University of Patras
Υπολογιστική διαπραγμάτευση των ταλαντώσεων και των κυμάτων βαρύτητας σε αστέρες νετρονίων
Abstract
dc:descriptionThis PhD Thesis consists of two parts. In the first part, we study thequasiradialpulsations of rigidly and slowly rotating neutron stars and wecalculate the eigenfrequences of these oscillations.As quasiradialpulsations (or, equivalently, oscillations), we characterizethe pulsations of neutron stars that rotate uniformly and slowly. Thestudy of the states of equilibrium of a rotating neutron star is importantbecause rotation is a possible mechanism to avoid gravitational collapse atthe onset of nuclear fusion in their interior, but it could also play an importantrole in the damping of radial oscillations, resulting in loss of energyin the form of gravitational radiation. The computation of such rotationalpulsations is based on Hartle’s perturbation theory, according to which,the study of relativistic rotating massive objects is based on a perturbativesolution of Einstein’s field equations, describing a static, sphericallysymmetric model.Our task is to calculate, lying on the study of Hartle and Hartle andFriedmann, the zeroth-orderand second-ordereigenfrequencies of thelowest three modes of radial pulsation for general-relativisticpolytropicmodels. The zeroth-ordereigenfrequences, [σ^2]^(0), are the eigenfrequences ofthe nonrotating model while the second-orderones [σ^2]^(2) are the rotation-inducedchanges in the eigenfrequencies. When a star is rotating rigidlywith a small angular velocity, the squared eigenfrequencies σ^2 of itspulsation modes can be expanded in powers of Ω,σ^2 = [σ^2](0) + [σ^2](2) + ...where the superscript ‘‘ (0) ’’ denotes terms of zeroth order in Ω and thesuperscript ‘‘(2)’’ terms of second order in Ω. Concerning the zeroth-ordereigenfrequencies, we begin with the so-called"Chandrasekhar operator" applied to the so-called"displacement function" U and set equal to zero.To compute the eigenvalues [σ^2](0) we start the numerical integration for atrial value [σ^2] and the proper initial conditions and we integrate towardsthe surface and then check if the resulting solution U satisfies the boundarycondition ΓPU(R)' = 0 and thus establishing a SturmLiouville(SL) boundary value problem with eigenvalues the pulsation eigenfrequencies[σ^](0), which can be transformed into a system of two first-orderdifferentialequations. The second-ordereigenfrequencies [σ^2]^(2) are then computed numerically.We resolve four nonrotating general-relativisticpolytropic models with polytropicindices n = 1.0, 1.5, 2.0 and 2.5. Each model is resolved for fivecases of central mass-energydensities. These values have been chosento be below and relatively close to the ‘‘maximum-massdensities’’ of thecorresponding models, being in fact the more interesting ones when consideringneutron stars. Each density case is resolved for the three lowestmodes of pulsation: Mode 0, Mode 1 and Mode 2. For each mode,we compute a rigidly rotating configuration with angular velocity equal tothe corresponding Keplerian angular velocity, ΩK. Hartle’s perturbationmethod uses proper expansions in the rotation parameter ε = Ω/Ωmax, where Ωmax = sqrt(GM/R^3) is the angular velocity for which mass sheddingstarts occuring at the equator of a star. Thus Ωmax describes theNewtonian balance between centrifugal and gravitational forces. However,this Newtonian upper bound appears to be a rather overestimated limit forneutron stars. For such relativistic objects, the appropriate upper boundis ΩK. Hence, if the angular velocity is slightly greater than ΩK, then massshedding occurs at the equator of a neutron star. ΩK can be computedby several methods. For our numerical calculations we use the packageMathemαtica® and the Keplerian angular velocities have been calculatedusing the package "Rotating Neutron Stars Package" (RNS).In the second part of the thesis we study the nonradial oscillations ofneutron stars. Specifically, we develop a new numerical method, the socalled "overall complex plane strategy" (OCPS) for the computation of the eigenfrequences of the w, f and p-modesof nonradial oscillations of general-relativistic,polytropic neutron stars which have no Newtonian analogue.Numerical integrations are resolved by the Fortran code DCRKF54which is a Runge-Kutta-Fehlbergcode of fourth and fifth order, modifiedso that to integrate "initial value problems" established on systems of firstorder ODEs of complex-valuedfunctions in one complex variable along prescribedcomplex results. We numerically integrate on the complex planethe four first order ordinary differential equations that describe the non radialoscillations of a general relativistic polytropic model. In order to avoidsingularities at the center of the neutron star, where r = 0, we introduce asmall imaginary part in the radial coordinate r and we perform the integrationsalong a path parallel to the axis of the real numbers on the complexplane. Using the numerical method OCPS, we start the integrations exactlyat the center of the neutron star, thus avoiding numerical problems nearthe center of the star and also achieving higher accuracy in our results.On the contrary, in all previous studies, efforts have been made to startthe integrations as close to the center of the star as possible, either byusing Taylor approximation and power series expansion orby choosing starting points very close to the center of the star in order toavoid the singularities, without though avoiding numerical problems.This is a major advantage of our numerical method. Moreover, using OCPS,we succeed in calculating with accuracy even the eigenfrequences withvery small imaginary parts, a point where previous studies also encounterednumerical difficulties. Last, the stability of our numerical methodis verified from the fact that our numerical results are not influenced fromthe point where the integrations terminate in the exterior of the neutronstar, neither from the point where we perform the numerical conjunction ofthe vectors of the solutions resulting form inward and outward integrationin the interior of the neutron star in order to find a unique solution. Thenumerical values of eigenfrequences of the w, f and p-modesof oscillationare compared to those of the bibliography in order to verify the reliabiltyand the stability of our numerical method.Einstein was the first to derive the equations describing small, nonradial,quasi-periodicoscillations of general relativistic stellar models. Thissystem can be transformed in a fourth order system of equationsand be written in the vectorial form.We resolve it according to the study of Kokkotas and Schultz; we initiallyrealize five integrations of the vector in the interior of the neutron star,two from the center of the star outwards and three from the surface of thestar inwards. We integrate on a straight line parallel to the axis of the realnumbers and very close to it (at a distance equal to the imaginary part weintroduce in the radial coordinate r) and the integrations terminate at theso called matching point. This is the point where the linear combinationof the five integrations is realized in order to find a unique solution in theinterior of the neutron star. We continue performing a sixth integration inthe interior of the neutron star, from the matching point to the surface of thestar. The solutions of this final integration in the interior of the neutronstar are used as initial conditions for the integration of the Zerilli equationin the exterior of the neutron star outwards. We numerically integrate, afterwe modify the Zerilli equation, up to a distance very far from the centerof the neutron star on a straight line with slope ωim/ωre to the axis of the realnumbers. This is the preferred "path" to integrate, according to the "Stokesphenomenon". Finally, we perform a last integration of the modified Zerilliequation up to a terminal point which is taken as a reference pointwhere we could say is the position of an inertial observer. At this point,we transform the Zerilli function to its initial form and we calculate therate γ(ω)/β(ω) which expresses the rate of the intensity of the incoming gravitationalradiation to the intensity of the outgoing gravitational radiation. Aphysically accepted solution gives no incoming gravitational radiation, sothe frequences, ω, that make this rate zero are the eigenfrequences of thefree oscillation of the neutron star, they are complex and they could haveeither large or small imaginary parts. Their real part is connected to thepulsating rate of the neutron star whereas the imaginary part is connectedto the damping of the oscillation due to gravitational radiation.
Degree
thesis:*- Grantor dc:publisher
- University of Patras
- Year dc:date
- 2016
Author and committee
dc:creator, dc:contributor.*- Authors dc:creator
-
- Tzelati, Eleftheria
- Τζελάτη, Ελευθερία
Subjects
dc:subject × 16- Αστέρας νετρονίων
- Κύματα βαρύτητας
- Μέθοδος διαταραχής Hartle
- Συνολική στρατηγική του μιγαδικού επιπέδου
- Ημιακτινικές ταλαντώσεις
- Μη ακτινικές ταλαντώσεις
- Neutron star
- Gravitational waves
- Hartle's perturbation method
- Overall complex plane strategy
- Quasi-radial oscillations
- Nonradial oscillations
- Φυσικές Επιστήμες
- Φυσική
- Natural Sciences
- Physical Sciences
Rights
- Language dc:language
- gre
Identifiers
dc:identifier.*- Identifier
- 10.12681/eadd/39174