Difference of iC4H10 in PyBoltz when compared to Magboltz v11.9

[PhilipHB]

We did a comparison of isoButane simulated with PyBoltz @ d6f3dbc and Magboltz v11.9 and found some deviation.

For fields exceeding ~7500V/cm some transport parameters start to systematically take off from the Magboltz "baseline". Longitudinal diffusion and drift velocity deviate, but transverse diffusion and first Townsend coefficient do not.

Is this expected behaviour in some way? From the magboltz source code (changelog at the top) I gather iC4H10 has not been touched since v10.5

If we can provide any more information, just let us know.

Pictures:

Configuration of our PyBoltz run:

        # Configure settings for our simulation
        [...]
        Boltzer.MaxNumberOfCollisions = 50 * 1e7
        Boltzer.Enable_Penning = 0
        Boltzer.Enable_Thermal_Motion = 1
        Boltzer.Max_Electron_Energy = 0
        Boltzer.TemperatureCentigrade = 298
        Boltzer.Pressure_Torr = 750.1
        [...]
        Boltzer.BField_Mag = 0
        Boltzer.BField_Angle = 0
        Boltzer.Console_Output_Flag = 1
        Boltzer.Steady_State_Threshold = 40
        Boltzer.Which_Angular_Model = 2

[benjpjones]

Hi Philip, Hmm, no I wold not expect this. It is very surprising to me to think that there is some effect that can influence some of the transport parameters like this and not others. Since its restricted to this gas I can only imagine its a cross sections issue. I’ll try to take a look today. Ben ===== Ben Jones [he/his/him] Asst Prof of Physics University of Texas at Arlington ***@***.*** =====

…

On Jun 17, 2021, at 6:24 AM, PhilipHB ***@***.***> wrote: We did a comparison of isoButane simulated with PyBoltz @ d6f3dbc <https://nam12.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgithub.com%2FUTA-REST%2FPyBoltz%2Fcommit%2Fd6f3dbc6d60c584edd5079cf60f86e8982fcf5b5&data=04%7C01%7Cben.jones%40uta.edu%7C050e4e1968694793f24708d931828048%7C5cdc5b43d7be4caa8173729e3b0a62d9%7C1%7C0%7C637595258831286968%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C1000&sdata=WNS%2BZJd6trS5xZs62Prmoa56LtiJyoY2LkCXMl4Yag8%3D&reserved=0> and Magboltz v11.9 and found some deviation. For fields exceeding ~7500V/cm some transport parameters start to systematically take off from the Magboltz "baseline". Longitudinal diffusion and drift velocity deviate, but transverse diffusion and first Townsend coefficient do not. Is this expected behaviour in some way? From the magboltz source code (changelog at the top) I gather iC4H10 has not been touched since v10.5 If we can provide any more information, just let us know. Pictures: <https://nam12.safelinks.protection.outlook.com/?url=https%3A%2F%2Fuser-images.githubusercontent.com%2F36764986%2F122385524-37d4a600-cf6d-11eb-908c-dc9b1c9731b0.png&data=04%7C01%7Cben.jones%40uta.edu%7C050e4e1968694793f24708d931828048%7C5cdc5b43d7be4caa8173729e3b0a62d9%7C1%7C0%7C637595258831286968%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C1000&sdata=xZ4oUxHcTPbQC5JRkOKDg3YapHMsF1st%2F1Tnk8xrluw%3D&reserved=0> <https://nam12.safelinks.protection.outlook.com/?url=https%3A%2F%2Fuser-images.githubusercontent.com%2F36764986%2F122385548-402ce100-cf6d-11eb-9730-6ef5289447c9.png&data=04%7C01%7Cben.jones%40uta.edu%7C050e4e1968694793f24708d931828048%7C5cdc5b43d7be4caa8173729e3b0a62d9%7C1%7C0%7C637595258831296963%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C1000&sdata=oPUJdbfTbvR6xvcDATUDcB%2FDxrkFW4tU%2FqXp06fxb3k%3D&reserved=0> <https://nam12.safelinks.protection.outlook.com/?url=https%3A%2F%2Fuser-images.githubusercontent.com%2F36764986%2F122385554-428f3b00-cf6d-11eb-889d-3d63a7f393e3.png&data=04%7C01%7Cben.jones%40uta.edu%7C050e4e1968694793f24708d931828048%7C5cdc5b43d7be4caa8173729e3b0a62d9%7C1%7C0%7C637595258831296963%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C1000&sdata=scg2kJE0YqXFMmdvtCTZAJQP9yuWJ6p%2Fke0fzyMM4mw%3D&reserved=0> <https://nam12.safelinks.protection.outlook.com/?url=https%3A%2F%2Fuser-images.githubusercontent.com%2F36764986%2F122385704-681c4480-cf6d-11eb-91e6-16d063ca79c8.png&data=04%7C01%7Cben.jones%40uta.edu%7C050e4e1968694793f24708d931828048%7C5cdc5b43d7be4caa8173729e3b0a62d9%7C1%7C0%7C637595258831306961%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C1000&sdata=Jf8e6hu%2BfEJnJ02VDvRVbwmOolCQTfUk%2BENVf5xgcVs%3D&reserved=0> Configuration of our PyBoltz run: # Configure settings for our simulation [...] Boltzer.MaxNumberOfCollisions = 50 * 1e7 Boltzer.Enable_Penning = 0 Boltzer.Enable_Thermal_Motion = 1 Boltzer.Max_Electron_Energy = 0 Boltzer.TemperatureCentigrade = 298 Boltzer.Pressure_Torr = 750.1 [...] Boltzer.BField_Mag = 0 Boltzer.BField_Angle = 0 Boltzer.Console_Output_Flag = 1 Boltzer.Steady_State_Threshold = 40 Boltzer.Which_Angular_Model = 2 — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub <https://nam12.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgithub.com%2FUTA-REST%2FPyBoltz%2Fissues%2F17&data=04%7C01%7Cben.jones%40uta.edu%7C050e4e1968694793f24708d931828048%7C5cdc5b43d7be4caa8173729e3b0a62d9%7C1%7C0%7C637595258831306961%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C1000&sdata=rBnqiOlVBAwm53%2BI1G1ThyZgfWaUMqAGF7UHN6r40x8%3D&reserved=0>, or unsubscribe <https://nam12.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgithub.com%2Fnotifications%2Funsubscribe-auth%2FACVWB2SMKV2TAIHDGX3SIL3TTHLPRANCNFSM463OZRZQ&data=04%7C01%7Cben.jones%40uta.edu%7C050e4e1968694793f24708d931828048%7C5cdc5b43d7be4caa8173729e3b0a62d9%7C1%7C0%7C637595258831506846%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C1000&sdata=1hfDEk9o6XaXMq%2B8pinjvJy9fCJR9GB6cQJwQv1wOPM%3D&reserved=0>.

[benjpjones]

Hi Philip, I wonder, is it possible you are using MagBoltz with isotropic angular distribution model, rather than Okkhrimovsky, as PyBoltz is set to in the example below? I see that when I run PyB with AngularDistModel 0 I get very close to your MagBoltz curve (I was trying to think of things that might disproportionately affect v, and this one came to mind). Below plot, two lines are yours (plot-clicked, forgive the imprecision), data points are my runs from PyBoltz with the different angular models. If I recall right the angular model setting is buried in the FORTRAN for MagBoltz, not an external parameter as in PyBoltz. Okkhrimovsky (which I probably misspell) is often considered the best option. Refs to those models are in the PyBoltz paper if you need them. Ben ===== Ben Jones [he/his/him] Asst Prof of Physics University of Texas at Arlington ***@***.*** =====

…

On Jun 17, 2021, at 6:24 AM, PhilipHB ***@***.***> wrote: We did a comparison of isoButane simulated with PyBoltz @ d6f3dbc <https://nam12.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgithub.com%2FUTA-REST%2FPyBoltz%2Fcommit%2Fd6f3dbc6d60c584edd5079cf60f86e8982fcf5b5&data=04%7C01%7Cben.jones%40uta.edu%7C050e4e1968694793f24708d931828048%7C5cdc5b43d7be4caa8173729e3b0a62d9%7C1%7C0%7C637595258831286968%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C1000&sdata=WNS%2BZJd6trS5xZs62Prmoa56LtiJyoY2LkCXMl4Yag8%3D&reserved=0> and Magboltz v11.9 and found some deviation. For fields exceeding ~7500V/cm some transport parameters start to systematically take off from the Magboltz "baseline". Longitudinal diffusion and drift velocity deviate, but transverse diffusion and first Townsend coefficient do not. Is this expected behaviour in some way? From the magboltz source code (changelog at the top) I gather iC4H10 has not been touched since v10.5 If we can provide any more information, just let us know. Pictures: <https://nam12.safelinks.protection.outlook.com/?url=https%3A%2F%2Fuser-images.githubusercontent.com%2F36764986%2F122385524-37d4a600-cf6d-11eb-908c-dc9b1c9731b0.png&data=04%7C01%7Cben.jones%40uta.edu%7C050e4e1968694793f24708d931828048%7C5cdc5b43d7be4caa8173729e3b0a62d9%7C1%7C0%7C637595258831286968%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C1000&sdata=xZ4oUxHcTPbQC5JRkOKDg3YapHMsF1st%2F1Tnk8xrluw%3D&reserved=0> <https://nam12.safelinks.protection.outlook.com/?url=https%3A%2F%2Fuser-images.githubusercontent.com%2F36764986%2F122385548-402ce100-cf6d-11eb-9730-6ef5289447c9.png&data=04%7C01%7Cben.jones%40uta.edu%7C050e4e1968694793f24708d931828048%7C5cdc5b43d7be4caa8173729e3b0a62d9%7C1%7C0%7C637595258831296963%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C1000&sdata=oPUJdbfTbvR6xvcDATUDcB%2FDxrkFW4tU%2FqXp06fxb3k%3D&reserved=0> <https://nam12.safelinks.protection.outlook.com/?url=https%3A%2F%2Fuser-images.githubusercontent.com%2F36764986%2F122385554-428f3b00-cf6d-11eb-889d-3d63a7f393e3.png&data=04%7C01%7Cben.jones%40uta.edu%7C050e4e1968694793f24708d931828048%7C5cdc5b43d7be4caa8173729e3b0a62d9%7C1%7C0%7C637595258831296963%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C1000&sdata=scg2kJE0YqXFMmdvtCTZAJQP9yuWJ6p%2Fke0fzyMM4mw%3D&reserved=0> <https://nam12.safelinks.protection.outlook.com/?url=https%3A%2F%2Fuser-images.githubusercontent.com%2F36764986%2F122385704-681c4480-cf6d-11eb-91e6-16d063ca79c8.png&data=04%7C01%7Cben.jones%40uta.edu%7C050e4e1968694793f24708d931828048%7C5cdc5b43d7be4caa8173729e3b0a62d9%7C1%7C0%7C637595258831306961%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C1000&sdata=Jf8e6hu%2BfEJnJ02VDvRVbwmOolCQTfUk%2BENVf5xgcVs%3D&reserved=0> Configuration of our PyBoltz run: # Configure settings for our simulation [...] Boltzer.MaxNumberOfCollisions = 50 * 1e7 Boltzer.Enable_Penning = 0 Boltzer.Enable_Thermal_Motion = 1 Boltzer.Max_Electron_Energy = 0 Boltzer.TemperatureCentigrade = 298 Boltzer.Pressure_Torr = 750.1 [...] Boltzer.BField_Mag = 0 Boltzer.BField_Angle = 0 Boltzer.Console_Output_Flag = 1 Boltzer.Steady_State_Threshold = 40 Boltzer.Which_Angular_Model = 2 — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub <https://nam12.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgithub.com%2FUTA-REST%2FPyBoltz%2Fissues%2F17&data=04%7C01%7Cben.jones%40uta.edu%7C050e4e1968694793f24708d931828048%7C5cdc5b43d7be4caa8173729e3b0a62d9%7C1%7C0%7C637595258831306961%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C1000&sdata=rBnqiOlVBAwm53%2BI1G1ThyZgfWaUMqAGF7UHN6r40x8%3D&reserved=0>, or unsubscribe <https://nam12.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgithub.com%2Fnotifications%2Funsubscribe-auth%2FACVWB2SMKV2TAIHDGX3SIL3TTHLPRANCNFSM463OZRZQ&data=04%7C01%7Cben.jones%40uta.edu%7C050e4e1968694793f24708d931828048%7C5cdc5b43d7be4caa8173729e3b0a62d9%7C1%7C0%7C637595258831506846%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C1000&sdata=1hfDEk9o6XaXMq%2B8pinjvJy9fCJR9GB6cQJwQv1wOPM%3D&reserved=0>.

[PhilipHB]

According to our Maboltz run logs, the model was:
ANISOTROPIC SCATTERING TYPE 2 (OKHRIMOVSKYY) USED IF AVAILABLE
I can not tell if it is available, though.

[PhilipHB]

We do not see any deviation up to our current simulation maximum of ~60kV/cm in Ar, CH4 or C3H8.

[PhilipHB]

A comparison of all available angular models in PyBoltz with the Magboltz simulation - for iC4H10 only. To me, it does not look like there's a problem with selecting the angular model at least. In my understanding, at least one of these curves should line up nicely.

Let me know if there's anything that could help investigating this.

[benjpjones]

Where is orange on the top fig? It looks like that one might match pretty well?

…

Sent from my iPhone On Jun 30, 2021, at 8:43 AM, PhilipHB ***@***.***> wrote:  A comparison of all available angular models in PyBoltz with the Magboltz simulation - for iC4H10 only. To me, it does not look like there's a problem with selecting the angular model at least. In my understanding, at least one of these curves should line up nicely. Let me know if there's anything that could help investigating this. [Screenshot from 2021-06-30 15-28-46]<https://nam12.safelinks.protection.outlook.com/?url=https%3A%2F%2Fuser-images.githubusercontent.com%2F36764986%2F123969262-58aee980-d9b8-11eb-9acc-73b4fe56dff9.png&data=04%7C01%7Cben.jones%40uta.edu%7C37953e501b924967a59608d93bcd0fde%7C5cdc5b43d7be4caa8173729e3b0a62d9%7C1%7C0%7C637606574183168963%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C1000&sdata=DB8%2F55lxSHvYgFy%2BDHm4lymIjiRozqPd%2FCxN%2B%2B2JpEY%3D&reserved=0> [Screenshot from 2021-06-30 15-29-06]<https://nam12.safelinks.protection.outlook.com/?url=https%3A%2F%2Fuser-images.githubusercontent.com%2F36764986%2F123969276-5c427080-d9b8-11eb-8ec5-44bf6c2b1f4c.png&data=04%7C01%7Cben.jones%40uta.edu%7C37953e501b924967a59608d93bcd0fde%7C5cdc5b43d7be4caa8173729e3b0a62d9%7C1%7C0%7C637606574183178958%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C1000&sdata=%2FF6vA7y5oNa71czXUUKVecqN8Dhe2n6b2nko9vZOdTs%3D&reserved=0> [Screenshot from 2021-06-30 15-29-32]<https://nam12.safelinks.protection.outlook.com/?url=https%3A%2F%2Fuser-images.githubusercontent.com%2F36764986%2F123969293-606e8e00-d9b8-11eb-9ecd-916a1b7a531e.png&data=04%7C01%7Cben.jones%40uta.edu%7C37953e501b924967a59608d93bcd0fde%7C5cdc5b43d7be4caa8173729e3b0a62d9%7C1%7C0%7C637606574183188953%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C1000&sdata=7f%2BC4ghq7qnTcX2gNNKrdhmv9%2BJXYku%2BMPY14haJQ2s%3D&reserved=0> — You are receiving this because you commented. Reply to this email directly, view it on GitHub<https://nam12.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgithub.com%2FUTA-REST%2FPyBoltz%2Fissues%2F17%23issuecomment-871416532&data=04%7C01%7Cben.jones%40uta.edu%7C37953e501b924967a59608d93bcd0fde%7C5cdc5b43d7be4caa8173729e3b0a62d9%7C1%7C0%7C637606574183188953%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C1000&sdata=L4AVM0XHzGo3GIzHJzbmqoh2%2Bx6ZNO995sTiWz0hNig%3D&reserved=0>, or unsubscribe<https://nam12.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgithub.com%2Fnotifications%2Funsubscribe-auth%2FACVWB2UXFS2Y322Q3FI3XS3TVMNQRANCNFSM463OZRZQ&data=04%7C01%7Cben.jones%40uta.edu%7C37953e501b924967a59608d93bcd0fde%7C5cdc5b43d7be4caa8173729e3b0a62d9%7C1%7C0%7C637606574183198953%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C1000&sdata=xYzwx56rxYdmpvrsvmWGeAJCzWDgexS1IysQjaBK0tY%3D&reserved=0>.

[PhilipHB]

Looks like models 0 and 1 predict essentially the same vd. A zoomed in version reveals that a bit.