User comments: O-O formation issues

oliver-phet commented 7 years ago

Comment from Sandor Sep 27, 2016:

I have just translated the HTML5 version of the old Java sim. (The old strings have not been included int he form.) Please pass the attached screenshot to the (theoretical/computational) chemist expert who takes care of your sims for reconsideration. The O-O example obviously shows molecule formation. It is very strange to me to describe it in terms of van der Waals force. On the other hand I like this example, but I think it is misleading in this form.

oliver-phet commented 7 years ago

Perhaps @arouinfar can comment?

ariel-phet commented 7 years ago

Not sure, but I think @amyh-phet will have more insight

arouinfar commented 7 years ago

@oliver-phet @ariel-phet @amyh-phet

One of the learning goals of the simulation is to

Contrast the potential well and behavior of a bonded pair of atoms with unbonded pairs.

This behavior was present in the Java sim, though sigma and epsilon were adjusted to be more realistic for the O-O pair. YYC stated in the design doc that:

If using LJ potential to approximate the O-O bonding interaction, epsilon should roughly correspond to the average bond energy (since that’s the well-depth), and sigma can be calculated based on the average bond length, since the bottom of the well should be located at the equilibrium bond distance. Consulted with Robert Parson who agrees that this sounds right that the dissociation energy can be used for epsilon.

This has also been noted in the Teacher Tips.

As far as I know, the Lennard-Jones potential is used to describe the interaction of two neutral atoms whether they are bonded or not. The two terms of the L-J Potential correspond to the van der Waals attraction and the Pauli repulsion (which we label as "electron overlap"), so I'm not sure I understand Sandor's comment...

amyh-phet commented 7 years ago

@arouinfar and @oliver-phet The Lennard-Jones potential is typically used to calculate attractions between two neutral particles and models intermolecular attractions, not covalent bonds. I think Sandor has a point that it is a bit unconventional to use L-J potentials for bonded atoms. We did attempt to address this in the Teacher Tips, but perhaps I need to reword how the O-O values for sigma and epsilon were calculated. From YYC's notes, the equation for the L-J potential was not used for this calculation, but information about bond energy/length was used to create the graph for oxygen.

Just curious, do we know if Sandor looked at the Teacher Tips?

amyh-phet commented 7 years ago

@arouinfar and @oliver-phet Can this issue be closed?

ariel-phet commented 7 years ago

There was a further concern raised in #68

I had some email discussion with the user, and I think I understand the concern better. Basically, when you displace the moveable atom from the potential well minimum in the O-O scenario, it causes the "pinned" atom to vibrate on "collision", this vibration then goes away, and the oscillation does not go as high on the walls of the potential well, implying a loss of kinetic energy.

It seems there are issues with showing O-O here, for multiple reasons. I need to discuss this KE loss more with @jbphet

ariel-phet commented 7 years ago

I am also going to probe Emily Borda regarding these questions, she might have some thoughts/insight.

ariel-phet commented 7 years ago

Emily Borda is on vacation but said she would give some input the week of 7/10

ariel-phet commented 7 years ago

Further comments by the user who raised concerns in #68

Having read through the thread in the hub, I ought to join the negative sentiment as to presenting the O+O dynamics on par with the noble pairs. To some extent the dynamics of the oxygen is not different, but the ensuing discussion may be out of scope. However, just one last response: if a bond is formed and KE release, it cannot go to translation of the molecule, because it is released in the center of mass frame of reference. In short, when 2 Oxygen atoms collide, it is impossible for a molecule to form. If the initial state is already a bond state (I.e. molecule) then we are in a vibrational state, which has to be periodic (unless one takes into account IR photon emission, or collisions with other particles.)

This is a good point to thank you and all of the team for providing us, STEM teachers, with good sims. Furthermore, PHET is becoming a standard for good sims, and that means you have added responsibility.

ariel-phet commented 7 years ago

Emily Borda responded:

The problem seems to be that the O-O part of the sim is able to go from an unbonded to bonded situation without the intervention of a third particle to take away excess energy, which seems to violate the law of conservation of energy. If PE decreases, KE must increase. This appears to be what happens, but KE should be so large that the O atom is able to escape its bonding partner after a collision. In order for it to become trapped in a bond, it needs to lose energy. This is one of the major goals of my Energy & Bonding sim; to model interactions where a third particle either inputs energy (in the case of bond breaking) or takes energy away (in the case of bond formation).

I see the Atomic Interactions sim as modeling what a bond is, once it is formed. It does a good job of showing the PE/KE tradeoffs and forces involved in the oscillation of two atoms involved in a bond. But without a third particle it can't model bond breaking or formation. To me, what this means for the Oxygen tab is that if I drag the unpinned O to the right, and it moves to the left, toward the pinned O, it should come back to the exact place I dragged it in the first place.

I hope this makes sense; I'd be happy to have a phone or zoom meeting to chat about this.

It seems we should revisit this behavior