon bounced of starts another leg of it's journey with wave function evolving - hits detector - another wave function collapse - another momentum / energy transfer. Each momentum/energy transfer with "macroscopic" object leads to decoherence and "wave function collapse". Since at slit there is no wave function collapse - we can conclude that there is no momentum / energy transfer between photon and macroscopic plate of slit / pinhole. I suspect same thing is happening with diffraction mirror - so in case light is shined on diffraction mirror - we also don't have function collapse - no energy/momentum transfer - another way to create engine - put diffraction mirror on opposite side of wall - fire laser beam - let it bounce back in such a way that it hits front of the ship again.

 

[andrew s]

What you propose is basically correct and has been demonstrated in several experiments. The emission and absorption of em photon involves a transfer of momentum. Any object with a net asymmetric emission or absorption feels a net force. Have a look at a Crooks radiometer. It also acts to shift asteroids in there orbits - the Yarkovsky effect. 

The issues about diffraction at the hole just complicate the issue but doesn't change the basic physics.

Regards Andrew

[vlaiv]

1 minute ago, andrew s said:

What you propose is basically correct and has been demonstrated in several experiments. The emission and absorption of em photon involves a transfer of momentum. Any object with a net asymmetric emission or absorption feels a net force. Have a look at a Crooks radiometer. It also acts to shift asteroids in there orbits - the Yarkovsky effect. 

The issues about diffraction at the hole just complicate the issue so just but don't change the basic physics.

Regards Andrew

You are talking about external source of momentum in these cases?

In the case I'm arguing, system does not exchange momentum with outside - photons (any particles that interfere / diffract) are both emitted and absorbed inside spacecraft / engine, but it appears that net momentum of whole closed system increases - using fan to blow into your sails paradigm. All of this has been inspired by EM drive - that seems to do exactly the same - violates momentum conservation law. It certainly violates momentum conservation law if it works, and can potentially explain dark energy (universe as closed system has positive net momentum increase because of this effect?). On the other case, it would be reasonable that it does not work - due to momentum conservation law, but in that case some interesting questions are raised about our understanding of QM, and it's interpretation. Or of course, there is a third option - I'm talking nonsense, and there is simple explanation for all of this . If we talk in language of QM - third option has highest probability, but I would really like to know where I'm wrong in my reasoning.

[andrew s]

Yes my examples have an outside sourse but they also absorb and emit radiation out of balance.

If you take a box that emits a flux of light through a hole (and assuming no other external forces acting on it) then that flux will carry away both energy and momentum and the box will both lose energy and gain momentum exactly equivalent to that carried away by the light flux. If it did not then energy and momentum would not be conserved. The light flux is no different to a flux of ions from an ion engine in this respect.

If you consider a microscopic example an atom recoils when it emits a photon.

To be pedantic I am assuming a flat space time where energy and momentum are conserved quantities.

Regards Andrew 

[andrew s]

I may have miss understood your proposal if nothing leaves the box then nothing happens.  In fact this is one way to derive E=MC^2 and if I remember correctly that is how Einstein did it.

If you beam light or anything else from one end of the box and absorb it at the other the box will move to keep its center of mass it the same location wrt an inertial frame at rest with the box before the beaming started.

Just the same thing happens if you have a weight on a compressed spring inside a box and let it go the "box" moves but its center of mass stays put.

Regards Andrew

[andrew s]

I was basically responding to you first most having looked more closely at your later post the issue of conservation laws in QM is raised.

In QED the best theory we have that satisfies special relativity and quantum mechanics both energy and momentum are conserved however, it can't be decomposed as in you discussion.

You have to remember that the Feynman many paths approach is a technique for doing the difficult sums that arise in QED and should not be taken literally any more than the terms in a series expansion of sin(x) ~ x - x^3/3! + x^5/5! -... means the ratio of the relevant sides of a triangle angle is made up of these bits.

Regards Andrew