måndag 16 oktober 2023

Towards Real Quantum Chemistry

Real Quantum Mechanics RealQM offers a model of an atom or ion as a positively charged kernel surrounded by a system of single non-overlapping electron densities interacting by Coulombic forces.  This is a classical continuum mechanics model, where the electron densities have individuality by occupying different regions of shared 3d space. RealQM is computable also for atoms with many electrons and shows good agreement with observation. 

RealQM for atoms/ions directly extends to Real Quantum Chemistry as interaction of atomic kernels surrounded by electron densities into molecules consisting of several kernels surrounded by non-overlapping electron densities. The constituents of molecules as atomic kernels and electron densities thus carry individuality, just like different material objects of continuum mechanics. 

In a world of RealQM there is thus no principal difference between physics as physics of atoms and chemistry as physics of molecules.  This not so in the world of standard Quantum Mechanics stdQM with its multidimensional wave function depending on $3N$ space coordinates for $N$ electrons, where electrons lack individuality and may overlap. This makes extension from atoms to molecules hard to realise because in the mind of a chemist, molecules occupy different regions in space with shape determining chemical properties.  

In stdQM there is thus a rift between physics of atoms without electron individuality and physics of molecules with kernel/electron individuality. Even more troubling, stdQM is uncomputable because of the high dimensionality making already the H2 molecule with two electrons into a computational challenge in six dimension, and already H2O with 10 electrons behind the capacity of any thinkable computer. 

Here you can yourself test RealQM for simple atoms/molecules by running (and inspecting) the (short) p5js-code (consisting of essentially 3 lines for updating electron densities, potentials and domains, click top name to see code and run by clicking arrow):

  

lördag 14 oktober 2023

Chemistry vs Quantum Mechanics

A basic mantra of modern physics is that standard quantum mechanics stdQM as the physics of atoms can explain chemistry as the physics of molecules, by the fact that the stdQM Schrödinger wave function $\Psi (x1,...,xN)$ for an atom/molecule with $N$ electrons contains everything there is to say about the atom/molecule and its interaction with other molecules as chemistry. But there is a (serious) caveat: $\Psi (x1,...,xN)$ cannot be computed because it depends on $3N$ space dimensions. Already $N=2$ presents a challenge and $N>6$ is beyond the capacity of any thinkable computer. 

The mantra that stdQM explains chemistry thus lacks real concrete support, as argued by e.g Eric Scerri. There is in fact a big leap from stdQM to chemistry in the sense that atomic kernels have individuality by position in space, and so also their surrounding electrons and connected molecules/chemistry, while electrons in stdQM lack individuality; they have no specific positions in space and are supposed to be identical. 

RealQM presents an alternative to stdQM in the form of a classical continuum model in 3 space dimensions where electrons have individuality by occupying different regions in physical space without overlap. RealQM is computable also for molecules with many electrons and so may have the capacity to explain chemistry.

Here you can follow RealQM for a Carbon2 molecule consisting of two kernels and 12 electrons.  

Here is Boron2 molecule with two kernels and 10 electrons. 

Recall that efforts to bring back stdQM to reality in 3d were made by Richard Bader with his Atoms in Molecules AIM theory and in Density Functional Theory DFT (Nobel Prize in Chemistry 1998) by integrating $\Psi (x1,...,xN)$ over all but one variable, however without electron individuality causing serious problems. 

Recall that the physical meaning of stdQM is still a mystery after 100 years of fruitless search and dispute, which can only mean that stdQM lacks (physical) meaning.

fredag 13 oktober 2023

Why Helium Does Not Form Helium2 Molecule?

RealQM may explain why Helium He does not form a He2 molecule like Hydrogen H forming a H2 molecule. In the RealQM model of the He atom its two electrons share the volume around the kernel in two half-spherical lobes meeting at a common separating plane appearing as a free boundary. In the plot below you see two nearby He atoms forming a He2 system with two of the electronic half-lobes meeting between the kernels and the other two on both sides of the system. 

The total energy of the He2 system is by RealQM computed to be -5.81 Hartree (see plot and code), which is very close to that of two separate He atoms of -5.806 Hartree, and so we find that He does not have tendency to form He2 as a molecule, according to RealQM. 



We now compare with H which forms H2 molecule from a concentration of electron density between the kernels with lower energy (-1.17) than two separate H atoms (-1), as shown in this computation: 


What is then the difference between H forming H2 and He not forming He2? What is the difference between a +1 kernel surrounded by a -1 charge and a +2 kernel surrounded by -2 charge? Inspecting the plots we understand that the two outer halt-lobes in He2 are repelled by the two inner half-lobes and so are prevented from concentrating between the kernels. This is not so for H2 since in RealQM electrons are not subject to self-repulsion, as discussed in previous post. 

There is thus a difference between a shell with several repelling electrons (2 for He)  and a homogenised shell without self-repulsion (1 for H). This is accounted for in this spherical shell model of RealQM.

Standard QM offers a qualitative explanation by He having a full 1st shell without incentive to combine,  while a quantitive computation requiring 12 space dimensions may be out of reach.    


torsdag 12 oktober 2023

Quantum Theory Paralysed by Self-Interaction

A basic idea/postulate of quantum mechanics is that electrons as particles lack identity/individuality, and so cannot be distinguished and surveyed, as if in face-recognition all faces are the same (or everybody wears a mask) which prevents survey. 

This  menas that the interaction between fields created by electrons/particles, and electrons/particles gets muddled and so opens to self-interaction with a particle interacting with the field it itself creates, which is (very) problematic.  

The same confusion would arise if you cannot distinguish yourself as a person from other persons. 

The previous post made the remark that there is no self-interaction in Newtonian mechanics since material particles, even of the same kind, can be distinguished by having different positions and can be surveyed under continuous motion. 

To see the interplay between particles and fields in Newtonian mechanics, recall that the gravitational potential field  $\phi_y (x)$ created by a point mass at $x=y$ is given by (modulo gravitational constant) for $x\ne y$:

  •  $\phi_y(x)=-\frac{1}{\vert x-y\vert}$   
with  
  • $-\nabla \phi (x) = -\frac{x-y}{\vert x-y\vert^3}$ 
the gravitational force between a point mass at $x\ne y$ and the point mass at $y$ (scaling with $\vert x-y\vert^{-2}$ as Newton's familiar law of gravitation.)

For a continuous distribution $\rho (y)$ of mass the gravitational potential is given by the integral 
  • $\phi_y (x)=-\int\frac{\rho (y)}{\vert x-y\vert}dy$ 
where thus the point value $\rho (y)$ gives no contribution under the integration, which signals no self-interaction; we can safely assume that $x\ne y$ thus avoiding infinities arising if $x=y$.  

In Newtonian mechanics mass particles occupy distinct positions and interact by Newton's law of gravitation without self-interaction. Particles and fields are separated in the sense that the potential field/force created by a particle does not act upon itself, only at particles with distinct positions. 

In quantum mechanics this is not so because particles lack individuality and so can occupy the same position (if spins are opposite). This means that the a particle and its field are intertwined and open to self-interaction with infinities causing severe problems. 

In RealQM electron densities do not overlap and there are no infinities. 

onsdag 11 oktober 2023

RealQM is Free of Self-Interaction

The big unsolved problem of Quantum Field Theory QFT is infinities arising from self-interaction, which have to be removed by techniques of renormalisation. Self-interaction arises because QFT is not a physical ontological theory of what is, but a mathematical formalism lacking physical reality and filled with mystery. Standard Quantum Mechanics also suffers from self-interaction, appearing as an unresolved main problem of Density Functional Theory. 

In Newton's theory of gravitation the gravitational force between two (point) masses at distance $r>0$ scales with $r^{-2}$, where infinity with $r=0$ does not occur, because the (point) masses do not occupy the same spot in space; they have individuality and do not sit on top of each other. This means that there is no self-interaction in Newton's mechanics. 

Quantum Mechanics QM is claimed to be different from Newton's mechanics in that electrons can overlay if they have opposite spin, that is, different electrons can occupy the same spot. This opens to self-interaction because overlaying electrons loose identity and so can interact with themselves. In short, QM is a non-physical (contradictory) theory and as such is very problematic, even if it sometimes produces results in agreement with observation (like a stand-still clock).

RealQM is a new approach to atom physics in the spirit of classical continuum mechanics based on non-overlapping electron densities without self-interaction. Like Newton's mechanics, RealQM is free of this defect. 

RealQM is computable for atoms/molecules with many electrons and seems to capture real physics, with code examples in preceding posts. The code essentially consists of 3 lines expressing energy minimisation by update of (i) electron densities, (ii) electron/kernel potentials and (iii) free boundary between different electrons to reach continuity of electron density.  Test yourself!   

lördag 7 oktober 2023

RealQM for Molecules

RealQM (see also earlier posts) offers ab initio computation/simulation of atoms and molecules in ground state, dissociation, ionisation, excitation and emission. This is made possible because RealQM is a classical continuum mechanics model in 3 space dimensions + time. An simple example of molecules of the form X=X with a double bond between atoms X with X = C, N, O, F...can be inspected by running the corresponding p5js code with typical output (coarse grid, also compare with X=H):


We compare with multidimensional Standard QM which for molecules boils down to a tricky play with s, p, d orbitals, and even more tricky the choice of spin up or down for each electron, to ad hoc find some agreement with observations. This technique was formed in the 1940s by Born-Oppenheimer, Hartree-Fock, Slater followed by Density Functional Theory in the 1970s without reaching ab initio simulation. 

Here is computation of dissociation energy for H20 (with p5js-code): 


O atom in the middle modeled with 2 + 4 + 2 shell electron distribution flanked by two H atoms with accumulation of electron density between O and H acting like a glue. The total energy of both molecule H2O and constituents 2H + 0 is around -77 Hartree and the dissociation energy of about 0.3 Hartree thus represents a small energy difference arising from the interaction of the 2 outer shell electrons of O with the  2 H electrons.  

Here is H2+ Ion (one electron) as a stable molecule with binding energy 0.1 Hartree (code);


 

  

onsdag 4 oktober 2023

Prototype RealQM for Helium Atom: Test Yourself!

Real Quantum Mechanics (RealQM) offers a new atom/ion/molecule model in the form of a system of non-overlapping electron densities as a classical continuum mechanics Schrödinger model in 3 space dimensions + time. This model describes a reality/actuality and is readily computable also for many electrons, which is not the case for standard QM (stdQM) as a multi-dimensional statistical model describing possibilities instead of actualities. See previous post for a connection to the 2023 Nobel Prize in Physics depicting electron charge densities, thus supporting RealQM, while real pictures of stdQM are missing. 

A prototype computational realisation of RealQM for the Helium atom with two electrons can be inspected here. The ground state of Helium is computed by minimising the total energy consisting of kinetic energy, attractive kernel potential energy of kernel at $x=0$ and repulsive mutual electron energy:  

  • $\frac{1}{2}\int w(x)\vert\nabla u(x)\vert^2 dx -\int\frac{2u^2(x)}{\vert x\vert}dx+\int\int\frac{u^2(x)u^2(y)}{\vert x-y\vert}dxdy$
over a decomposition over a fixed grid of $u(x)=u_1(x)+u_2(x)$ into two electron density functions $u_1(x)$ and $u_2(x)$ with disjoint supports in 3d space given by characteristic functions $w_1(x)$ and $w_2(x)$ with $u_1^2(x)$ the charge density of electron 1 and $u_2^2(x)$ that of electron 2, with the boundary between $w_1$ and $w_2$ acting as a free boundary determined to achieve that u_1 and u_2 agree on the free boundary so that the electrons meet with same density.

The code consists of a couple of lines expressing gradient minimisation involving
  1. Explicit relaxation of Hamiltonian in u + charge density normalisation (involves homogeneous Neuman conditions for u_1 and u_2 on free boundary enforced by the presence of $w(x)$ in the kinetic energy).
  2. Explicit relaxation of Poisson problem for electron potentials.
  3. Explicit level set front tracking of w to reach continuity of u over fixed grid.
which can be seen as a Bernoulli free boundary problem with homogeneous Neumann + continuity as free boundary condition.  

Run the code by clicking the arrow in the p5js code to see electrons initiated away from the kernel approaching the kernel and meeting at a free boundary with continuity (and approx homogeneous Neumann condition because of coarse resolution):   


and experiment further by modifying the code. It is fun and illuminating!

Compare with 1st excited state with 1 electron in 1st shell and 1 electron in 2nd shell (with p5js code):


Here you can test RealQM for the H2 molecule including minimisation over kernel distance:


Here you can test 2-shell atoms/ions starting with Lithium with 2 electrons in 1st shell and 1 electron in 2nd shell, continuing with always 2 electrons in 1st shell and Beryllium/ions with 2 electrons in 2nd shell.

This is a preparation for atoms/ions with more than 2 shells starting with Boron with 2+2+1 in 3 shells, Carbon 2+2+2, Nitrogen 2+2+3, Oxygen 2+2+4, Fluorine 2+3+4 and Neon 2+4+4 filling the first period.
And so on through the whole table... 

The electrons in each shell are captured by one electron density function carrying the total electron charge.

Here a test (with p5js-code to test) for Beryllium with 4 electrons in 2 shells initiated with a gap between of the 2 electrons in 2nd shell which closes under free boundary level set tracking:



See further RealQM results:
Computations in 3d without assuming spherical symmetry. Compare with earlier Atom Simulator in spherical symmetry.


Nobel Prize Physics 2023 vs RealQM

The Nobel Prize in Physics 2023 has been awarded to Pierre Agostini, Ferenc Krausz and Anne L’Huillier for

  • Experimental methods that generate attosecond pulses of light for the study of electron dynamics in matter.
  • The laureates’ experiments have produced pulses of light so short that they are measured in attoseconds, thus demonstrating that these pulses can be used to provide images of processes inside atoms and molecules.
This directly connects to Real Quantum Mechanics, which describes the dynamics of N interacting electrons as a 3d continuum mechanics system of N non-overlapping distributed charge densities, instead of the standard QM statistical particle model in 3N spatial dimensions. 

The Prize shows that pictures of electron charge densities can be taken, which gives support of an idea that they do indeed exist, as the essence of RealQM.  

This is to be compared with pictures of stdQM electron particles, which have not been taken, which gives support of an idea that they do not exist.

Hopefully, the 2023 Physics Prize can give an incentive to take a look at RealQM. See also this recent post  and earlier posts on RealQM.
 
Recall Wittgenstein: What you can take a picture of, you can speak of and believe to exist.

Here is an illuminating picture of two interacting electron charge densities from the presentation of the Prize by the Royal Swedish Academy of Sciences (assuming a preschool level audience?): 


 Compare with a RealQM picture of the interacting electrons of a H2O molecule


which confirms the chemist (but not physicist!) conception of H2O:


PS A closer look at the work awarded the prize, somewhat disappointingly reveals that it does not really contain images of processes inside atoms and molecules. Strange.

String Theory = Theory of Everything?

Peter Woit on Note Even Wrong recalls a podcast with Eli Frenkel about string theory:

  • In particular, one thing that happened to Frenkel since last spring is that he attended Strings 2023 and gave a talk there (slides here, video here). The experience opened his eyes to just how bad some of the long-standing problems with string theory have gotten, and starting around here in the podcast he has a lot to say about them.
  • It’s pretty clear that his reaction to what he saw going on at the conference was colored by his experience growing up in late Soviet-era Russia, where the failure of the system had become clear to everyone, but you weren’t supposed to say anything about this. He pins responsibility for this situation on senior leaders of the field, who have been unwilling to admit failure.
String theory was initiated in the 1970s as a grand attempt to unite all forces into a Theory of Everything ToE. My good friend Lars Brink at Chalmers was one of the early pioneers with a thesis in 1973 and subsequent work with all the senior leaders during a 50 year career until Sept 2022 when he had to give up his search for a ToE. A Memorial Meeting took place on the web Febr 18 2023 remembering a great physicist.

According to Eli Frenkel, sad to say, string theory has not delivered a credible ToE and so it is time to move on to something different building on the experience of a failed attempt. But this is not what the  leaders of string theory still alive are saying, which creates a big problem for the young generation asking for guidance and hope/inspiration for a career in physics. String theory? If not string theory, what then? What has fundamental physics de facto delivered the last 50 years? Why are the foundations of modern physics (quantum mechanics and relativity) still incompatible after 100 years? Why is there a crisis in fundamental physics if its leaders are smartest on Earth?

Eli Frenkel gives a perspective that is important to listen to if you are a young scientist in search of a mission. Old professors/theories never die, they just fade way.



måndag 2 oktober 2023

What Can Quantum Mechanics Predict?

The previous post recalled the fact that standard Quantum Mechanics QM based on a $3N$ spatial dimensional Schrödinger equation for a system with $N$ electrons is effectively uncomputable for almost all atoms in the periodic table. With a resolution of $10^2$ in each spatial dimension, the total number of mesh points is daunting $10^{6N}$ which fills any thinkable computer already for $N=5$ (Boron). The wave function for all atoms beyond Boron is thus uncomputable. This means that the wave function for almost all atoms in the period table is uncomputable!

We compare with the standard message by physicists that all predictions of QM perfectly match with observation, or more precisely that there are no observations which contradict QM. 

The crucial question is then what predictions QM offers? Does QM predict the periodic table? A physicist would say that certainly this is so, in principle, while precise computations are lacking, because the full wave function is uncomputable. 

What can be computed are various low-dimensional approximations based on a variety of ad hoc assumptions. If such an ad hoc approximate solution happens to match observation, it is accepted as a demonstration that the full wave function (although unknown) also matches observation. If the ad hoc solutions does not match observation, it is discarded. With this strategy it is impossible to find a contradiction between theory and experiment! Success story!

But there is a caveat. Do you see it? Compare with How far does quantum mechanics explain the periodic table?

Real Quantum Mechanics presents a new Schrödinger equation as a non-linear system of non-overlapping electron charge densities sharing a common 3d space, thus a classical continuum model, which is computable for many electrons. A laptop computation for H2O is reported here with more to come.

Summary: QM in fact predicts very little (ab initio without ad hoc assumptions), which is consistent with the fact that there is no observation in contradiction with a QM prediction. But a theory which predicts nothing is not falsifiable and as such not a scientific theory. This is the reason behind the crisis of modern physics based on QM witnessed by so many physicists. Real QM opens a door out of the crisis. Why not give it a try? Permanent QM crisis over 100 years is not healthy to physics.