08:40:11 Okay, Welcome back everyone.
08:40:15 And we have excellent Evan, keeping with the themes. Very good.
08:40:23 Evan Schneider, Dr. Evan Schneider is going to talk to us about mass loading. I'm particularly excited about this one, and she's doing so as a cat.
08:40:33 So please welcome Evan Schneider. Oh I am I am so sorry. Sorry about that. Don't know how that happened.
08:40:46 Okay, can we all, you can see me now, great, great, great.
08:40:50 Um, so right I was tasked with telling you about mass loading and galactic outflows, and I'm trying to keep this sort of tutorial Lee, and less talk, he.
08:41:10 But we'll see, we'll see how that turns out. I didn't need an assistant to turn it off.
08:41:13 Okay, so I made myself a little outline to try to keep us on track here so first try to present some context, even in a tutorial I can't give a talk without giving some context.
08:41:23 And then some basic definitions for masturbating I think a lot of the confusion around this particular topic comes from the fact that this is not always defined, the same way.
08:41:33 Then I'll spend a little bit of time talking about how we measure mass loading, with an emphasis on simulations, since that's what I'm most familiar with.
08:41:42 I'll talk a little bit about measurement and observations, but I'm sure we're going to hear a lot more about that from from Tim tomorrow so I won't spend too much time there and then I'll end if we have time with a little bit on, you know, additional
08:41:54 loading parameters that I think can also be very helpful and trying to understand the sort of broad topic.
08:42:02 Okay, so for some context. This, this slide is all you need.
08:42:08 The point being, right we've seen this several times. At this point, but what we're interested in is, you know, how do galaxies remove their gaps. And we're interested in that, because we think that the process of removing gas from galaxies or at the
08:42:25 very least removing gas from halos is an important one in galaxy formation, and, you know, perhaps is responsible for setting. Many of the properties of galaxies that we see today.
08:42:39 And so, you know, one of the ways that we try to parameter eyes. This is by measuring mass outflow rates and associated with that this parameter that we call mass loading.
08:42:53 Okay, so what is mass loading, going from, I guess, terminology that you might be more familiar with the terminology that perhaps are less familiar with most commonly in our field I think mass loading is defined as just, it's a way to measure how much
08:43:09 mass is leaving a galaxy, relative to its star formation right that's that's the most common definition I think that you'll see.
08:43:17 But in the wind community is also very commonly defined as a way to measure how much additional mass has been added to a wind, typically we're talking about the low density phase of a wind, relative to what you might expect to see from, say, the byproducts
08:43:33 of stellar evolution supernova Jacques de stellar wins.
08:43:38 When I was starting this presentation I decided to go back and see if I could figure out like when this term mass loading entered the the galaxy wind literature and actually I found that it was really prevalent in the planetary science and plasma physics
08:43:54 community.
08:43:55 So it starts showing up in abs, not in the context of galactic winds at all but you know to describe how for example, a comet that's outgassing might be adding mass to the, you know, to the solar wind plasma and how that might changes properties.
08:44:10 So suffice it to say that what mass loading is depends very much on the context in which you hear the term.
08:44:17 So that's something to keep in mind.
08:44:21 I'll talk just quickly about those first two two scenarios because this is something that confused me a lot as a graduate student when I was starting to read these papers.
08:44:30 So scenario. What I said is I think the most common mass loading as a way to measure how much mass is leaving a galaxy relative to its star formation right.
08:44:39 So this is the typical formula you'll see Can you all see my cursor. Yeah, great.
08:44:45 Right so m dot wind just meaning the mass outflow rate in the wind that you're measuring is equal to some factor here I'm calling it eight times, the M dot star formation rate.
08:44:59 So just the masses formed in stars as a function of time. And this parameter ADA is what we call the mass loading factor so often, you know, we will present values for ADA, rather than just say outflow rates.
08:45:15 And it's worth noting that Ada in this context can have any value, right. So, if you have, you know, high star formation rate and not very much mass in the wind, you can have a very small value for ADA, in a sense, you could even have a negative value
08:45:31 for Ada right if you have an inflow. And so there's nothing particularly physically meaningful in this parameter Ada it just relates these two quantities.
08:45:42 Okay scenario to that I said you know comes up a lot in the winds community is, is, you know, how, how much additional NASA has been added to a wind, relative to these byproducts of stellar evolution I attribute this to the fact that this is actually
08:45:57 the way the mass limiting factor is defined in the in the famous Scalia and Clegg, and paper.
08:46:04 Right, so there The idea was you're measuring mm dot wind say in the hot plasma, the you know, the very low density wind phase, which, for the record is the one that we are the most certain is actually a wind test high velocity use supersonic, you know,
08:46:22 will actually escape.
08:46:24 actually escape. And so there you know the mass loading factor which I'm going to call beta for this presentation just to just, you know, clarify between the two is, you know, is a factor that scales between the mass that you see in that when say the
08:46:39 hot plasma, and the mass injection rate from these stellar processes so commonly supernova that you could also include stellar wins, etc. And so something to note right is that they're actually beta now has a sort of a more physical meaning.
08:46:55 For example, it's typically greater than one.
08:46:58 You can only add mass into this face.
08:47:02 So that would reflect, for example, in training, additional is in material into the hot wind.
08:47:09 But it's also worth noting that you know in order to measure beta, you have to make an assumption about the right side of this equation. Right, so you're making an assumption about the mass injection rate into the wind from these underlying solar processes.
08:47:25 So I think, you know what, what I want to highlight here right is that it's, it's important to actually think about what you are trying to get out when you are measuring and looting factor because that may determine what kind of mass limiting factor you
08:47:38 are you are interested in.
08:47:41 Okay, so this is a tutorial so I decided to give a quick example to, to, kind of, you know like highlight the difference between these two actually I think near the school answer your question.
08:47:54 So here's an example that just kind of shows you the difference between say beta and ADA, and these two definitions. Let's say you have a Star Wars 99 model, it gives a constant rate of mass loss or mass injection however you think of it.
08:48:09 After 20 million years, and that rate scales with the star formation rate with the scaling so m dot star here you can just think of this as supernova plus stellar wins whatever goes is approximately point two five times the star formation rate and solar
08:48:24 masses per year. Okay, so you can imagine here if you if you, you know, this is what you would get if he ran like a cluster model, you can put whatever mass you want for the cluster in here this will give you the mass injection right from that cluster.
08:48:38 Now let's say you have, you know, some region with a star formation rate you think you've measured say five solar masses per year and the nuclear region of him at two.
08:48:48 And, you know, measure or fit for a value of beta, in the hot plasma, and you get to.
08:48:55 Okay, then. How would you convert that to ADA. In this context, this is a complicated problem, I just again wanted to highlight that these are two different measurements and they have two different values.
08:49:06 Right. So, if we know beta is to that's essentially just saying that the mass that we measure in the hot plasma is twice what we would expect, given the, you know, given the star formation rate and adjust the objective from from supernova installer wins.
08:49:23 wins. So we get a total mass outflow rate in that case of 2.5, solar masses per year, which then, you know, if you convert it to our more traditional definition Ada here, you would get point five.
08:49:34 Right, so in this case the scaling between beta and ADA is just getting by this this factor of point two five in front of the star formation right.
08:49:41 Great. And in this case also note right the physics comes in here is the Star Wars 99 model that's giving the relationship between beta and ADA, in this case.
08:49:54 Okay.
08:49:52 Just a warning.
08:49:54 We have not settled on the symbols that all in the literature. So, we will often use the symbol, beta, when we actually mean what I just called ADA. So just watch out, right, just make sure when you're reading a paper, you've read the definition of the
08:50:10 mass limiting factor before you make assumptions, you know, a factor of four might not sound like much but you know if you start to really scale that you could you could easily be off by an order of magnitude if you've misunderstood what was actually
08:50:26 being refined. Okay, so to summarize what is mass loading mass loading is just a scaling between some measured m dot, and something else. Also in units of mass per time, and I you know I want to say like, I actually really like mass loading as a concept
08:50:44 I think when you normalize it thoughtfully.
08:50:47 For example, as in the case of the beta example, it can give you some actual physical insight into, you know what, what is the physical phenomena that's leading to this mass load and.
08:50:58 But as Dylan pointed out rightfully in his in his keynote yesterday. If you normalize it carelessly or you're not really thinking about what it means.
08:51:09 Then it can lead you to some very confusing conclusions about about the driving business so right as stone pointed out, if you at, you know, late times are looking at winds that are primarily driven by supernova one a, but you're scaling your mass loading
08:51:23 factor by the star formation right, this could this could lead you to some, some confusing conclusions, so worth again pointing out just just when you're using mass loading it's useful to think about what what it is that you're trying to trying to understand.
08:51:40 Okay, so now on to how we actually measure mass loading in in simulations first So in principle, this is, this is very straightforward right, this is why we like simulations, you can measure whatever you want.
08:51:52 Generally, so in this case right, all we want to do is measure the flux that's in the wind or the outflow. I'm kind of using those interchangeably here.
08:52:02 In practice, of course there are a lot of choices left to be made.
08:52:07 In particular, right, you might have to decide whether you're going to measure in a plane parallel sense right just as a function of height for example from a disk or radial flux outward from the galaxy.
08:52:19 Once you've made that decision you have to decide where to measure at what height or radius. Are you going to measure this flux, you have to also decide if you're going to measure the instantaneous rate.
08:52:31 Right, or the time averaged values. Yeah, Exactly. Maybe, maybe a custom was the right shape for you.
08:52:41 So, the instantaneous or time average I'll touch on a little bit again later but you know this is really important one because I think often the time average value gives you more insight into the physics, but the instantaneous value is what observations
08:52:53 measure. Right, so again we're thinking about when you're going to compare. And of course you have to decide, you know what phase Are you going to measure and simulations and principal you can measure the total mass outflow right and all phases, but if
08:53:04 you want to compare it observations that may not be the right thing to measure.
08:53:09 I'll just run through a few examples. I'm not going to touch on cosmological simulations because Dylan basically did that yesterday, but you know so starting kind of from bigger and going and here's an example from from the fire collaboration.
08:53:22 This was from paper in 2015, just showing you how for example zoom simulation might measure the mass loading factor. So I would say on average cosmological in zoom simulations are usually measuring radio outflow rates and radio shells.
08:53:38 So in this case, say you have SPH particles. This is again pretty straightforward. You're just going to some the outflow rate in the spherical shell. So here this is giving you know m.is just the sum of this first term v dot you know our overall is just
08:53:55 the radial velocity times the mass of BSPH particles, some all those up in some, you know, in some radial shell divided by the width of the shell this gives you an M dot n m.in the show.
08:54:11 Um, it's worth noting right this is a plot then on the right of the outflow rate measured this way at a quarter of the radius that's the red dashed line, and at the burial radius that's the, you know, the less dashed line.
08:54:28 These values can be very high. And if you are going to measure an instantaneous an instantaneous mass learning rate from this simulation here the black line is the star formation rate actually scaled up by a factor of 10, you can get huge numbers and
08:54:44 particularly in the cosmological context, this is very likely to happen, because we think outflows right are actually one of the processes that's responsible for decreasing star formation.
08:54:55 So if you're looking at an outflow. That was rejected by a previous epic of star formation, then you're likely to see a really large value of beta.
08:55:06 Okay, so often actually in simulations we average and averaging. You know, em dots over some period of time here from this fire simulation they averaged from redshift to so redshift point five star formation rate over the same period of time.
08:55:25 This allows you to get a sort of time average value of the mass loading factor, and you know here, this is just showing that value of eight as a function of Halo mass and these fires fire simulations.
08:55:37 You guys are killing me with the chat is very hard not to laugh.
08:55:42 Okay, so again I'm going to point out I don't think this is what observations are typically measuring right observations are typically trying to get a measure of the current star formation rate, and then they have some measurement of the current outflow.
08:55:57 Great.
08:55:59 So, worth talking about how to connect these two things, I think in his keynote Dylan just put x. And this is what we're comparing and we moved on.
08:56:09 Okay so moving in a little bit and scale an example from, from my work from the seagulls simulations.
08:56:14 Same idea, we can measure the outflow rate in radio buttons here we're integrating over cells, instead of over, SPH particles.
08:56:24 But, you know, it's the it's the exact same idea I take some then I measure all of the mass.
08:56:31 That's out flowing in that been to answer marks question, it doesn't have to be only VR greater than zero, that's a choice you can make.
08:56:42 But you know, if you, if you don't make that choice then you could find that it's on radio and maybe it isn't it inflow.
08:56:49 And, you know, you get the same you get the same kind of results Here I'm showing a plot. Just, just plotting these mass outflow rates as a function of radius.
08:56:58 You can see I've broken this up into different phases. And again, once you've measured the mouse outflow rate converting to a mass loading is basically just, you know, what's the star formation rate.
08:57:11 Another thing I want to note though is that in a lot of these small scale simulations mine included on the star formation rate is not a self consistently calculated thing, right, we just put in clusters that correspond to some star formation rate that
08:57:26 we want to measure.
08:57:29 Whereas, you know, in reality there's there's some physical processes in the ISS, that are setting that star formation rate. And so that's that's an important thing to be aware of most of the simulations are not doing this so consistently.
08:57:46 Of course, that brings me to highlight one of the simulations that is doing the self consistently.
08:57:51 Really nice example from that from the Tigris simulation suite, which we've seen also in this workshop so theory you're looking at a plane parallel simulation.
08:58:03 So, a box that represents a patch of disk, so you can get nice and high resolution.
08:58:09 And there you can do the same kind of thing but you're going to be measuring horizontal averages, so they take the average rho V over some, you know, horizontal slice in this box, so at a given height, z, that gives a mass flux through the surface that's
08:58:25 what's being plotted here. And you can see that, depending on what height you pick here's flexes for one kilo bar set to kill carsick three kilo per second this box, you again are going to get different values for the mass outflow Right, right.
08:58:40 This is similar to the way in the last slide.
08:58:43 I was showing you get different values for him.at different radio.
08:58:49 If you want to convert that into a mass loading factor, fine, multiply the mass flux by the area of the surface, that gives you an M dot that can be used to calculate ADA.
08:59:01 Here it's plotted as beta given whatever the star formation rate is in the box. Right. And so here we see right there are some, there are some trends in this simulation.
08:59:12 In particular, you know the hot phase. Seems like maybe it has a sort of constant value of mass limiting factor with height. It's around point one, whereas the you know, the cold and the warm phases here are clearly dropping off.
08:59:27 If height.
08:59:30 So, right, a point to note is that measurements at different heights or radio will generally give you different values for Maslen factor. Okay, that's important.
08:59:41 And in particular, this is often because some of the gas that's out flowing at low Heights is just not going to make it to larger distances right we talked about this, this is what Filippo talked about his cloak as well if you think of the sort of ballistic
08:59:53 ballistic trajectory, or a fountain flow, then if you're measuring it at low heights, you will be capturing gas that's out flowing, that by the time you get to a higher height that fat gas is no longer a part of your outflow.
09:00:05 So this is not concerning, it's just something to be aware of.
09:00:09 But something else to note is if we go to larger scales right guess that's outside of the disk can also get swept up in outflows, which will change the mass loading factor.
09:00:17 And I think this is actually a major difference between the values of mass winning factor that are typically measured in the small scale simulations versus in the cosmological simulations, right, if you're looking at, you know, a quarter of the burial
09:00:42 not all of that outflow and gas was in the galaxy. So we don't need to push as much, much gas out of galaxies, as you know as as seen in the in the mass loading factors measured and these cosmological simulations. Right. So to put up, put a picture on it right this is where Tigris is
09:00:45 it right this is where tigress is measuring massive massively factors. This is where seagulls is measuring mass living factors. These are both very small scales compared to, you know, where one will typically measure a mass loading factor in a, in a cosmological
09:00:59 or zoom simulation. Okay, um this slide is because just asked this question, you know, do the simulations agree on, on these sorts of things.
09:01:08 And there's lots to say about this maybe we can talk about in the discussion I'm already running long on the small scale, I would say actually mostly we do agree.
09:01:17 In particular, some statements, I think we mostly agree on our that values of mastering factor for the holidays of about point one are pretty common so coming out of the disk, you know, a little ways away from the desk.
09:01:27 We're all getting values of order point one.
09:01:32 We get higher values when we include cool gas that has low velocities. But that's gas that that generally will not escape, and we still aren't, you know, generally getting values of masculinity factors that are above say 10.
09:01:46 There's a really nice plot this is from chain goose 2020 paper for the Tigers suite of simulations just showing values of the mass loading factor as a function of surface density of star formation, with different velocity, cut offs.
09:02:01 So you can see right if you include all out flowing gas, you get this really steep correlation between star formation surface density and mass alpha rate, really high rates at low circulation surface density, but as you increase that cut off right to
09:02:17 gas that will actually escape, so you could pick whatever your circular velocity is for your potential this this correlation kind of goes away.
09:02:26 So I would say, we're not quite sure how, how to best interpret this, but it's it's something that it's probably true.
09:02:34 On the halo scale, I don't know, I didn't have time to do the literature review so maybe they could tell me in discussion.
09:02:41 Okay, do I can I go on for another like five minutes.
09:02:46 Okay, I wanna I want to just hit on the observations briefly, um, fundamentally, it's the same idea we want to know what the outflow rate is relative to the star formation rate there's additional challenges though right star formation rate is is usually
09:02:58 what I'll call an instantaneous measurement right yes you can do integrated star formation but in these big samples I think we're generally just looking at, you know, something that gives us a measure of the current star formation, you often don't know
09:03:10 the geometry of the measuring measured out playing gas This is particularly true it, you know, higher redshift, and typically in observations you're measuring a single face.
09:03:19 So it's often very difficult to get at the hot gas we can only do that for really close galaxies, and even in a single phase say the more mine is gas you usually need to make a number of assumptions just to get at the mass that you're measuring nevermind
09:03:34 the mass outflow rate.
09:03:36 Right. And so the example that I'll show is absorption lines because this is where we actually have the best statistical samples, right if you know these are your absorption lines that you're looking at from a down the barrel observation of the galaxy.
09:03:50 The easiest thing to measure is the outflow velocity.
09:03:53 Nevertheless, there are choices to be made, right, you could use the maximum velocity of this line, but that's a signal to noise dependent quantity so often in trying to measure mass outflow rates you know observers will use something like the line centroid
09:04:06 instead, but worth noting that right you see there's there's very fast moving gas in this, in this phase.
09:04:14 Then this is you know this is straight from Tim Hawkins review right the simplest way to get an M dot and mass out for rate given this absorption line is to relate it to you know this these these quantities the solid angle the column density and hydrogen.
09:04:30 I mean mass, or hydrogen the velocity the outflow velocity that's the one you just pick based on that line, and are, which is the column density waited radius and be absorbing material.
09:04:42 So, why is this tricky well because we really don't know most of these things. Most of the time, right. These lines are often optically thick so we're not sure what is the ionization correction might be unknown so we don't know how to go from Silicon
09:05:15 silicon. We're not sure what the middle is it is so we're not sure how to map silicon column density hydrogen column density, the radius of the outflow depends on the flow structure where is the outflow and gas, you have to pick something, and the solid
09:05:09 angle of the outflow is often unknown there's some opening angle there might be some covering fraction. Okay, so there's a lot of uncertainties that go into measuring these mass out for rates.
09:05:21 Great. But let's say we've measured in dot, now we're good now we're off to the races we can measure Ada I've been told that instantaneous information re indicators are 100% accurate so this is no problem.
09:05:34 I'm obviously I'm being a little bit facetious, there's going to be a lot of scatter, and the observations. Nevertheless, if you make a plot like this this is a plot from Tim happens.
09:05:45 Paper looking at where am I and is war I mean kind of the 4k.
09:05:49 And just plotting star formation rate versus outflow rate. You can see this medium line is a is a massive leading factor of that too.
09:05:57 So that's kind of good, it's the right order of magnitude based on what we are seeing in the in the small scale simulations.
09:06:05 I also want to throw in a disclaimer here that that observers are working really hard on this and there's some really beautiful examples particularly of individual objects where a lot of these assumptions can be tested directly, and I want to call out
09:06:17 some work by white by john Chisholm he's been doing a really nice job of this. Nevertheless, I'm getting, you know robustly measured mass alpha rates for statistical samples which is what we really need is still is still a big challenge.
09:06:34 Okay, so finishing up to simulations observations agree, as I said on small scales I think things look kind of okay, we don't have great ways to measure right now the mass alpha reason the hot phase which we think is generally driving the winds.
09:06:48 But, you know, for the, for the warm ionized phase maybe things are kind of okay on large scales I don't know again I didn't have time to do literature review this is a plot from Dylan's 2019 paper where he's plotting mass leading factor in illustrious
09:07:03 versus host of observations and I don't know, do these agree, what does this mean.
09:07:11 No idea.
09:07:11 Again, topic for discussion, perhaps.
09:07:15 I was going to talk a little about other loading factors but I'll skip it in the interest of time, but I just want to know, we can play the exact same game with lots of other variables.
09:07:23 So, energy loading momentum loading metal content. These are all things that we can we can measure, and with some assumptions about how much there should be given the driving physics of the outflow.
09:07:38 You can, you know, you can try to try to make some, some physical interpretations of what's going on and the reason I want to highlight it is because I actually think this might be the way forward.
09:07:49 Mass is just one very specific thing but we can probably start to distinguish more between different physical models for example for our those if we start looking at things like energy and momentum loading as well.
09:08:02 You can imagine the energy loading factor between an AGM driven outflow starvation driven outflow might look very different, for example.
09:08:10 Okay, so summarizing mass loading is confusing, thumbs down mass shootings useful. Thumbs up.
09:08:18 I don't know, mass loading straightforward measuring simulations we're not always comparing the same thing. It's tricky to measure and observations, but it's easier to compare.
09:08:26 And maybe we can get some interesting physical constraints, if we include these additional mass limiting factors, this is really just supposed to be a point to bring up for discussion.
09:08:38 So I'll stop I haven't been able to keep up with the chat so we can talk about these about Forrester, so.
09:08:47 Excellent, thank you Thank you very much and that was really enlightening.
09:08:51 Very well done.
09:08:52 And, yeah, I think we can chat briefly.
09:08:59 Well no, I guess we should probably just be consistent and and and send people who want to discuss this to, to the breakout room.
09:09:11 And also a suggestion for, for, for relevant comments that bring up questions and whatnot throughout the the tutorials, it may be useful to write those questions in the slack, as opposed to just the zoom, since the zoom doesn't get necessarily recorded
09:09:28 indefinitely. I'll try and copy over relevant questions, not ones about thruster to the, to the slack. But, yeah, so we will restart in 10 minutes with Maria allows Marie Lau job Burgett and wrong on board lawyers presentation on white profile fitting
09:09:49 observational data. But until then, if you want to, to discuss this with Evan and have questions about mass loading and its various different incarnations, I encourage people to join the relevant breakout room right now.
09:10:05 Thanks Evan. That was awesome.