8.044 Lecture Notes Chapter 7: Thermal Radiation

1y ago
8 Views
1 Downloads
802.96 KB
23 Pages
Last View : 29d ago
Last Download : 3m ago
Upload by : Troy Oden
Transcription

8.044 Lecture NotesChapter 7: Thermal RadiationLecturer: McGreevy7.1Thermodynamics of blackbody (thermal) radiation . . . . . . . . . . . . . .7.2Statistical treatment of thermal radiation . . . . . . . . . . . . . . . . . . . . 7-107.3Phonons in a solid . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-19Reading: Baierlein, Chapters 4.1 and 6.7-17-4

Thermal radiation (a.k.a “blackbody” radiation) is the answer to the following simplequestion:What is the state of the electromagnetic (EM) field in equilibrium with its surroundings attemperature T ?It is what happens when you heat up empty space.body at temperature Tcut a hole in the box,observe radiation coming outthermal radiationin cavity3 (equivalent) viewpoints:1. As a particular thermodynamic system (This is §7.1).2. Via normal mode solutions to Maxwell’s equations in the cavity are standing EM waves.In §7.2, we will do stat mech from this starting point.3. As a gas of photons (subsequently, and in Chapter 9 this will be our viewpoint).We’re going to try to figure out:u(ν, T )dν energy of thermal radiation with ν frequency ν dνunit volumein thermal equilibrium at temperature T . Here ν (‘nu’) indicates the frequency of the EMradiation. u is called the “spectral energy density”.[u] energyvol · frequency7-2

To begin, here are some observed facts about thermal radiation (which we’ll come to understand): u(ν, T ) is independent of the cavity shape and the wall material. The walls of thecavity play the role of the heat bath, keeping the EM field at temperature T . The radiation is isotropic and unpolarized. The shape of u(ν, T ) looks like this:Shown are T 1 and T 1.5.1At T TRoom , νpeak 1013 Hz (λpeak 50µm) is in the IR (not visible).[End of Lecture 18.]1in some units with 1 c; the energy density and the frequency are both measured in units of thetemperature.7-3

7.1Thermodynamics of blackbody (thermal) radiationFirst: How far can we go with just macroscopic (i.e. thermodynamic) arguments?Relate energy density u to energy flux.cavity,filled withthermalradiationdetector,with area A E energy that hits the detector in time tenergy flux E power incident on detector intensityA tCalculate E: consider just the radiation with frequency ν and given momentum direction.( p hν/c always.)In time t all the radiation in this cylinder that is moving in this direction will hit the detector:Integrate over directions (θ and ϕ):π/2ZZdθ E 2πdϕ (energy in cylinder)00p(θ, ϕ) {z }prob that radiation is going in the right dir {z }if θ π/2, we’re outsideZπ/22πZdϕ u ·dθ 0 0· c t A cos{z θ}area of slice to p sin θ 4π{z }isotropic radiation cuA t4πZ0 π/2Z2πdθ cos θ sin θdϕ .0{z} {z } 1/2 2π1 energy flux per unit freq cu(ν, T )4On the LHS is the energy per unit volume per unit frequency, so the units work out.7-4

Next let’s calculate theRadiation pressurepressure p/ tforce areaareapz p · cos θRecall: for a chunk of EM radiation, p energy/cFor definiteness, assume reflecting walls (doesn’t affect answers).The change in momentum of the radiation that reflects off the wall in time t:Z pz π/2Z2π(energy in cylinder) · {zdθ00momentum normal to wall cuA t 2 4π cZ 0π/2Z2dθ cos θ sin θ{z} 1/3cos θc }cos θ creflects as beforeenergy2πdϕ . 0 {z } 2π pressure exerted by radiation with freq ν 7-5· {z}2 · p(θ, ϕ) . {z } pz1 u(ν, T ) .A t3

ThermodynamicsThat was for each frequency. Now let’s integrate over ν.Z1 dν u(ν, T )P {z}3 0the familiar thermodynamic pressureZU/V {z} dν u(ν, T ) u(T ) 0U is the familiar thermodynamic energyi.e. U V u(T ), and this is the only V -dependence in u – it’s extensive.2So, thermal radiation satisfies:P 1U3V(Compare to results on pset for ultra-relativistic gas.)Thermal radiation is described by P, V, T, U, S. – it is a hydrostatic system.Remarkably, we can go most of the way toward finding u(T ).2Actually, the fact that the energy density is independent of V is not clear from what we have said so far.A simple argument for it relies on the scale invariance of Maxwell’s equations – there is no quantity withthe dimensions of length appearing in Maxwell’s equations. Therefore, when you heat up a patch of emptyspace, the only quantity with units of length is the size of the patch of empty space. The energy per unitvolume must be independent of the volume.7-6

dU T dS P dV U S T P V T V T {z } {z } u(T ),since U V u ( Pby Maxwell reln) T V u(T ) T T P T PV1 du(T )3 dT 1 u(T )34T du(T )u(T ) 33 dTduu 4 u(T ) bT 4dTTfor some constant b. b cannot depend on T or V , or the properties of the cavity. It must besome constant of nature.bU bV T 4P T43 energy flux, out of a hole in the cavity cb 4T .4We’ve determined the pressure and the energy density (i.e. all the thermodynamics) ofthermal radiation, in terms of one constant b. Conventionally:1cb σ,4Stefan-Boltzmann constantNext we’ll show that :4kB {z}σ 42π 5 kBπ 2 stat mechW 8 10·5.67.3215h3 c260 {z}m2 K 4 {z}cQM EM, SR7-7

Why thermal radiation “blackbody radiation”?A body (i.e. a thing of finite size, i.e. a thing) can becharacterized by α(ν, T ), e(ν, T ):energy absorbedenergy incidentenergy radiatedEmissivity e(ν, T ) unit time · unit area · unit freqAbsorptivity α(ν, T ) both defined when the body has temperature T .Thought experiment: put the body in the cavity we discussed before. Wait long enoughthat the cavity, radiation and body are in equilibrium.In equilibrium, the body emits and absorbs the sameamount of radiation, at each frequency. Else, if the bodywere emitting radiation in some frequency more than itabsorbed, the radiation would not be in equilibrium –the state of the radiation field would be changing. Thisis called “detailed balance”.(?) :e(ν, T ) ·Eemitted (ν) Eabsorbed (ν)A {z}surface area of body · t α(ν, T ) · (energy flux hitting body) · A · t.{z} 1cu(ν,T )4e(ν, T )1 cu(ν, T )α(ν, T )4“Kirchoff’s Law”This is an amazing fact: the LHS has two material properties, each of which varies wildlywith material. The RHS is independent of the material, of the shape, of anything. 33There is a notational issue with Baierlein’s chapter 6: Baierlein uses a nonstandard definition of emissivity, in his (6.33). What Baierlein has done instead is to define:eBaierlein (ν, T ) eStandard (ν, T )1[u(ν, T )c/4]which means that for Baierlein, Kirchhoff’s Law (which he states in words right after (6.33)) is:eBaierlein (ν, T ) α(ν, T ).You need to be aware of Baierlein’s nonstandard definition if you compare any equation that has emissivityin it between Baierlein’s version and the version in my lecture notes.7-8

Conclusion: a good absorber (at freq ν) is a good emitter (at freq ν).A perfect absorber “black body” is a body with α(ν, T ) 1 for all ν, T . Such a body emits radiation with an emitted flux1e(ν, T ) cu(ν, T )4for a blackbody.But this is the flux out of a little hole in the cavity (Note: if the hole is small enough, waves incident fromoutside the little hole go in and never come out: so this isconsistent with the statement that the hole has αhole 1.)So: thermal radiation is blackbody radiation.This proves that the size, shape, reflectivity of the cavitydon’t affect the spectrum.energy radiatedfor any blackbody, in terms of one unknown conWe have determined the unitarea·unit timestant b. To find b, we need stat mech, next.7-9

7.2Statistical treatment of thermal radiationFor some of you, this will be your first exposure to quantum field theory.We want to calculate u(ν, T ) and u(T ) Rdν u(ν, T ).Consider a cubic cavity. We will build up the answer mode by mode (not point by point).EM facts [8.02 or 8.03]:Maxwell’s equations in a cavity have standing wave solutions, e.g.:Ey ( r, t) sin kx x E(t),kx mx πLMaxwell’s equations include: Bz ( r, t) 1 B 1 t E ckx cĖ(t) {z}cos(kx x)so far unknownĖ 2cos2 kx xE 2 sin2 kx x (kx c)2!11Ė(t)2V E 2 (t)228π(kc)x {z} R1 2 2 1energy density, u E B 8π8πZu Hone mode vol of cavityR!sin2 12 V cos2The EM dynamics of each mode E is an SHO! HSHO 21 mẋ2 21 κx2 with ω 2 κ/m; we haveω 2 (kx c)2 . Pick a mode number, mx , this gives kx mLx π which gives ωm kx c mxLπc .The solutions for the time evolution are E(t) a sin (ωm t ϕ) where a, ϕ are integrationconstants determined by initial conditions.But we know how to do stat mech for SHOs!7-10

Comments: Rather than springs at each point in space, these oscillators that we just found areeach modes which fill the box – they are distinguished by theirwavenumber k (kx , ky , kz ) π (mx , my , mz ),Lnot by their location. Crucial point: Maxwell’s equations are linear. This means that to find the generalsolution, we can just add up solutions for each mode:X kEEkand furtherH X(energy of each mode)kso in the canonical ensemble, the modes will be statistically independent – the partitionfunction will factorize. Claim: boundary conditions are not important for determining the thermal spectrum.They affect at most a few modes and their contribution becomes negligible in thethermodynamic limit of a large enough box (compared to the length scale set by T ).(You can (and will in recitation) redo the following analysis with e.g. periodic boundaryconditions.)7-11

Outline of calculation of thermal radiation spectrumThe calculation of the thermal radiation spectrum takes two steps:1. Count modes to get D(ω), the density of states:D(ω)dω # of modes with ω freq ω dω2. Calculate h (ω)i, the mean energy in the mode with frequency ω, in thermal equilibrium.Then the things we want are:ZU u(ω, T ) dωh (ω)iD(ω)h (ω)iD(ω) {z}1Venergy in modes with freq ωNote that I am using the angular frequency ω and the ordinary frequency ν interchangeably.They differ by a factor of 2π.7-12

Step 1: Counting modes Modes are labelled by k mLx π , mLy π , mLz π with mx 1, 2.; my 1, 2.; mz 1, 2. 4 Foreach choice of (mx , my , mz ), there are two modes, because there are two polarizations of theEM field. In the example above, we had kx , Ey , Bz nonzero; the other polarization state withwavenumber along kx has nonzero Ez , By .ω depends on k, i.e. is different for different modes, as we determined above: c2 π 2 ω 2 c2 kx2 ky2 kz2 2 m2x m2y m2zLThis is called the dispersion relation. Theallowed ks for a given box, make a 3d grid(a lattice). Here is what this grid wouldlook like for a 2-dimensional box:(The 3d version is on the next page.)Cumulative number of states:N (k) # of modes with k k 2 {z} 18 43 πk 3 π 3polarization L{z vol of one octant of sphere with radius k vol in k-space around each lattice point}# of lattice points 1 L3 k 33 π2Convert to N (ω) using the dispersion relation 3 3ω 2 c2 kx2 ky2 kz2 , i.e. ω ck: N (ω) 13 Lπ2ωc3 The density (of states) is obtained fromthe cumulative number (of states) in the usual way:D(ω) dN (ω)dω 4D(ω) L3 2ω .π 2 c3The example in the picture above had mx 5 and didn’t say what my and mz were. With hard-wallboundary conditions (the parallel component of the electric field vanishes at the walls), each of the threemodenumbers have to be nonzero in order to have a mode with nonzero amplitude (sin(0) 0). Actually,the normal component of the electric field need not vanish in general – so there can be e.g. a mode of Eywith (mx 6 0, my 0, mz 6 0) – but this depends on the polarizability of the walls. This kind of thing isthe subject of 8.03, I’m told. As you’ll see in recitation, this difference of one mode is not important forthermodynamics.7-13

Here I have drawn (told Mathematica to draw) : a blue dot for each mode of a 3d box, the region of k space with k 7 (shaded yellow) – its volume is the numerator in theDoS, the region of k space occupied by a single dot (red cube), in particular the one at(kx , ky , kz ) Lπ (2, 1, 2) – its volume is the denominator in the DoS.7-14

Step 2: Computing the energy in each mode (classical version)Before the advent of QM, we would have saidh (ω)i kB Tclassical equipartition for a classical oscillatorD(ω)h (ω)iL31ω 2 kB T 2 3 ω 2 kB T 3 Vπ cLπ 2 c3Z kB T ω 3dω u(ω, T ) 2 3 u(T ) .π c 3 00 u(ω, T ) Oops. In a classical world, an empty cavity at any nonzero temperature contains an infiniteenergy per unit volume. Each little patch would have an infinite energy. It’s not just aninfinite constant that you could subtract off either, because it depends on the temperature.This seems to be in conflict with observation. This was rightly called “the UltravioletCatastrophe”.This is the problem that Planck solved in 1895, by an inspired guess which began theunderstanding by humans of quantum mechanics.Next: the correct quantum treatment, which is already clear to you from the fact that theEM field in box is just a bunch of quantum harmonic oscillators.7-15

Step 2: Computing the energy in each mode (quantum version)12 , n 0, 1, 2. 1h (ω)i h n i ω hni 2 ω1 β ω ωe 12 {z}The allowed energy levels are n ω n “zero point energy”We drop the zero point energy. In the integral over ω, it gives an infinite constant. Wecan safely drop this by choosing the zero of energy, as long as we’re not worrying about itsgravitational effects.51u(ω, T ) h (ω)iD(ω) 3Lu(ω, T ) ω3π 2 c3 e k ωBT 1the Planck distribution[End of Lecture 19.]6On the right: For ω kB T , the oscillator at frequency ω is likely to be in its groundstate,is rarely excited; therefore it contributes kB T worth of energy.On the left: for ω kB T , these modes have classical occupation numbers, contribute kB Tper mode.5(Trust me that you don’t want to worry about its gravitational effects right now.) This process of takingadvantage of an ambiguity in the definition of the microscopic model (adding an (infinite) constant to theenergy of every state) is an example of a procedure called “renormalization” which is crucial in studyingsystems with lots of degrees of freedom distributed over space, like a quantum field theory.7-16

duuLocation of maximum: 0 dω ωkmax' 2.82. At Troom , λmax u ' 10 5 · 4.6 mBTwhich is 100 times lower freq (and 100 times longer wavelength) than visible light.Energy density: Z dω ω 3dω u(ω, T ) 2 3π c 0 eβ ω 10 4 Z dx x3kB T 2 3.π c ex 1 0 {z }U u(T ) VZ π 4 /154π 2 kB 3 c3} 15 {zU VT4 4σ/c, Stefan-Boltzmann constas promised. We can also find all the other thermodynamic quantities (Z, F, P, S). Since His a sum of the energies of each mode,XYdegeneracy(ω) ln Z(ω)Zi F kB T ln Z kB TZ ωall modesZ F kB Tdω D(ω) ln Z(ω) Z L3 21 kB Tdω 2 3 ω lnπ c1 e ω/kB T0 3 Z kB T L3 kB Tdx x2 ln 1 e x . 23π c 0{z}0 π 4 /45 F 41 π 2 kBT 4V .45 c3 3Negative F means T S E. 41 π 2 kB F P T433 V T45 c 4 F4 π 2 kBS T 3V 33 T V45 c 2 4414 π kB 41 π 2 kBU F TS TV T 4V45 45 3 c315 3 c3 as before.P 1U3Vas before, andσ 4π 2 kB.60 c2 37-17

Compare to experimentThe best known blackbody is the universe as a whole. At some time in the recent past (13billion years ago) there was a hot big bang, and all the stuff was in thermal equilibrium ata very high temperature; since then it’s been expanding and cooling. The universe is filledwith thermal radiation at TCMB 2.726 K, whose peak is in the microwave frequency range7 .A lot has been learned about the early universe by studying the fluctuations in the tem 10 5 .perature from different directions, which have a size of order δρρThe same spectrum also applies to gluons at tuniverse microseconds (the presence of quarksleads to corrections at the 10-20% level; we’ll learn to include them in Chapter 9).It’s a really good blackbody.7This value TCMB comes from the temperature above which electrons and protons decide to no longerform neutral atoms (this a chemistry energy, of order a few eV 104 K. This then gets divided by a factor of1000 to account for the redshift due to the expansion of the universe since the time when the temperaturewas so high.7-18

7.3Phonons in a solidSoundwaves in a crystal are a lot like lightwaves in a cavity. csound replaces c. So the dispersion relation is ω csound k . In counting modes for each wavenumber, for light there were two (transverse) polarizations. For sound, there are three – two transverse and one longitudinal. (In a gas,only the longitudinal mode is present.) There is one more distinction which will be important below, which is that (the quantum version of) Maxwell’s equations are valid down to as short a wavelength as weusually care about. On the other hand, soundwaves only are well-described as such iftheir wavelength is much longer than the spacing between the atoms of the solid.longitudinaltransverseWith these replacements, we can immediately write down the density of states:D(ω) 3V· 2 3· ω22 π csoundfor the range of frequencies ω where the soundmodes exist as such. A simple way to see thatthis formula can’t persist up to ω is that these are modes of vibration of the positionsof N atoms; these positions are specified by 3N coordinates. There are therefore exactly3N modes! [Recall from 8.03 that there are two useful perspectives (i.e. bases) here: thevibrations of each atom, or the Fourier basis of normal modes which diagonalize the energy.]7-19

Debye model of solidDebye suggested a reasonable approximate way totake this into account:Z ωD3N dω D(ω)0this defines ωD , the Debye frequency, which then servesas a high-frequency cutoff on all the integrals, a “UVcutoff”. This gives 1/3V33N 2 3 ωDcs /a ωD 6π 22π cs V 1/3where a Nis the average interatomic spacing.ωD defines a temperature ΘD , the Debye temperature, ωD kB ΘD ,i.e. ΘD 6π 2 1/3 cs /akBThis approximation does a good job if we ask the right questions: it makes an error at thehigh-frequency end of the ω integrals; but the large-ω physics is suppressed by the Boltzmannfactor, as long as T ΘD . A more correct answer depends on details of the crystal structure– e.g. is it a cubic lattice, is it a diamond lattice, is it a pyrochlore lattice.Following our previous analysis of light in thermal equilibrium, the energy density in atomicvibrations (‘phonons’) in the Debye model is:Z ωDZ ωD3 Udω ω 3dω D(ω)h (ω)i V2 π 2 c3s 0 e ω/kB T 10Numerically, this gives a heat capacity which looks like this:(the quantity ? which sets the units on the vertical axis is Vintuition for this answer, we consider .7-203 2 π 2 c3s kB 4.) To get some

LimitsThis problem is a little more complicated than the EM problem – there are two energyscales involved: T and ΘD .High T : T ΘD(This regime is less interesting since it’s where we know the model is wrong.) In this regime,we can Taylor expand the Boltzmann factor e β ω 1 β ω.:Z ωD3U3 3 kB T ωDdω ω 3 V2 π 2 c3s 0 ω/kB T2 π 2 c3s 3The s cancel! 3U N kB T,233N3 kB T 12 cs(6π)N kB T232 π cs 3V2V3CV N k B2classical equipartition of 3N oscillatorsLow T : T ΘDAt low enough temperatures, the Debye cutoff doesn’t matter, so the model does a goodjob. 4 Z ωD /T 4 43 kB TTdx x33 π 2 kBU 23x3V2 π cs e 12 15 c3s 0{z} π 4 /15 againCV V46π 2 kBT315 3 c3sThe answer is the same as blackbody radiation with c cs and a factor of 3/2 to accountfor the longitudinal polarization.V k3Translating this using the def of ΘD ( 3 cB3 s12π 4CV N kB5 6π 2 N)Θ3DTΘD 3for T ΘDThis is the dominant contribution to the heat capacity at low temperatures of solids whichare not conductors. (As we’ll see in Chapter 9, conduction electrons contribute CV T atlow temperatures. This is larger.)7-21

Compare to experiment (Figures from Baierlein)The slope determines ΘD .ωD ΘD cS.aMore rigid solids give bigger cs and hence bigger ΘD . Notice diamond.7-22

The Debye model does a lovely job of parametrizing our ignorance of the microscopicphysics in terms of just one parameter, which can then be measured. This is a nice exampleof an effective field theory: it describes the long-wavelength physics of a system in terms ofa continuum of modes; we know this description breaks down at some UV cutoff scale (herethe Debye frequency); dependence on the physics at such short wavelengths is parametrizedin terms of a few parameters (here, e.g. the sound speed). Other examples of such theoriesinclude the Standard Model of particle physics.7-23

energy of thermal radiation with frequency d unit volume in thermal equilibrium at temperature T. Here ('nu') indicates the frequency of the EM radiation. uis called the \spectral energy density". [u] energy vol frequency 7-2. To begin, here are some observed facts about thermal radiation (which we'll come to un-

Related Documents:

Introduction of Chemical Reaction Engineering Introduction about Chemical Engineering 0:31:15 0:31:09. Lecture 14 Lecture 15 Lecture 16 Lecture 17 Lecture 18 Lecture 19 Lecture 20 Lecture 21 Lecture 22 Lecture 23 Lecture 24 Lecture 25 Lecture 26 Lecture 27 Lecture 28 Lecture

Part One: Heir of Ash Chapter 1 Chapter 2 Chapter 3 Chapter 4 Chapter 5 Chapter 6 Chapter 7 Chapter 8 Chapter 9 Chapter 10 Chapter 11 Chapter 12 Chapter 13 Chapter 14 Chapter 15 Chapter 16 Chapter 17 Chapter 18 Chapter 19 Chapter 20 Chapter 21 Chapter 22 Chapter 23 Chapter 24 Chapter 25 Chapter 26 Chapter 27 Chapter 28 Chapter 29 Chapter 30 .

TO KILL A MOCKINGBIRD. Contents Dedication Epigraph Part One Chapter 1 Chapter 2 Chapter 3 Chapter 4 Chapter 5 Chapter 6 Chapter 7 Chapter 8 Chapter 9 Chapter 10 Chapter 11 Part Two Chapter 12 Chapter 13 Chapter 14 Chapter 15 Chapter 16 Chapter 17 Chapter 18. Chapter 19 Chapter 20 Chapter 21 Chapter 22 Chapter 23 Chapter 24 Chapter 25 Chapter 26

GEOMETRY NOTES Lecture 1 Notes GEO001-01 GEO001-02 . 2 Lecture 2 Notes GEO002-01 GEO002-02 GEO002-03 GEO002-04 . 3 Lecture 3 Notes GEO003-01 GEO003-02 GEO003-03 GEO003-04 . 4 Lecture 4 Notes GEO004-01 GEO004-02 GEO004-03 GEO004-04 . 5 Lecture 4 Notes, Continued GEO004-05 . 6

DEDICATION PART ONE Chapter 1 Chapter 2 Chapter 3 Chapter 4 Chapter 5 Chapter 6 Chapter 7 Chapter 8 Chapter 9 Chapter 10 Chapter 11 PART TWO Chapter 12 Chapter 13 Chapter 14 Chapter 15 Chapter 16 Chapter 17 Chapter 18 Chapter 19 Chapter 20 Chapter 21 Chapter 22 Chapter 23 .

Dr. S. Geetha lakshmi 044-22353595 - 044-2235369 8 vc@tnmgrmu. ac.in Registrar Dr. D. Balasubrama nian 044-22353574 - 044-2235369 8 vc@tnmgrmu. ac.in Dean Dr. Sukumaran Annamalai O:0431-3058687 999464555 4 0431-3058877 deancmchrc@ gmail.com Vice Principal Dr. Guru Datta Pawar O:

Mercedes-Benz Schweiz AG Bernstrasse 55 8952 Schlieren Telefon 044 755 80 00, Telefax 044 755 80 80 www.mercedes-benz.ch Leasing, Finanzierung und Fahrzeugversicherung Mercedes-Benz Financial Services Schweiz AG Bernstrasse 55 8952 Schlieren Telefon 044 755 99 99, Telefax 044 755 99 88 www.mercedes-benz-financialservices.ch

Lecture 1: A Beginner's Guide Lecture 2: Introduction to Programming Lecture 3: Introduction to C, structure of C programming Lecture 4: Elements of C Lecture 5: Variables, Statements, Expressions Lecture 6: Input-Output in C Lecture 7: Formatted Input-Output Lecture 8: Operators Lecture 9: Operators continued