I am trying to convert the dipole moment vector of csv files from x, y, z coordinates to a scalar value using formula: Square root of (x^2 + y^2 + z^2)\nThis formula does not get affected by Cartesian system.
\n\nNow using this formula I am getting magnitude of dipole moment for ammonia as (dsgdb9nsd_000002) as Sqrt(-0.0002^2 + 0^2 + 1.6256^2) = 1.6256\nHowever standard value of magnitude of dipole moment for methane is 1.46 which I found on multiple pages (for example https://www.quora.com/Whose-dipole-moment-is-greater-NH3-or-NF3)
\n\nSimilar is the case with water (getting 1.8511 instead of 1.8546) and almost all the structures that I tested.
\n\nCan anyone suggest where am I going wrong here. Is it the formula I am using?
", "messages": [ { "id": "580221", "postDate": "07/19/2019 21:22:47", "content": "I am trying to convert the dipole moment vector of csv files from x, y, z coordinates to a scalar value using formula: Square root of (x^2 + y^2 + z^2)\nThis formula does not get affected by Cartesian system.
\n\nNow using this formula I am getting magnitude of dipole moment for ammonia as (dsgdb9nsd_000002) as Sqrt(-0.0002^2 + 0^2 + 1.6256^2) = 1.6256\nHowever standard value of magnitude of dipole moment for methane is 1.46 which I found on multiple pages (for example https://www.quora.com/Whose-dipole-moment-is-greater-NH3-or-NF3)
\n\nSimilar is the case with water (getting 1.8511 instead of 1.8546) and almost all the structures that I tested.
\n\nCan anyone suggest where am I going wrong here. Is it the formula I am using?
", "rawMarkdown": "I am trying to convert the dipole moment vector of csv files from x, y, z coordinates to a scalar value using formula: Square root of (x^2 + y^2 + z^2)\nThis formula does not get affected by Cartesian system.\n\nNow using this formula I am getting magnitude of dipole moment for ammonia as (dsgdb9nsd_000002) as Sqrt(-0.0002^2 + 0^2 + 1.6256^2) = 1.6256\nHowever standard value of magnitude of dipole moment for methane is 1.46 which I found on multiple pages (for example https://www.quora.com/Whose-dipole-moment-is-greater-NH3-or-NF3)\n\nSimilar is the case with water (getting 1.8511 instead of 1.8546) and almost all the structures that I tested.\n\nCan anyone suggest where am I going wrong here. Is it the formula I am using?", "votes": null }, { "id": "580273", "postDate": "07/19/2019 23:30:11", "content": "If I am looking at the right molecule I believe I got the same answer as you for methane and water. Guess that could mean we both goofed the same way.
\n\nI like to think I did not - here's what I believe. The scalar coupling value we are trying to predict are values from a non-standard set of conditions (a different space time - perhaps a multi-verse). They are not measured - they were calculated. The same muti-verse supplied all the additional supplemental data - ie, it was calculated under the same set of parameters as the coupling values.
\n\nSo values you might find in a reference are different because they are not from the same set of conditions - I assume temperature being the big one.
", "rawMarkdown": "If I am looking at the right molecule I believe I got the same answer as you for methane and water. Guess that could mean we both goofed the same way.\n\nI like to think I did not - here's what I believe. The scalar coupling value we are trying to predict are values from a non-standard set of conditions (a different space time - perhaps a multi-verse). They are not measured - they were calculated. The same muti-verse supplied all the additional supplemental data - ie, it was calculated under the same set of parameters as the coupling values.\n\nSo values you might find in a reference are different because they are not from the same set of conditions - I assume temperature being the big one.", "votes": null }, { "id": "580410", "postDate": "07/20/2019 05:59:41", "content": "@pcjimmmy I was also assuming the same. Thanks for confirming! However the differences are almost double than the reference value in some cases. I hope the test set is also calculated keeping conditions of training set intact.
", "rawMarkdown": "pcjimmmy I was also assuming the same. Thanks for confirming! However the differences are almost double than the reference value in some cases. I hope the test set is also calculated keeping conditions of training set intact.", "votes": null }, { "id": "580506", "postDate": "07/20/2019 08:52:46", "content": "I am not a quantum chemist at all, just have some background in physics, therefore take the following with a (large) grain of salt.
\n\nIt is not the first time people in this forum notice differences between physical/chemical values here and common values found in publications. Every time the reason given is that the molecules we are looking at were simulated at 0 kelvin. As a result, values are different from the ones measured at usual temperature. For those not familiar with it, 0 Kelvin is about minus 273 celsius degrees (and I don't know how many farenheit). At 0 kelvin molecules and atoms aren't vibrating or rotating, they are fixed in space.
", "rawMarkdown": "I am not a quantum chemist at all, just have some background in physics, therefore take the following with a (large) grain of salt.\n\nIt is not the first time people in this forum notice differences between physical/chemical values here and common values found in publications. Every time the reason given is that the molecules we are looking at were simulated at 0 kelvin. As a result, values are different from the ones measured at usual temperature. For those not familiar with it, 0 Kelvin is about minus 273 celsius degrees (and I don't know how many farenheit). At 0 kelvin molecules and atoms aren't vibrating or rotating, they are fixed in space.", "votes": null }, { "id": "580950", "postDate": "07/21/2019 06:01:58", "content": "I also initially thought this, but as far as I could find, the dipole moment of a single molecule shouldn't be affected by temperature (the interactions between multiple molecules will definitely be affected as you mentioned). I guess maybe if the molecule has free rotors that might have an effect at 0K, but then why are the dipole moments for simple molecules like water and methane so different? More head-scratching to do! 🤔
", "rawMarkdown": "I also initially thought this, but as far as I could find, the dipole moment of a single molecule shouldn't be affected by temperature (the interactions between multiple molecules will definitely be affected as you mentioned). I guess maybe if the molecule has free rotors that might have an effect at 0K, but then why are the dipole moments for simple molecules like water and methane so different? More head-scratching to do! 🤔", "votes": null }, { "id": "580964", "postDate": "07/21/2019 06:49:36", "content": "Let's hope a chemist will enlighten us!
", "rawMarkdown": "Let's hope a chemist will enlighten us!", "votes": null }, { "id": "581391", "postDate": "07/21/2019 21:27:50", "content": "Your equation is correct to get the norm of a vector. You could use numpy also, which does the same formula by default: https://docs.scipy.org/doc/numpy/reference/generated/numpy.linalg.norm.html
\n\nWhen you say \"methane\" you mean ammonia, as methane has zero dipole moment. Ammonia might be a bit special, so comparing these QM simulations with real-world values could lead to discrepancy... I'd like to see if your dipole moments differ for bigger molecules too.
\n\nThe two values for water (1.8511 and 1.8546) are pretty close, probably within error for the simulation, so you can congratulate yourself there =)
", "rawMarkdown": "Your equation is correct to get the norm of a vector. You could use numpy also, which does the same formula by default: https://docs.scipy.org/doc/numpy/reference/generated/numpy.linalg.norm.html\n\nWhen you say \"methane\" you mean ammonia, as methane has zero dipole moment. Ammonia might be a bit special, so comparing these QM simulations with real-world values could lead to discrepancy... I'd like to see if your dipole moments differ for bigger molecules too.\n\nThe two values for water (1.8511 and 1.8546) are pretty close, probably within error for the simulation, so you can congratulate yourself there =)", "votes": null }, { "id": "581634", "postDate": "07/22/2019 07:31:51", "content": "@phucco thanks pointing out! molecule \"dsgdb9nsd_000002\" is actually NH3 (ammonia). I've updated the description.\nIn practice I always prefer numpy.linalg.norm for finding norm of vector or even distance between two vectors.
", "rawMarkdown": "phucco thanks pointing out! molecule \"dsgdb9nsd_000002\" is actually NH3 (ammonia). I've updated the description.\nIn practice I always prefer numpy.linalg.norm for finding norm of vector or even distance between two vectors.", "votes": null }, { "id": "582946", "postDate": "07/23/2019 20:29:24", "content": "The scalar coupling constant is independent of temperature first off. \nAnd, to answer the original question, did you factor in electronegativity difference between the atoms and the angles among the atoms in the molecule. Because if you didn't your dipole moment will be off.
", "rawMarkdown": "The scalar coupling constant is independent of temperature first off. \nAnd, to answer the original question, did you factor in electronegativity difference between the atoms and the angles among the atoms in the molecule. Because if you didn't your dipole moment will be off.", "votes": null }, { "id": "583023", "postDate": "07/23/2019 22:56:12", "content": "OK - not dependent on temperature - than what is it dependent on that the sponsor's are not worried that we can find the data sitting on a web site somewhere.
", "rawMarkdown": "OK - not dependent on temperature - than what is it dependent on that the sponsor's are not worried that we can find the data sitting on a web site somewhere.", "votes": null }, { "id": "583027", "postDate": "07/23/2019 23:08:56", "content": "I think it is independent of most things. If it were not, then, it would not be of much value to the host. It is supposed to give insight on the structure of the molecule and if many variables can change it, then it would not be able to give accurate insight on the structure of the molecule.
", "rawMarkdown": "I think it is independent of most things. If it were not, then, it would not be of much value to the host. It is supposed to give insight on the structure of the molecule and if many variables can change it, then it would not be able to give accurate insight on the structure of the molecule.", "votes": null }, { "id": "583032", "postDate": "07/23/2019 23:20:40", "content": "One of us is very confused about the nature of the scalar coupling - but guess its not totally relevant to the original question so might not be worth finding out which of us is the confused one. While a chemistry graduate from college - that was 50 years ago and the darn thing was barely invented at the time.
", "rawMarkdown": "One of us is very confused about the nature of the scalar coupling - but guess its not totally relevant to the original question so might not be worth finding out which of us is the confused one. While a chemistry graduate from college - that was 50 years ago and the darn thing was barely invented at the time.", "votes": null }, { "id": "583170", "postDate": "07/24/2019 06:01:10", "content": "It's just because the whole QM9 dataset has been calculated using an approximation of the Schrodinger equation (called hybrid DFT), so every small (or large, for that matter) difference between QM9 and experimental values just stems from our not solving the exact equation.
\n\nDipole moments are not so easy to calculate in a full theoretical manner. They require knowing the charge of the atom (called the Born effective charge), which can be calculated using perturbation theory, and which is linked to the response of the system wrt a small displacement of each atom, in each direction. These kind of calculations are not straightforward at all.
\n\nWe could of course use the experimental values of the atom charge instead of these complex calculations, but it wouldn't be consistent to mix the experimental value of something with other theoretical quantities, we prefer to have a full 100% theoretical model
", "rawMarkdown": "It's just because the whole QM9 dataset has been calculated using an approximation of the Schrodinger equation (called hybrid DFT), so every small (or large, for that matter) difference between QM9 and experimental values just stems from our not solving the exact equation.\n\nDipole moments are not so easy to calculate in a full theoretical manner. They require knowing the charge of the atom (called the Born effective charge), which can be calculated using perturbation theory, and which is linked to the response of the system wrt a small displacement of each atom, in each direction. These kind of calculations are not straightforward at all.\n\nWe could of course use the experimental values of the atom charge instead of these complex calculations, but it wouldn't be consistent to mix the experimental value of something with other theoretical quantities, we prefer to have a full 100% theoretical model", "votes": null }, { "id": "583217", "postDate": "07/24/2019 07:06:21", "content": "Thanks Boris! I've been using OpenBabel to calculate dipole moments and I've found that different charge models can give wildly different values. I guess as long as they correlate with the QM data then it might not be much of an issue
", "rawMarkdown": "Thanks Boris! I've been using OpenBabel to calculate dipole moments and I've found that different charge models can give wildly different values. I guess as long as they correlate with the QM data then it might not be much of an issue", "votes": null } ], "comments": [ { "id": 580273, "author_name": "pcjimmmy", "author_url": "", "post_date": "07/19/2019 23:30:11", "content": "If I am looking at the right molecule I believe I got the same answer as you for methane and water. Guess that could mean we both goofed the same way.
\n\nI like to think I did not - here's what I believe. The scalar coupling value we are trying to predict are values from a non-standard set of conditions (a different space time - perhaps a multi-verse). They are not measured - they were calculated. The same muti-verse supplied all the additional supplemental data - ie, it was calculated under the same set of parameters as the coupling values.
\n\nSo values you might find in a reference are different because they are not from the same set of conditions - I assume temperature being the big one.
", "votes": null, "replies": [ { "id": 580410, "author_name": "rajparmar", "author_url": "", "post_date": "07/20/2019 05:59:41", "content": "@pcjimmmy I was also assuming the same. Thanks for confirming! However the differences are almost double than the reference value in some cases. I hope the test set is also calculated keeping conditions of training set intact.
", "votes": null, "replies": [] }, { "id": 582946, "author_name": "petersk20", "author_url": "", "post_date": "07/23/2019 20:29:24", "content": "The scalar coupling constant is independent of temperature first off. \nAnd, to answer the original question, did you factor in electronegativity difference between the atoms and the angles among the atoms in the molecule. Because if you didn't your dipole moment will be off.
", "votes": null, "replies": [] }, { "id": 583023, "author_name": "pcjimmmy", "author_url": "", "post_date": "07/23/2019 22:56:12", "content": "OK - not dependent on temperature - than what is it dependent on that the sponsor's are not worried that we can find the data sitting on a web site somewhere.
", "votes": null, "replies": [] }, { "id": 583027, "author_name": "petersk20", "author_url": "", "post_date": "07/23/2019 23:08:56", "content": "I think it is independent of most things. If it were not, then, it would not be of much value to the host. It is supposed to give insight on the structure of the molecule and if many variables can change it, then it would not be able to give accurate insight on the structure of the molecule.
", "votes": null, "replies": [] }, { "id": 583032, "author_name": "pcjimmmy", "author_url": "", "post_date": "07/23/2019 23:20:40", "content": "One of us is very confused about the nature of the scalar coupling - but guess its not totally relevant to the original question so might not be worth finding out which of us is the confused one. While a chemistry graduate from college - that was 50 years ago and the darn thing was barely invented at the time.
", "votes": null, "replies": [] } ] }, { "id": 580506, "author_name": "cpmpml", "author_url": "", "post_date": "07/20/2019 08:52:46", "content": "I am not a quantum chemist at all, just have some background in physics, therefore take the following with a (large) grain of salt.
\n\nIt is not the first time people in this forum notice differences between physical/chemical values here and common values found in publications. Every time the reason given is that the molecules we are looking at were simulated at 0 kelvin. As a result, values are different from the ones measured at usual temperature. For those not familiar with it, 0 Kelvin is about minus 273 celsius degrees (and I don't know how many farenheit). At 0 kelvin molecules and atoms aren't vibrating or rotating, they are fixed in space.
", "votes": null, "replies": [ { "id": 580950, "author_name": "anjum48", "author_url": "", "post_date": "07/21/2019 06:01:58", "content": "I also initially thought this, but as far as I could find, the dipole moment of a single molecule shouldn't be affected by temperature (the interactions between multiple molecules will definitely be affected as you mentioned). I guess maybe if the molecule has free rotors that might have an effect at 0K, but then why are the dipole moments for simple molecules like water and methane so different? More head-scratching to do! 🤔
", "votes": null, "replies": [] }, { "id": 580964, "author_name": "cpmpml", "author_url": "", "post_date": "07/21/2019 06:49:36", "content": "Let's hope a chemist will enlighten us!
", "votes": null, "replies": [] }, { "id": 583170, "author_name": "borisdee", "author_url": "", "post_date": "07/24/2019 06:01:10", "content": "It's just because the whole QM9 dataset has been calculated using an approximation of the Schrodinger equation (called hybrid DFT), so every small (or large, for that matter) difference between QM9 and experimental values just stems from our not solving the exact equation.
\n\nDipole moments are not so easy to calculate in a full theoretical manner. They require knowing the charge of the atom (called the Born effective charge), which can be calculated using perturbation theory, and which is linked to the response of the system wrt a small displacement of each atom, in each direction. These kind of calculations are not straightforward at all.
\n\nWe could of course use the experimental values of the atom charge instead of these complex calculations, but it wouldn't be consistent to mix the experimental value of something with other theoretical quantities, we prefer to have a full 100% theoretical model
", "votes": null, "replies": [] }, { "id": 583217, "author_name": "anjum48", "author_url": "", "post_date": "07/24/2019 07:06:21", "content": "Thanks Boris! I've been using OpenBabel to calculate dipole moments and I've found that different charge models can give wildly different values. I guess as long as they correlate with the QM data then it might not be much of an issue
", "votes": null, "replies": [] } ] }, { "id": 581391, "author_name": "phucco", "author_url": "", "post_date": "07/21/2019 21:27:50", "content": "Your equation is correct to get the norm of a vector. You could use numpy also, which does the same formula by default: https://docs.scipy.org/doc/numpy/reference/generated/numpy.linalg.norm.html
\n\nWhen you say \"methane\" you mean ammonia, as methane has zero dipole moment. Ammonia might be a bit special, so comparing these QM simulations with real-world values could lead to discrepancy... I'd like to see if your dipole moments differ for bigger molecules too.
\n\nThe two values for water (1.8511 and 1.8546) are pretty close, probably within error for the simulation, so you can congratulate yourself there =)
", "votes": null, "replies": [ { "id": 581634, "author_name": "rajparmar", "author_url": "", "post_date": "07/22/2019 07:31:51", "content": "@phucco thanks pointing out! molecule \"dsgdb9nsd_000002\" is actually NH3 (ammonia). I've updated the description.\nIn practice I always prefer numpy.linalg.norm for finding norm of vector or even distance between two vectors.
", "votes": null, "replies": [] } ] } ], "raw_markdown_by_id": { "580221": "I am trying to convert the dipole moment vector of csv files from x, y, z coordinates to a scalar value using formula: Square root of (x^2 + y^2 + z^2)\nThis formula does not get affected by Cartesian system.\n\nNow using this formula I am getting magnitude of dipole moment for ammonia as (dsgdb9nsd_000002) as Sqrt(-0.0002^2 + 0^2 + 1.6256^2) = 1.6256\nHowever standard value of magnitude of dipole moment for methane is 1.46 which I found on multiple pages (for example https://www.quora.com/Whose-dipole-moment-is-greater-NH3-or-NF3)\n\nSimilar is the case with water (getting 1.8511 instead of 1.8546) and almost all the structures that I tested.\n\nCan anyone suggest where am I going wrong here. Is it the formula I am using?", "580273": "If I am looking at the right molecule I believe I got the same answer as you for methane and water. Guess that could mean we both goofed the same way.\n\nI like to think I did not - here's what I believe. The scalar coupling value we are trying to predict are values from a non-standard set of conditions (a different space time - perhaps a multi-verse). They are not measured - they were calculated. The same muti-verse supplied all the additional supplemental data - ie, it was calculated under the same set of parameters as the coupling values.\n\nSo values you might find in a reference are different because they are not from the same set of conditions - I assume temperature being the big one.", "580410": "pcjimmmy I was also assuming the same. Thanks for confirming! However the differences are almost double than the reference value in some cases. I hope the test set is also calculated keeping conditions of training set intact.", "580506": "I am not a quantum chemist at all, just have some background in physics, therefore take the following with a (large) grain of salt.\n\nIt is not the first time people in this forum notice differences between physical/chemical values here and common values found in publications. Every time the reason given is that the molecules we are looking at were simulated at 0 kelvin. As a result, values are different from the ones measured at usual temperature. For those not familiar with it, 0 Kelvin is about minus 273 celsius degrees (and I don't know how many farenheit). At 0 kelvin molecules and atoms aren't vibrating or rotating, they are fixed in space.", "580950": "I also initially thought this, but as far as I could find, the dipole moment of a single molecule shouldn't be affected by temperature (the interactions between multiple molecules will definitely be affected as you mentioned). I guess maybe if the molecule has free rotors that might have an effect at 0K, but then why are the dipole moments for simple molecules like water and methane so different? More head-scratching to do! 🤔", "580964": "Let's hope a chemist will enlighten us!", "581391": "Your equation is correct to get the norm of a vector. You could use numpy also, which does the same formula by default: https://docs.scipy.org/doc/numpy/reference/generated/numpy.linalg.norm.html\n\nWhen you say \"methane\" you mean ammonia, as methane has zero dipole moment. Ammonia might be a bit special, so comparing these QM simulations with real-world values could lead to discrepancy... I'd like to see if your dipole moments differ for bigger molecules too.\n\nThe two values for water (1.8511 and 1.8546) are pretty close, probably within error for the simulation, so you can congratulate yourself there =)", "581634": "phucco thanks pointing out! molecule \"dsgdb9nsd_000002\" is actually NH3 (ammonia). I've updated the description.\nIn practice I always prefer numpy.linalg.norm for finding norm of vector or even distance between two vectors.", "582946": "The scalar coupling constant is independent of temperature first off. \nAnd, to answer the original question, did you factor in electronegativity difference between the atoms and the angles among the atoms in the molecule. Because if you didn't your dipole moment will be off.", "583023": "OK - not dependent on temperature - than what is it dependent on that the sponsor's are not worried that we can find the data sitting on a web site somewhere.", "583027": "I think it is independent of most things. If it were not, then, it would not be of much value to the host. It is supposed to give insight on the structure of the molecule and if many variables can change it, then it would not be able to give accurate insight on the structure of the molecule.", "583032": "One of us is very confused about the nature of the scalar coupling - but guess its not totally relevant to the original question so might not be worth finding out which of us is the confused one. While a chemistry graduate from college - that was 50 years ago and the darn thing was barely invented at the time.", "583170": "It's just because the whole QM9 dataset has been calculated using an approximation of the Schrodinger equation (called hybrid DFT), so every small (or large, for that matter) difference between QM9 and experimental values just stems from our not solving the exact equation.\n\nDipole moments are not so easy to calculate in a full theoretical manner. They require knowing the charge of the atom (called the Born effective charge), which can be calculated using perturbation theory, and which is linked to the response of the system wrt a small displacement of each atom, in each direction. These kind of calculations are not straightforward at all.\n\nWe could of course use the experimental values of the atom charge instead of these complex calculations, but it wouldn't be consistent to mix the experimental value of something with other theoretical quantities, we prefer to have a full 100% theoretical model", "583217": "Thanks Boris! I've been using OpenBabel to calculate dipole moments and I've found that different charge models can give wildly different values. I guess as long as they correlate with the QM data then it might not be much of an issue" }, "source": "meta" }