SSC/Stability/n3PolytropeLAWE
Radial Oscillations of n = 3 Polytropic Spheres
Background
Our Formulation of the Problem
In an accompanying discussion, we derived the socalled,
whose solution gives eigenfunctions that describe various radial modes of oscillation in spherically symmetric, selfgravitating fluid configurations. Because this widely used form of the radial pulsation equation is not dimensionless but, rather, has units of inverse lengthsquared, we have found it useful to also recast it in the following dimensionless form:
where,
and
In a separate discussion, we showed that specifically for isolated, polytropic configurations, this linear adiabatic wave equation (LAWE) can be rewritten as,






where we have adopted the dimensionless frequency notation,



In this chapter we carry out a numerical integration of this governing LAWE for polytropes. The results are presented below.
Schwarzschild (1941)
We can directly compare our results with Schwarzschild's (1941) published work on "Overtone Pulsations for the Standard [Stellar] Model." To begin with, it is straightforward to demonstrate that the last form of the LAWE, provided above, matches equation (2) from Schwarzschild (1941), if is set to 3 — see the boxedin excerpt, immediately below. Note as well that Schwarzschild's dimensionless oscillation frequency — defined in his equation (1) and which we will label, — is related to our dimensionless frequency via the expression,



Schwarzschild (1941) numerically integrated the LAWE for polytropic spheres to find eigenvectors (i.e., the spatially discrete eigenfunction and corresponding eigenfrequency) for five separate oscillation modes (the fundamental mode, plus the 1^{st}, 2^{nd}, 3^{rd}, and 4^{th} overtones) for models having four different adopted adiabatic indexes .
Paragraph extracted from M. Schwarzschild (1941)
"Overtone Pulsations for the Standard Model"
ApJ, vol. 94, pp. 245  252 © American Astronomical Society 
^{3}A. S. Eddington (1930), The Internal Constitution of the Stars, pp. 188 and 192. 
Drawing from our discussion of the historical treatment of boundary conditions, we presume that Schwarzschild imposed the following constraint at the surface:












Recognizing from an accompanying tabulation that, for polytropes,



we presume that the surface boundary condition imposed by Schwarzschild was,



Our Table 1 catalogs the eigenfrequencies that Schwarzschild determined (drawn from his Table 1) for these twenty different models/modes.
Table 1: From Table 1 of M. Schwarzschild (1941) 


Mode  
0  0.00000  0.05882  0.10391  0.64414  0.13670 
1  0.16643  0.19139  0.21998  0.47003  0.25090 
2  0.3392  0.3648  0.3920  0.2120  0.4209 
3  0.5600  0.5863  0.6136  +0.1204  0.6420 
4  0.8283  0.8554  0.8832  +0.5248  0.9117 
Schwarzschild (1941) also documented the radial structure of the eigenfunction that is associated with each of these twenty model/mode eigenfrequencies. Each column of his Table 4, except the first, presents numerical values of the amplitude of a specific model/mode at 84 discrete radial locations throughout the n = 3 polytrope; the first column of the table lists the corresponding radial coordinate, . Focusing on the model that he analyzed assuming , we have typed his five columns of data into an Excel spreadsheet and have used this data to generate the pair of plots displayed, below, in Figure 1. The lefthand panel displays the eigenfunction amplitude versus radius, , for the fundamental mode as well as for the first four overtones; it essentially replicates Figure 1 from Schwarzschild (1941). The righthand panel displays the same data, but as a semilog plot; specifically, it displays , where,
Each sharp valley in this semilog plot highlights the location of a node in the corresponding eigenfunction, that is, it identifies where crosses through zero.
Figure 1:
Schwarzschild's Eigenfunctions for an n = 3 Polytrope with 

Numerical Integration
From the Core to the Surface
Here we use the finitedifference algorithm described separately to integrate the discretized LAWE from the center of the polytropic configuration, outward to its surface, which in this case — see, for example, p. 77 of Horedt (2004) — is located at the polytropiccoordinate location,
It is assumed, at the outset, that we have in hand an appropriately discretized description of the unperturbed, equilibrium properties of an polytrope; specifically, at each radial grid line, we have tabulated values of the radial coordinate, , the LaneEmden function, , and its first radial derivative, .
The algorithm is as follows (substitute everywhere):
 Establish an equally spaced radialcoordinate grid containing grid zones (and, accordingly, grid lines), in which case the gridspacing parameter, .
 Specify a value of the adiabatic exponent, , which, in turn, determines the value of the parameter,
 Choose a value for the (square of the) dimensionless oscillation frequency, , which we will accomplish by assigning a value to the parameter,
 Set the eigenfunction to unity at the center of the configuration, that is, set .
 Determine the value of the eigenfunction at the first grid line away from the center — having coordinate location, — via the derived powerseries expression,
 At all other grid lines, , determine the value of the eigenfunction, , via the expression,
We divided our model into radial zones and, using this algorithm, integrated the LAWE from the center of the configuration to the surface, for , and approximately 40 different chosen values of the frequency parameter across the range, . The radial displacement functions resulting from these integrations are presented in Figure 2 as an animation sequence. The specified value of is displayed at the top of each animation frame, and the resulting displacement function, , is traced by the small, red circular markers in each frame.
Figure 2: Numerically Determined Eigenfunctions for Various 



Each frame of the Figure 2 animation also displays, as smooth solid curves, the radial eigenfunctions that Schwarzschild (1941) obtained for the fundamental mode (blue curve) and the first three overtone modes (green, purple, & orange curves, repectively) for his model with . These are the same curves that appear in the lefthand panel of Figure 1, but here the displacement amplitude has been renormalized such that , and, along the horizontal axis, the radial location is marked in terms of the fractional radius, . In our examination of this model, as we approached each specific value of a modal eigenfrequency identified by Schwarzschild — see the frequencies highlighted in pink in our Table 1 — we finetuned our choice of the eigenfrequency in order to find a displacement function whose surface amplitude matched, to a high level of precision, the surface amplitude associated with Schwarzschild's corresponding published eigenfunction. The column of our Table 2 whose heading is "Match Schwarzschild" identifies — to at least 10 digits precision — the frequency choice that was required in order for these surface amplitudes to match in each case.
It is gratifying to see that our resulting frequencies match well the values published by Schwarzschild (as highlighted in pink, above). But this does not satisfactorily explain why, among the entire range of displacement functions displayed (in red) in the Figure 2 animation, Schwarzschild labeled these specific ones as the eigenmodes. As we shall now demonstrate, his eigenmode identifications resulted from the imposition of a specific, physically justified constraint on the slope, rather than the value, of the displacement function at the surface of the configuration. (See also our separate brief answer to the question, "What makes this an eigenvalue problem?".)
Surface Boundary Condition
As was stated, above, we presume that as Schwarzschild searched for natural modes of oscillation in isolated, polytropes, he imposed the following boundary condition at the surface of the configuration:



In order to duplicate his findings, then, we need to fine tune our specification of the oscillation frequency such that the resulting displacement function presents this behavior at the surface of our model. A finitedifference expression of this logarithmic derivative that is consistent with the abovedescribed finitedifference algorithm, is,



Everything is known here, except for the quantity, , which can be evaluated using the last expression in our algorithm one more time to, in effect, evaluate the eigenfunction just outside the surface. That is, we obtain and, in turn, obtain a value for the logarithmic derivative at the surface, via the expression,



We added to our numerical algorithm a step that evaluates, in this manner, the logarithmic derivative of the displacement function at the surface of our polytropic configuration. The eigenfrequencies that generated displacement functions with this surface behavior are listed for four separate modes in the column of Table 2 titled, "Match B.C." In every case the values agree to at least five decimal places with the "Match Schwarzschild" eigenfrequencies. We conclude, therefore, that it was the implementation of this surface boundary condition that permitted Schwarzschild to quantitatively identify the properties of the eigenvectors associated with natural radial modes of oscillation in polytropes.
Our Results
Table 3: Our Results (to be compared w/ Table 1, above) 


Mode  
0      0.311782342981  0.058812  0.644131577959  0.103912     
1  +0.24946512002  0.166310  0.113086698932  0.191276  0.47013975308  0.219907     
2  +0.50882623652  0.339217  +0.14705874055  0.364706  0.2121282667  0.391914     
3  +0.83977118  0.559847  +0.479241829  0.586161  +0.120257856  0.613505     
4  +1.24253191  0.828355  +0.8832297  0.855486  +0.52498863  0.883326     
Truncated n = 3 Polytropes
We understand that, for isolated polytropic spheres, the value of the adiabatic exponent for which the configuration is marginally unstable is , which is equivalent to, . This critical condition is identified by examining when the oscillation frequency of the fundamental mode goes to zero. Let's use our numerical integration tool to determine what this critical value of the adiabatic exponent is for truncated, polytropes. We will accomplish this as follows:
 At various truncation radii, …
 Force ;
 Iterate on the choice of until the displacement function with no radial nodes (i.e., the fundamental mode) satisfies the surface boundary condition of , to a desired level of accuracy.
The following table shows the values of — and associated values of — that we obtained for nine different values of ; in each case, iterations were continued until the desired surface boundary condition was satisfied to six significant digits.
PressureTruncated n = 3 Polytropes  
Edge Zone  Surface B.C  
200  1.00  0  0    
180  0.90  0.000201541  3.00000  1.333244  
160  0.8  0.00327575  3.00000  1.331879  
150  0.75  0.00808603  3.00000  1.329749  
120  0.60  0.0576031  3.00000  1.308214  
100  0.5  0.159111  3.00000  1.266179  
80  0.40  0.405712  3.00000  1.174497  
50  0.25  1.74909  3.00000  0.842266  
20  0.10  14.6648  3.00000  0.226439 
Analytic Inquiry
NOTE (from J. E. Tohline in April, 2017): The following subsections present some exploratory ideas that were pursued while I was searching for analytic solutions to the polytropic LAWE. For the most part this material has been superseded by a separate discussion in which we describe the desired analytic solution, which we discovered in March, 2017.
FundamentalMode, Homentropic Oscillations
The LAWE, presented above, that is relevant to polytropic spheres, may be rewritten as,
Now, if we assume that oscillations occur adiabatically with an adiabatic index that is consistent with the chosen polytropic index — that is to say,
in which case the configuration remains homentropic as it oscillates — and if we look only for a (marginally unstable) configuration that has , then the relevant LAWE is,



Specific case of n = 3 Polytropes
Homologous Collapse
If we examine only an polytropic configuration, then the last term disappears. This means that in this very special case, a perfectly valid solution to the LAWE is . This is presumably the eigenfunction that Schwarzschild deduced; the fundamentalmode "oscillations" are perfectly homologous. Given that the model is marginally unstable, an ensuing dynamical collapse will presumably begin in a perfectly homologous fashion. This is precisely the type of "freefall" collapse that was discussed and modeled by Goldreich & Weber (1980).
Another Potential Option
We have wondered whether, in this very special case, one or more additional fundamentalmode eigenfunction(s) might satisfy the governing LAWE. Here is a relevant line of arguments, beginning with the LAWE for the n = 3 polytropic sphere.









where,
But, by definition, the function is a logarithmic derivative of the LaneEmden function. Hence, we also can write,









Integrating this equation once gives,









Referring to the powerseries expansion of the polytropic LaneEmden function, , about the configuration's center, we see that the product, , goes to zero as the first power of . This means that the righthand side of this last differential equation blows up at the center. This, therefore, does not appear to provide a physically viable avenue by which to identify an alternative fundamentalmode eigenfunction.
Play With Form of LAWE
Logarithmic Derivative Rewrite
We have noticed that the LAWE that governs the eigenfunction associated with the fundamental mode of the marginally unstable model (FMMUM),



may be rewritten entirely as an expression that relates the logarithmic derivatives of and . Multiplying through by , then drawing on a differential relation that has been derived in a separate context, namely,



this LAWE associated with the FMMUM becomes,









I'm not sure if anyone else has previously appreciated that the "fundamental mode" polytropic LAWE can be written in this form. I'm even less sure that this form sheds light on its solution.
Play a little more … Start by letting, in which case we have,















One Feeble Guess
Now, what if, ?



in which case, the LAWE becomes,









Behavior for Known n=5 Solution
We know that the FMMUM for pressuretruncated, n = 5 polytropic configurations takes the form,



where,
The governing LAWE therefore gives,





















Let's check to see whether the known function properly satisfies this last ODE when n = 5.





Yes! 
Given that,






we could presume that, when defined in terms of , the defining function,





In this case, the governing LAWE becomes,












Now, from the polytropic LaneEmden equation, we also know that,
That is,






So, the LAWE becomes,



Again, let's check to see if the case of n=5 works …








Yes! 
Related Discussions
 Radial Oscillations of Uniformdensity sphere
 Radial Oscillations of Isolated Polytropes
 Setup
 n = 1: Attempt at Formulating an Analytic Solution
 n = 3: Numerical Solution to compare with M. Schwarzschild (1941)
 n = 5: Attempt at Formulating an Analytic Solution
Appendices:  VisTrailsEquations  VisTrailsVariables  References  Ramblings  VisTrailsImages  myphys.lsu  ADS  