PX 389: Cosmology Andrew J. Levan A.J.Levan@warwick.ac.uk May 4, 2016 Contents 1 Overview and key observations 1.1 Outline . . . . . . . . . . . . . 1.2 The night sky is dark . . . . . . 1.3 Redshift and Hubble Law . . . 1.4 Scale factor → redshift . . . . . 1.5 The cosmological principle . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 5 6 8 10 11 space-time . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12 12 12 13 14 15 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19 19 20 20 22 22 23 fluid and acceleration equations The Fluid Equation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The Acceleration Equation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . and finally ........ equations of state . . . . . . . . . . . . . . . . . . . . . . . . . . 24 24 25 26 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 Describing the evolution of the universe 2.1 Some basic relativity . . . . . . . . . . . 2.1.1 Faster than the speed of light . . 2.1.2 The equivalence principle . . . . 2.2 Curvature . . . . . . . . . . . . . . . . . 2.3 The Friedman equation . . . . . . . . . . . . . . basics of . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 Measuring Curvature 3.1 The relativistic Freidman equation . . . . . 3.2 Critical density . . . . . . . . . . . . . . . . 3.2.1 Mass density – stars . . . . . . . . . 3.2.2 Dark Matter . . . . . . . . . . . . . 3.2.3 The Cosmic Microwave Background 3.2.4 The flat universe . . . . . . . . . . . 4 The 4.1 4.2 4.3 5 Model Universes 5.1 Matter Dominated . . 5.2 Radiation Dominated 5.3 Cosmological Constant 5.4 Curvature Dominated . . . . . . . . . . . . . . Dominated . . . . . . . . . . . . . . . 1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28 28 29 30 30 6 Multi-component universes 6.1 Different densities at different times . . . . . . . . . . . . . . . 6.2 Solving the Friedman equation for a multi-component Universe 6.3 Observationally: Measuring expansion and acceleration . . . . . 6.3.1 The Hubble parameter - H . . . . . . . . . . . . . . . . 6.3.2 The deceleration parameter q . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33 33 34 36 36 36 7 Recent Observations of the Universe at large 7.1 Testing the Benchmark model . . . . . . . . . . . . . . . . . . 7.2 Future tests of benchmark cosmologies . . . . . . . . . . . . . 7.2.1 Large scale supernova searches . . . . . . . . . . . . . 7.2.2 High resolution Cosmic Microwave Background maps . 7.2.3 Baryon Acoustic Oscillations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39 39 40 40 41 41 8 Dark Matter 8.0.4 Dark matter temperature . . . . 8.1 MACHOS . . . . . . . . . . . . . . . . . 8.2 WIMPS . . . . . . . . . . . . . . . . . . 8.2.1 Supersymmetry . . . . . . . . . . 8.2.2 Axions . . . . . . . . . . . . . . . 8.2.3 Searches for particle dark matter 8.3 Modified Newtonian Dynamics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43 44 45 47 47 47 48 49 9 Structure in the universe 9.1 The Jeans Mass . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9.2 Formation in a radiation or matter dominated Universe . . . . . . . . . . . . . . 51 51 52 10 Inflation 10.1 The horizon problem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10.2 The flatness problem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10.3 Acceleration expansion at early times – Inflation . . . . . . . . . . . . . . . . . . 54 54 54 55 11 Baryogenesis and Nucleosynthesis 11.1 Baryogenesis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11.2 Nucleosynthesis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58 58 59 12 The 12.1 12.2 12.3 63 63 63 64 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Big Bang The Planck scale . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Vacuum fluctuations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Branes and the Epkyrotic Universe . . . . . . . . . . . . . . . . . . . . . . . . . . A The universe on a side of A4 65 2 B Problems and Answers B.1 Problem Set . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B.2 Problem Set with Answers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 67 67 69 Health Warning I have put this together with the aim of helping to clarify key points of the course, and to provide a repository for the aspects of the cosmology which are directly covered. The idea is that this will provide a basis for further reading, directing you to the areas which will aid your ultimate performance in the exam. In the past I have primarily used this as a aide memoir for myself, but am posting it online at the request of several people taking the course. It is inevitable that some typos and errors will have slipped through (this has not had the level of editing which any of your textbooks will have, not is it intended to be your lecture notes, which should be largely free of errors). If, at any time the contents of this appear to contract what you read in the textbooks (below) it is likely that these notes are in error. If you are worried about any possible errors, or on careful examination cannot understand any differences, then please see/email me. Also, please note that these notes are an overarching set of material that may be covered in the course. The exact flavour of the course may vary from year to year somewhat, and so the content of the lectures (including what is said and not written) should be viewed as the final arbiter of what is required (and examinable) for the course. Textbooks The lectures are intended to be contained, and from 2014 onwards are also available via Echo360. However, should you wish to consult textbooks then I can recommend the following. B. Carrol & Ostlie – Introduction to Modern Astrophysics This is not a formal textbook on cosmology but is the recommended text for the second year stars and galaxies course and also contains some useful information for this course M. Roos – Introduction to cosmology B. Ryden – Introduction to cosmology Andrew Levan - Vlissingen, the Netherlands (December 2014). 4 Chapter 1 Overview and key observations Cosmology is principally concerned with the origin and evolution of the universe as a whole. It is a relatively new subject, at least in the way which it is presented now, as a quantative science. Nonetheless it roots reach back well into the past, since its fundamental objective – understanding where the universe in which we live came from, and where it is going – have been central questions since the dawn of civilization. The earliest cosmologies came from the Ancient Greeks, who placed the earth at the centre of the universe, with the sun, moon and other planets revolving around it in a series of different “circles” (Ptolemy’s epicycles). Although contrived this model was understandable since the conceptual leap from viewing stars merely as points on the sky, to viewing them as other suns, perhaps complete with their own solar systems, is a huge one. Copernicus was the first to suggest that the sun, and not the earth lay at the centre of the universe, but this was not broadly accepted until some time after, Newton recast our vision of the universe by providing a mathematical underpinning to the Copernican view. The discovery of external galaxies again questioned our location in the universe, although, the natural assumption then became that our own galaxy much lie the centre. Only the the 1950’s, when Baade finally showed that the Milky Way was really rather like other galaxies did arguments assigning us a special place in the universe finally give way, allowing the picture of modern cosmology, which we will study during this course, to emerge. 1.1 Outline The questions and aims which will be addressed over the timescale of this course will be: • Key observations - the birth of cosmology • Describing the universe and space time • Model universes - describing their evolution • Measuring the key properties of our Universe • The formation of structure 5 • The early universe Although much of this course will be describing the models which explain the above questions there will also be a focus on the observations which provide evidence for these; why are they important? what are their flaws? and what future directions will help us to better understand cosmology? At the end of the course you should have a quantative understanding of the current status of cosmology, focussed on the breakthroughs in the 20th century which enable us to describe its evolution, but sufficiently up to date to take into account recent research highlights, including those which have happened as recently as 2007. Modern cosmology is built on several key observations, and it important to understand these, and their origin before moving on to more complex descriptions of the universe. 1.2 The night sky is dark This, remarkably simple observation actually provides an important cosmological lesson, and was first studied by Olbers in (1826). It essentially takes the Newtonian view, that the universe is an infinite space occupied by an infinite number of ”other suns”, and demonstrates that, despite its simplicity the fact that the night sky is dark has profound implications for the nature of the universe. Imagine that the universe consists of an infinite number of stars. The flux (F ) from each star, as received at the earth is simply L , (1.1) 4πr2 where L is the luminosity of the star. These stars are arranged in some configuration about the sun (which we integrate from since it is our viewpoint, not because it occupies and special place in the universe). Suppose we split up the sky into a series of spherical shells. The light from each steradian of the shell, nr2 dr is simply, F = L nL × nr2 dr = dr 2 4πr 4π and therefore the brightness of the night sky per unit area is dJ = Z ∞ (1.2) nL ∞ dr = ∞ (1.3) 4π 0 0 Thus, if the universe did consist of an infinite number of stars, then the sky would not be dark, but blindingly bright. This is known as Olbers paradox. In fact, as we shall see later the solution to this is simple. J= Z dJ = The universe is a finite age, Light from sources at distance greater than the speed of light time the current age cannot yet have reached us. 6 Figure 1.1: Olbers paradox. In a universe filled uniformly with stars or galaxies the sum of the light from each star, integrated over the universe is infinite. The solution to this problem is that the age of the universe is finite. In this systems objects over our horizon are invisible, since their light has not had time to reach us since the big bang. 7 1.3 Redshift and Hubble Law Spectrographs, which spread the light from a given source over a range of wavelengths allow precise measurements of the radial velocities of stars and galaxies from the us. This is possible because of discreet spectral features in both stars and galaxies, such as absorption lines (in stellar atmospheres) and emission lines (from excited gas). The relative location of these lines, with respect to that measured in the laboratory is known as the redshift. In simple observer terms (1 + z) = λobs λrest (1.4) Locally this enables us to, for example, measure the orbits of stars around the Milky Way1 . n a larger scale it also allows us to measure the velocities of extragalactic systems. As it happens most galaxies (aside from our nearest neighbours) are moving away from us. Edwin Hubble realised in 1929 that this isn’t simply a random motion, but that the redshift is correlated with its distance, that is that more distant galaxies are moving away from us more rapidly. He coined this into the now famous Hubble’s Law v = H0 d (1.5) Where H0 is a number known as the Hubble Constant, and is currently constrained, principally by observations of Cephied variables to be H0 = 72 ± 2 km s−1 Mpc−1 . Thus a galaxy moving away from us at 500 km s−1 would be a distance of ∼ 7 Mpc away. In practice the determination of H0 has been one of the major challenges in astronomy since it was first discovered by Hubble in 1929. Indeed, the Hubble Space Telescope is so named because one of its chief goals was to improve the value of the Hubble constant. The principle difficulty in this determination is not in measuring redshifts, which are straightforward, but in measuring distances. All distance measurements are built from the distance ladder, but the direct measurements which underpin this are only possible in our own galaxy (and even then only in stars close to the sun). The value of the Hubble constant is now determined principally from observations of Cepheid variable stars (whose period correlates with luminosity), allowing a means of measuring the distances of galaxies of to ∼ 30 Mpc, and therefore placing much better constraints on H0 than was possible in the era of Hubble. Hubbles law states that there is a linear relationship between velocity and distance, with a proportionality constant H0 = 72 km−1 Mpc−1 . This law holds out to z ∼ 0.2 One consequence of Hubble’s law is that it seems to refute the claim that we are not at a special place in the universe. Surely if every other galaxy is moving away from us then we do occupy some prime position. This may be true if the observed redshift was purely the result of relative motions in a fixed space, however, the truth is that the space itself is expanding. The common analogy is that of blowing up a balloon. Imagine sticking pictures on the side of a 1 Sometimes rather than plot z the velocity cz is plotted instead, note that this can be larger than c! 8 Figure 1.2: A triangle of galaxies, expanding uniformly. balloon and then inflating it. The pictures remain the same size, but each of them moves away from each other one, much as each other galaxy appears to move away from us. This can be shown with slightly more mathematical rigour. Suppose the triangle shown in Figure 2.2 expands uniformly. In this case the relative shape of the triangle is preserved in time. i.e. r12 (t) = a(t)r12 (t0 ) r23 (t) = a(t)r23 (t0 ) r31 (t) = a(t)r31 (t0 ) (1.6) Where a(t) is the so called scale factor which will be important later. Imagine an observer in galaxy one. The relative velocities observed for galaxies 2 and 3 are dr12 = ȧr12 (t0 ) = dt dr31 v31 (t) = = ȧr31 (t0 ) = dt v12 (t) = ȧ r12 (t) a ȧ r31 (t) a (1.7) Where we have made the obvious substitution that r12 (t0 ) = r12 (t)/a(t). The form of this equation is clearly that of Hubble’s law. In this case value of H0 can be equated with ȧ/a, at the time t = 0 (i.e. the current time). More generally the value H is known as the Hubble parameter, and is variable with time. The Hubble parameter is simply the rate of change of the scale factor of the universe 9 An important point here is that the expansion of the universe implies that at some point the two galaxies were in contact with one another. Using this we can calculate the time at which the two galaxies were in contact. Since the time the galaxies were in contact they have expanded to r21 = v21 t0 , therefore: t0 = r12 r12 = = H0−1 v12 H0 r12 (1.8) This is known as the Hubble time, and for H0 = 72 km s−1 Mpc− 1 then t0 , also known as tH = 14.0 ± 1.4 Gyr. This is the natural explanation of Olbers paradox. Note that this also gives a natural distance for the universe, the distance which light can travel in a Hubble time, simply c/H0 = 4300 ± 400 Mpc. 1.4 Scale factor → redshift The above shows how we can write the evolution of an expanding Universe based on the change of the scale factor. This in practice actually directly relates to the more measurable parameter, the redshift, or velocity of the object. Because of this, if we can measure the redshift of a galaxy we can directly infer the relative size of the Universe at the point that light was emitted. Consider two adjacent galaxies which are moving apart from one another in the Hubble flow, with a relative velocity dv. For this galaxy we can write Hubble Law as ȧ dr a = dv/c , we can substitute in above for dv dv = Hdr = now, from the doppler effect dλ/λemit ȧ da dλ ȧ dr = dt = = λemit a c a a (1.9) (1.10) This can be integrated Z dλ = λemit Z da a (1.11) so ln(λ) = ln(a) + c, or λ ∝ a We now can return to our definition of redshift λobs aobs = (1.12) λemit aemit since by definition aobs = 1 (i.e. the Universe is defined to have a scale factor of 1 at the present time), the simplifies to 1+z = 1 (1.13) a In other words. If we observe a galaxy at z = 1 we observe it when the Universe was exactly half of its present size. If we look at the Cosmic Microwave background at a redshift of z = 1100 we see the Universe when it was ∼ 1/1110th of its present size, and 109 times denser then it is today. 1+z = 10 1.5 The cosmological principle The discovery by Baade in the 1950’s that the Milky Way was a galaxy very much like any other led to the acceptance of the view that we do not occupy a priveldged space in the universe. This has in turn led to a stronger statement, known as the cosmological principle. This states that. The universe is homogeneous – no preferred location The universe is isotropic – no preferred direction When looking around us day to day this is patently ridiculous. The world we inhabit is clearly neither homogeneous or isotropic, there are clearly different (and distinct) locations, and directions which can easily be determined. Indeed, not only does this not hold true on the scale of humans it does not hold true on the scale of normally studied astronomical objects. The solar system is clearly not the same on average – the earth is not the same as Jupiter, or interplanetary space. Even the galaxy, or the set of galaxies which form the local group are not especially homogeneous, and have obvious preferred directions (e.g. towards the Milky Way or M31). However, while this concept may seem odd, when viewed on the largest scales (which are, of course the scales we are most interested in cosmology) the universe is both homogeneous and isotropic – the relevant scales are > 200 Mpc. 11 Chapter 2 Describing the evolution of the universe - basics of space-time In the last lecture we set up some of the basics in the universe. From this we now know. • The universe is both isotropic and homogeneous • The universe is expanding, and its current rate of expansion is 72 km s−1 Mpc−1 We now want to know how we can describe this expansion in more general terms. In particular, in the first lecture we looked at describing the universe in terms of observational parameters, such as distance and velocity. We now want to know how it evolves on a more theoretical level, based on force, mass and energy. The first thing to consider is that the only force of consequence in the evolution of the universe (at this stage of its life) is gravity. The strong and weak nuclear forces act only at short distances, and despite the fact that the electromagnetic force is 1036 stronger than gravity on a pair of individual protons, the universe, on large scales is neutral. In order to understand gravity, we need to understand, at least to first order something about general relativity. 2.1 Some basic relativity Before we can set about understanding the universe we need to understand some of the basics of space time, so that we can set up the models that we need to describe the evolution of the universe as a whole. This course will, in general, not worry about relativistic effects, but we do need to understand at least some of the basics of general relativity, and to set up the necessary maths to enable us to describe the universe. This is not the most exciting section of the course, but it is necessary to understand it before we can move on. 2.1.1 Faster than the speed of light Firstly, before moving on to some basic of general relativity, a note on special relativity. Galaxies can be moving apart from one another at faster than the speed of light. In the last lecture we 12 saw that the Hubble time t0 is defined as 1 H0 (2.1) dH = ct0 (2.2) t0 = and a corresponding Hubble distance Essentially (and this is simplified) it means that as v = c = H0 dH (2.3) The velocity of these galaxies is the speed of light, anything beyond them is moving at v > c. This does not violate special relativity since it is due to the space between the galaxies expanding. Two galaxies, moving apart from one another at faster than the speed of light are not causally connected. 2.1.2 The equivalence principle This statement is central to how Einstein created General Relativity. At its most simple, it say that the force due to gravity, GMg mg r2 where mg is the gravitational mass, is equal to the force due to uniform acceleration Fg = F = mi a (2.4) (2.5) where mi is the inertial mass. The equivalence principle states simply that mi = mg (2.6) That is to say that how strongly an object is pulled on my gravity is equal to its resistance to acceleration by any other force – it is not a natural assumption to make. This statement essentially means that it is impossible to tell between a uniformly accelerating frame, and one in a gravitational field. The classical thought experiment to prove this is that if one were in a spacecraft uniformly accelerating then it should be impossible for an astronaut within that craft to perform a local experiment (i.e. one inside the ship) to determine if he or she was uniformly accelerating, or in a gravitational field. The natural experiment would be to shine a light at the opposite wall. Since light travels in a straight line, in the accelerating case the light would strike the wall at a lower point than it was started from, due to the motion of the craft. However, in the Newtonian view at least, the light, in the gravitational frame would still hit the same point on the wall. Since this would provide a means of determining the difference between acceleration and gravity, it would disobey the equivalence principle. Rather than take the easy option, and abandon this, Einstein to the drastic, but apparently correct, view that in 13 Figure 2.1: The basic equivalence principle. There are no local experiments that observers in either a uniformly accelerating frame, or a frame sitting in a constant gravitational field can do which will determine which frame they are in. As a consequence light must be bent by the presence of the gravitational field (or more generally, by mass energy). fact the light would still be bent down by the gravity, because of the curvature of space-time itself. Under this model, all masses bend space-time, and light follows geodesic lines through this. This is well put in Barbara Ryden’s introductory cosmology book: Newton Mass tells gravity how to exert a force (F = −GM m/r2 ) Force tells mass how to accelerate (F = ma) Einstein Mass–energy tells space time how to curve Curved space time tells mass energy how to move 2.2 Curvature The effect that mass can have on the universe, and its curvature from this, can at crudely be understood in terms of some simple geometries. Flat, is the standard Euclidean geometry which we are familiar with. A simple way of understanding this is to imagine a triangle drawn on a surface. In a flat systems the angles add up to π e.g. in Figure 3.1. 14 Figure 2.2: Flat (left), negative (centre) and positive (right) curved spaces. α+β+γ =π (2.7) In a positively curved surface - a spherical one α + β + γ = π + A/R2 (2.8) Where A is the area of the triangle, and R the radius of the sphere. In a negatively curved space, which is hyperbolic α + β + γ = π − A/R2 (2.9) Geometrically many different possibilities exist, however because the universe is homogeneous and istotropic (yes we come back to this again). There are only three options, the universe has constant positive curvature, constant negative curvature, or is flat. 2.3 The Friedman equation We are interested in how a particle behaves under the influence of gravity in an expanding universe. This is essentially described by how its potential and kinetic energy evolve. Note, that because of the cosmological principle we can use any particle we like, since there is no difference between them. It can be shown (and probably has been in earlier dynamics lectures) that in a sphere mass outside the observer does not contribute to the gravitational force. This means that we are only concerned about matter inside the radius of the particle. Assuming uniform density (again an assumption we can make from the cosmological principle), then its mass is simply given by, M = 4/3πρr3 , so F =− 4πGρmr GM m =− 2 r 3 15 (2.10) Figure 2.3: A uniform density sphere contracting or expanding under its own gravity. Figure 2.4: As co-moving co-ordinate system, which expands with the universe. 16 and the gravitational potential energy is given by V = −GM m/r = − 4πGρmr2 3 (2.11) Kinetic energy is just just 1 T = mṙ2 2 As energy is conserved the total energy is just (2.12) 1 4πGρmr2 U = T + V = mṙ2 − (2.13) 2 3 Now we can make a clever switch of co-ordinates, which makes understanding this simple. These are called co-moving co-ordinates. This is just (thinking back to lecture 1) r(t) = a(t)x (2.14) Where r(t) is the co-moving co-ordinate and x is the physical (non-moving co-ordinate) so that ẋ = 0. Substituting this in above gives: 1 4π U = mȧ2 x2 − Gρa2 x2 m 2 3 (2.15) We can now multiply each side of this equation by 2/ma2 x2 to get 2 ȧ a = 8πG 2U ρ+ 3 mx2 a2 (2.16) Commonly a substitution kc2 = −2U/mx2 is made, which reduces the equation to the standard Newtonian Freedman equation: 2 ȧ a = 8πG kc2 ρ − 2 = H2 3 a (2.17) Note that the form (ȧ/a)2 is that of the Hubble parameter (this is not the Hubble constant H0 , which we examined in the previous lecture as the current value of ȧ/a, but is a more general Hubble parameter (H), describing this ratio at different times. In this case, k is called the curvature constant, and defines one of three possibilities for the curvature of the universe. Since the universe is homogeneous (yes, we come back to this yet again) the curvature has to be constant, and cannot vary across the universe. Therefore there are essentially only three possibilities for the curvature; uniform positive, uniform negative, and flat. We can examine the consequences of each in turn. Let us start by imagining a expanding universe, where ȧ > 0. In cases where k is negative the right had side of the equation is always positive, and so the universe will continue to expand forever. If k is positive then the equation begins positive, but with time a(t) reaches a maximum. 17 k k>0 k<0 k=0 Geometry Spherical Hyperbolic Flat Type Closed Open Flat Fate Collapse Expansion Expansion Table 2.1: The different types of universe depending on the value of the curvature constant. Figure 2.5: The evolution of the universe with time for different values of the curvature. amax = − GMS 3kc2 = Ux 8πGρ (2.18) At this point the expansion will stop and the second derivative of a will be negative, and so the universe will contract. Finally we can consider the case where k = 0. Since the universe is shaped by its mass energy this only happens at a certain mass/energy density. ȧ 8πGρ = H2 = a 3 This can be re-arranged to give a critical density, 3H 2 8πG If the universe is at this critical density then k = 0 and it will expand forever. ρcrit = 18 (2.19) (2.20) Chapter 3 Measuring Curvature In the last lecture we saw how mass/energy causes curvature in the universe, and set up (from a Newtonian viewpoint), the Friedman equation, which describes the evolution of the size of the universe. We also saw broadly how the curvature constant can impact the behaviour of the universe. In this lecture we will take this a step further, looking at how we measure curvature, and what observations today imply about the curvature of the universe. 3.1 The relativistic Freidman equation The equation we derived in the last lecture is known as the Newtonian Freidman equation, because it was derived from basic Newtonian physics. There also exists a fully relativistic version (which we shall not derive). The form of this is: 2 ȧ a = 8πG κc2 Λ − + = H2 2 2 2 3c 3 R0 a (3.1) The three changes here are due to the replacement of the mass density with the energy density, since E 2 = m2 c2 + p2 c2 (3.2) so that photons can in fact contribution to the curvature of the universe since they carry momentum and energy. The second is to replace k with κ, and subsequently re-scale using R0 (currently observed scale factor). This means that rather than worry about a range of positive or negative values for k, it is either 1,0,-1, when normalised with the current scale factor. The final addition is the term Λ/3, this term was added by Einstein when it became clear that the “original” Friedman equation required the universe to be either expanding or contracting. As the viewpoint at the time favoured a steady state Universe he introduce the Λ term, which he called the cosmological constant. Although this was thought to be his biggest blunder after the discovery of universal expansion in recent years it has become clear that the universe does have cosmological constant (which will be discussed in more detail later). 19 3.2 Critical density Since the Universe is curved by the impact of its mass energy, it will be flat if it achieves a critical density, which is derived by setting k = 0 in the non-relativistic Friedman equation. 3H 2 8πG Commonly this critical density is re-written as ρcrit = Ω= ρuniverse ρcrit (3.3) (3.4) and the contribution of individual components (e.g. wM ) is indicated separately. We can of course calculate a numerical value of ρcrit now by substituting in the Hubble constant ρcrit,0 = 3H02 = 1.4 × 1011 M M pc−3 = 9.2 × 10−27 kgm−3 8πG (3.5) Simply using E = mc2 this can also be written as an energy density. crit,0 = 3c2 H02 = 8.3 × 10−10 Jm−3 8πG (3.6) So, an important component of the curvature of the universe is its mass/energy density. A simple (at least in principle) way of measuring this is to “count-up” the obvious sources of mass energy within the universe. 3.2.1 Mass density – stars A first idea is simply to count up the mass. This isn’t actually as easy as it looks. We cannot trivially measure the mass of each star in the universe, and add it up. What is much simpler is to measure the integrated light, and then turn this into a mass. To do this, we need to define something call a mass to light ratio. We can easily define this based on the sun, which has a mass of one solar mass (M = 2 × 1030 kg) and a luminosity of one solar luminosity (L ), which in the B-band equates to 4.7 × 1025 Watts s−1 . We can now measure the luminosity of galaxies within several Mpc of our own and average to get the mean B-band luminosity j = 1.2 × 108 L,B M pc−3 (3.7) If we assume that all stars have the same mass to light ratio we can trivially convert this into a mass. Of course, in truth stars do not have a uniform mass to light ratio. Low mass stars put out much less energy than high mass stars, in general this can be approximated by a power law dL ∝ M3 dM 20 (3.8) However, low mass stars greatly outnumber the higher mass stars, their initial mass function is dN ∝ M −2.35 (3.9) dM In practice this means that much of the light comes from high mass stars, while most of the mass lies in the lower mass objects. Nonetheless the assumption of a mass to light ratio similar to the sun is not a bad one, and hence assuming a mass to light ratio of one gives a mass density of 1.2 × 108 M M pc−3 . This provides a fraction of the critical density of ∼ 0.0001. In practice a slightly more complete estimate is that the mass of stars locally is ∼ 5 × 108 M , and thus 5 × 108 ∼ 0.004ρcrit (3.10) 1.4 × 1011 It is clear that stars alone do not provide the necessary mass-energy to close the universe. Of course, not all the mass in the universe lies in stars, we know that there exists dust, while there are also stellar remnants, such as black holes and neutron stars which can increase this number, but only by a factor of a few, not sufficient to close the universe. The total estimated density in this form of matter (known as the baryon density) is ρstars = Ωb = 0.04 (3.11) an order of magnitude larger than our estimate of the stellar density, but not enough to flatten the universe. If this were the true picture of the universe then it would be open, and expand forever. However, we have not yet explored all the possible sources of mass-energy within the universe. The orbital velocity of an object around a mass is simply given by a= v2 GM (R) = R R2 (3.12) i.e. s v= GM (R) R (3.13) Therefore, the velocity in orbit around a point mass (or any object whose entire mass is enclosed within radius R) should fall of as R−1/2 . Therefore, this is what we might expect to observe for stars far out from the cores of galaxies, where there is little mass. A typical galaxy light profile can be described as an exponential disk, where the intensity of light varies as I(R) = I(0) exp −R , RS (3.14) where RS is the scale length of the disc, and for typical galaxies is a few kpc. Since the light falls off exponentially there is little light (and from our arguments above, little mass) outside this point. 21 3.2.2 Dark Matter As luminous matter apparently provides very little of the critical density of the universe we are naturally led to the question of whether any other sources of matter exist. Indeed, evidence for some form of dark matter can be found in the rotation curve of the Milky Way, and other galaxies. The rotation curve of a given galaxy should be described simply by its orbital rotation velocity s v= GM R (3.15) Where M is the mass of the galaxy and R the radius from its centre. If the galaxy were dominated by starlight then beyond the visible stars M would be constant and the velocity should fall as R−0.5 . However, in the majority of galaxies the rotation curves are flat out to large radii indicating that the mass is continuing to increase with increasing R. This component is known as dark matter. There are various possible explanations for the origin of dark matter, and we will return to these later in the course. From the perspective of measuring curvature the important factor is that dark matter is the dominant source of matter in the universe, but nonetheless only provides around 25% of the critical density. Not sufficient to result in a spatially flat universe. 3.2.3 The Cosmic Microwave Background It turns out that possibly the best means of measuring the mass density of the universe is the Cosmic Microwave Background. This is actually a very useful tool for several branches of cosmology and we will come back to it later in the course. Here we investigate briefly how it places limits on the curvature. The Cosmic Microwave Background (CMB) is the imprint of recombination of electrons and protons which occurred roughly 300,000 years after the big bang. The image essentially contains a measure of the structure in the universe at that time, and indicates that the universe was highly uniform. The scale of the inhomogenities in the CMB temperature map is only < δT T >∼ 1 × 10−5 ). None-the-less there does exist structure within the CMB, and this structure can place limits on the curvature of the universe. The size of these imhomogenities is set by time which they have had to grow. Because of the age of the universe at the time of CMB formation this sets a rather stringent limit on this size of vp × t, where vp is the speed of sound in the plasma and t is simply the age of the universe at CMB formation. vp is a significant fraction of the speed of light (∼ 0.6c), and implies that the largest structures in the CMB can only have reached physical sizes of ∼ 200, 000ly. Thus, by measuring the size of the largest angular structures in the CMB maps we can obtain a physical measure of the angular size distance at the redshift of the CMB (z = 1100). This makes the CMB anisotropies a standard ruler. The measured size of the CMB anisotropy can then be compared to the expectations under different cosmological models. Since the light has travelled through the entire universe it size is modified by the mass–energy of the universe. So, measuring its apparent size, and comparing to the known size enables the total mass energy of the universe to be measured. 22 Recent results from the Wilkinson Microwave Anisotropy Probe (WMAP), indicate that the universe indeed does attain critical density with wtot = 1.02 ± 0.02. 3.2.4 The flat universe When all the sources of mass energy within the universe are co-added it appears that the universe does attain (within observational errors) the critical density. However, the visible light of the universe is unable to explain this, since all the stars/dust/gas only create a few percent of the critical density. Even allowing for a component of the universe to be made up of dark matter ∼ 25% of the total density is accounted for. The rest of the critical density is in some largely unknown form, for which the term dark energy has been coined. This was first discovered from supernova observations, and we will return to dark energy, and its impact on the evolution of the universe in later lectures. 23 Chapter 4 The fluid and acceleration equations We have now derived the fundamental equation for the evolution of the universe - the Friedman equation - this, when solved allows us to compute the evolution of the universe under a myriad of different physical models. However, we are still not in a position to solve it, largely since it contains two unknowns the scale parameter (a(t)) and the density (ρ or ). We therefore need to search for more fundamental laws which will allow us to solve the Friedman equation. 4.1 The Fluid Equation The Friedman equation is basically a statement of energy conservation (total = potential + kinetic = constant). Equally, the first law of thermodynamics is similar, linking the heat flow in and out of a system (dQ) with a change in the internal energy (dE), the pressure p and the change in volume dV . dQ = dE + pdV (4.1) In a homogeneous universe dQ = 0 (it is an adiabatic expansion - adiabatic expansions do not increase the entropy of the universe since dS = dQ/T .) Therefore Ė + pV̇ = 0 (4.2) The universe can essentially be considered an expanding sphere, with co-moving radius r(t) = a(t)x, and therefore volume, V (t) = 4π a(t)3 x3 3 (4.3) so 4π 3 2 V̇ = x (3a ȧ) = V 3 ȧ 3 a (4.4) The internal energy is then E(t) = V (t)(t) 24 (4.5) The total rate of change of the internal energy of the universe is given by Ė = V ˙ + V̇ = V ȧ ˙ + 3 a (4.6) This can be written more simply, combining the above equations as V ȧ ȧ ˙ + 3 + 3 P a a =0 (4.7) or even ȧ ˙ + 3 ( + P ) = 0 a (4.8) and is known as the fluid equation. 4.2 The Acceleration Equation The acceleration equation is a third equation, which can be derived from the fluid equation and the Friedman equation. It tells us if the expansion of the universe is slowing down, or speeding up. The standard Friedman equation, multiplied by a2 takes the form ȧ2 = 8πGa2 kc2 Λa2 − 2 + 3c2 3 R0 (4.9) If we differentiate this we get 2ȧä = 8πG 2 2Λȧa (a ˙ + 2aȧ) + 2 3c 3 (4.10) Which, when divided by 2ȧa gives us ä 4πG a Λ = ˙ + 2 + 2 a 3c ȧ 3 (4.11) ȧ ˙ + 3 ( + P ) = 0 a (4.12) But the fluid equation Can be re-arranged to give a = −3( + P ) ȧ Which can be substituted into equation 4.11 to give the acceleration equation ˙ ä 4πG Λ = − 2 ( + 3P ) + a 3c 3 25 (4.13) (4.14) Note that any positive energy density contributes to a negative acceleration (the relative velocity between two points is getting less). The pressure component is due to the thermal motions of the particles which make up the universe. Normal baryonic matter exhibits a positive pressure, and also contributes to the universe slowing down. However, if the universe actually had a different pressure, say one which have (4.15) 3 The it would in fact have a positive acceleration (a tension). This might sound odd, but it is exactly what the cosmological constant provides (it has P = −). P < 4.3 and finally ........ equations of state We are now just one step away from having the mathematical machinery to describe (at least at a basic level) the evolution of the universe. What we need now is an equation of state, which enables us to solve our 3 equations (but only independent equations) in terms of only 2 unknowns, rather than the current three (a(t), (t), P (t)). Thankfully this can be fairly straightforward. Essentially we need to know a relation of the form P = P () (4.16) Generally in cosmology we write a simple form P = w (4.17) where w is a dimensionless number. For example an ideal gas (which it turns out isn’t such a bad approximation for the universe would have), P = ρ kT µ (4.18) Note that here k is the Boltzman constant, not the curvature constant. Because of relativity we need to be thinking in terms of mass-energy, rather than just mass, so that ∼ ρc2 , i.e. most of the mass energy comes from the mass of the particles, and not from their motion (you can check this just by calculating mc2 for a proton, and comparing that with its kinetic energy at a given velocity < c. Therefore P = kT µc2 (4.19) We can simplify this further by noting that for a non-relativistic gas the relation D 3kT = µ v 2 E (4.20) Thus if we define an equation of state parameter, w w= 2 v 3c2 26 << 1 (4.21) Then we can write the equation of state of a a non-relativistic gas (i.e. a matter dominated universe) as Pnonrel = wnonrel (4.22) In the relativistic case, which is for example where photons (or neutrinos) dominate the universe the equation of state is 1 Prel = rel 3 (4.23) i.e. w = 1/3. In general cosmologists can then distinguish between different model universes simply by differing values of the equation of state parameter. A matter dominated universe, to first order has w = 0. Radiation dominated universes have w = 1/3, and mildly relativistic systems have some value in between. It is interesting to note that there is nothing to stop negative values of w, which give rise to negative pressure (tension). For example any universe where 1 (4.24) 3 will provide an positive acceleration. Indeed the cosmological constant has w = −1, and is thought to be responsible for the currently observed acceleration of the universe. w<− The last couple of lectures have been hard going, but this is the basic maths we need to explain the universe. In the next few lectures we will apply these rules, and try to understand how the universe at large evolves. 27 Chapter 5 Model Universes 5.1 Matter Dominated For matter, the equation of state parameter w=0. Thus matter has no contribution to the pressure. Therefore the fluid equation can be re-written as 3ȧ 3ȧ ( + P ) = ˙ + = 0 a a This can be rewritten, using some slight of hand as. ˙ + (5.1) 1 1 d (3a2 ȧ + a ˙ 3 ) = 3 (a3 ) = 0 (5.2) 3 a a dt In, other words, the energy density of a matter dominated universe, falls off as the cube of its volume. When you think about it this really isn’t very surprising. We can now simplify the problem somewhat. In the case where κ = 0, the form of the Friedman equations doesn’t change when they are multiplied by a constant, because in essence they are normalised against the current scale factor. Therefore we are free to “scale” the Friedman equation to make it easier to solve, the most obvious scaling being a0 = 1 (i.e. the universe measured one universe across at the current time). In this case we can relate everything relative to this time, so; 0 a3 Substituting this into the Friedman equation gives = (5.3) 8πG0 1 (5.4) . 3c2 a This can formally be separated, but is easiest to solve by substitution. We make a first guess that the evolution of the universe is a power-law (as you may have already noticed, everything in astronomy is a powerlaw). So we take the form a(t) ∝ tq . In this case the exponents on each side of the equation are ȧ2 = 28 LHS = t2q−2 RHS = t−q (5.5) The solution is when these two match. Namely when q = 2/3. So a(t) = 5.2 t t0 2/3 (5.6) Radiation Dominated We can perform essentially the same analysis above for a radiation dominate universe. In this case w = 1/3 and the fluid equation becomes 3ȧ 4ȧ ( + P ) = ˙ + = 0 a a Which can be rewritten as for matter as ˙ + 1 1 d (4a3 ȧ2 + a ˙ 4 ) = 4 (a4 ) = 0 4 a a dt (5.7) (5.8) So the density of a radiation dominated universe falls off as 1/a4 , faster than a matter dominated universe. Since we know the universe contains both matter and radiation, the fact that they behave differently with increasing scale factor immediately implies that at early times in the universe the radiation component may be much more important than it is now. In any case, we are currently concerned with a universe containing only one component. In this scenario we can now solve the Friedman equation for a radiation dominated universe by assuming that the evolution follows a powerlaw a ∝ tq , and substitute in as we did in the case of matter domination. LHS = t2q−2 RHS = t−2q (5.9) The solution to this being q = 1/2, so in a radiation dominated universe the scale factor varies as a(t) = t t0 29 1/2 (5.10) 5.3 Cosmological Constant Dominated The case of a cosmological constant dominated universe is slightly different. In this case w = −1. Since, P = − and ˙ = 0, the Fluid equation 3ȧ ( + P ) (5.11) a essentially reduces to zero. What this means is that the energy density from Λ is constant (this is why it is called the cosmological constant). However, this is also slightly odd. Unlike a matter or radiation dominated universe in which the energy density varies with scale factor as one might expect, the actual contribution from the cosmological constant increases with the scale factor since the energy density per unit volume is constant. Considering a flat universe we have all the energy density provided by Λ so that ˙ + 3c2 H 2 8πG (5.12) 8πGΛ 2 a 3c2 (5.13) Λ,crit = and the Friendman equation is simply ȧ2 = which can be re-written as ȧ = H0 a (5.14) Which has the simple solution of a(t) = exp H0 (t − t0 ) 5.4 (5.15) Curvature Dominated This scenario sounds rather bizare; a completely empty universe consisting entirely of curvature. It is also known as the Milne Universe. In a universe consisting only of curvature the Friedman equation is −κc2 ȧ2 = (5.16) R0 This can be solved for different values of κ. The simplest is of course κ = 0, there is nothing (mathematically) wrong with this solution, but a static, flat and empty universe is not the most exciting object to study. If κ is positive then there is no solution allowed since once gets an imaginary ȧ. This makes sense since it is not possible to fill the universe with sufficient mass energy to make it positively curved, while also making it empty! If κ is negative then the solution becomes ȧ = ± 30 c R0 (5.17) and integrating gives a(t) = t t0 (5.18) For all of these scenarios we can also calculate the behaviour of the Hubble parameter H = ȧ/a. This yields Hmatter = 2/3t (5.19) Hradiation = 1/2t (5.20) HΛ = H0 (5.21) This essentially covers all the bases for single component universes. The various different possible evolutions of the universe in each of the cases described above is shown in Figure 5.1, note that not all of them require a big bang (i.e the cosmological constant dominated universe is infinitely old). Of course, the real universe is more complicated than these simple models, and we consider that next. 31 Figure 5.1: The possible fate of the universe under the assumption that it consists entirely of either mass, raditation, cosmological constant or curvature 32 Chapter 6 Multi-component universes 6.1 Different densities at different times The differing dependencies on the energy density derived in the last lecture point to something obvious, and yet highly important. At different times in the Universe’s history, it is dominated by different components When the Universe is very small, radiation, which is of very little importance today, is dominant, simply because r ∝ 1/a4 , at intermediate times matter is dominant, while as the Universe gets progressively bigger, the contribution from Λ, which doesn’t fall off as the Universe is expands, becomes the driving force in its evolution. This is interesting, but also problematic. It implies that the single component model Universe’s that we considered last lecture aren’t necessarily representative of the real Universe. Again we know this isn’t surprising, its obvious from the world around us that the Universe contains matter and radiation. It also seems likely from various experiments that it also contains Λ. Today we believe that the relative contributions of radiation, matter and cosmological constant can be summed up as part of our benchmark cosmological model. In particular: ΩM,0 = 0.23, ΩR,0 = 1 × 10−4 , ΩΛ = 0.73 (6.1) From this, and the relative dependencies on density with scale factor we can estimate the points at which the Universe switches between different periods of dominance. Since there is a flat Universe ΩM + Ω R + Ω Λ = 1 (6.2) In practice only two of these are ever important in our Universe (i.e. at the point at which matter and Λ are equivalent the contribution from radiation is negligible. The same is true 33 of Λ when radiation and matter are contribute equally). Hence, we can calculate the point of equality by where ΩM,0 = M /c and ΩR,0 ΩR,0 ΩM,0 = 4 , 3 a a = R /c . Hence, ΩR,0 = aM R ≈ 10−4 . ΩM,0 (6.3) (6.4) This is an epoch well before the formation of the Cosmic Microwave background, which occurred at a ∼ 10−3 . Similarly, we can calculate the epoch when matter and Λ were equal contributors to the Universal density ΩM,0 = ΩΛ,0 , a3 (6.5) so ΩM,0 ΩΛ,0 !1/3 = aM Λ ≈ 0.72. (6.6) The Universe can then be broken down into three distant eras, radiation dominance, matter dominance and Λ dominance. In most cases then a single solution may be effective, although there are switching points at which the scenario may be less straightforward. 6.2 Solving the Friedman equation for a multi-component Universe If we want to solve the Friedman equation for these multiple component Universe’s the it is useful to re-write the Friedman equation in terms of Ω, so that the classic Friedman equation, 2 ȧ a = 8πG κc2 Λ − + = H2 2 2 2 3c 3 R0 a (6.7) divided by H 2 looks like. 1= and since 8πG 3c2 H 2 8πG κc2 − 3c2 H 2 R02 a2 H 2 (6.8) = 1/c , this can be written as 1= κc2 − 2 2 2 c R0 a H (6.9) κc2 R02 a2 H 2 (6.10) or 1−Ω=− 34 An important point here is that the RHS cannot change sign. κ = −1, 0, 1, and do so always, for the entire life of the Universe. This leads to an important conclusion to start with. Changing the dominant components of the Universal energy density does NOT change its curvature Now, we can evaluate this at the current time, when a0 = 1. κ H02 (Ω0 − 1) = c2 R02 (6.11) Substituting this back into the Friedman equation gives H2 = 8πG H02 − (Ω0 − 1) 3c2 a2 (6.12) which, upon division by H02 is H2 1 − Ω0 + = 2 c,0 a2 H0 (6.13) Separating this into separate energy components gives. H2 (1 − Ω0 ) Ωm,0 Ωr,0 = 3 + 4 + ΩΛ + 2 a a a2 H0 (6.14) This is a general form of the Friedman equation, that can be integrated to obtain the solution for the dependence of a on t, for any arbitrary combination of energy densities (one can even easily add more exotic equations of state if needed). Multiplying by a2 and taking the square root gives an equation that can be integrated. Z t Z a H0 dt = 0 0 da h Ωm,0 aa + Ωr,0 a2 + ΩΛ a2 + (1 − Ω0 ) (6.15) i1/2 The main problem with this is that in most cases it must be integrated numerically. Analytical solutions can only be obtained in cases where there are one or two terms. For example, in the current Universe, where the contributions from matter and Λ are of the same order of magnitude, we have ΩΛ,0 = 1 − Ωm,0 , and the analytical solution is a 2 log H0 t = p a 3 1 − Ωm,0 mΛ s 3/2 + 1+ a amΛ 3 (6.16) where amΛ = Ωm,0 ΩΛ,0 35 !1/3 (6.17) Feel free to try and re-produce this as an exercise, but in practice you won’t be asked to solve the Friedman equation for more than a single component. Fortunately, this actually provides a pretty good estimate most of the time. After all, for most of the time, either a/amΛ >> 1 or a/amΛ << 1. In other words the solution reduces to that for either a matter dominated, or Λ dominated Universe. i.e. the above equations reduce to a(t) ≈ amΛ exp q (1 − Ωm,0 )H0 t (6.18) for a/amΛ >> 1, or 3q Ωm,0 H0 t)2/3 . (6.19) 2 So, while the behaviour at the transition is more complex, for most of the Universe’s history it behaves rather simply. This is shown graphically in Figure 6.1, and in many ways is the punchline of this lecture. a(t) ≈ ( The Universe consists of many components, but for the majority of its history a single one is dominant, and it expands or contracts based on that component alone. 6.3 Observationally: Measuring expansion and acceleration 6.3.1 The Hubble parameter - H 6.3.2 The deceleration parameter q In a sense, what we want to look at with the deceleration parameter is for an explanation of the expansion or contraction of the universe which is physics free, in other words, in an observable which is not determined purely by the physics driving the evolution, which we may or may not understand. In practice it is something that we can determine if we can measure a change in the Hubble parameter as we move away from the local Universe. To get a handle on this we can expand the form of a(t) around t = t0 . In other words, a(t) = a(t0 ) + da dt 1 da2 (t − t0 ) + 2 d2 t t=t0 ! (t − t0 )2 + ......... (6.20) t=t0 dividing this by a(t0 ) gives us 1 a(t) = 1 + H0 (t − t0 ) + q0 H02 (t − t0 )2 2 (6.21) where äa ä q0 = − 2 = − ȧ aH 2 q0 is potentially an observable, but is also a link to the make up of the Universe. 36 (6.22) Figure 6.1: The evolution of different components of the energy density with scale factor. As can be seen Λ = m at a ∼ 0.7), while m = r at a much earlier epoch a ∼ 3 × 10−4 . 37 The acceleration equation is 4πG ä = − 2 ( + 3P ) a 3c (6.23) This can be re-arranged into a form for q0 by dividing by H 2 and multiplying by −1 to give −ä 1 8πG = ( + 3P ) 2 aH 2 3c2 H 2 (6.24) Note that the form 3c8πG 2 H 2 is equal to 1/c , where c is the critical density of the universe. This allows us to re-write the acceleration equation in terms of the critical density parameter Ω = w /c , where w is the energy density in a component of the universe with equation of state parameter w. Summing over the possible components we get −ä 1X 1 ( + 3P ) = q0 = 2 aH 2 w c (6.25) Since the equation of state links P and , we can substitute in for P = w to get −ä 1X Ωw (1 + 3w) = q0 = 2 aH 2 w (6.26) Now, summing over the possible values of w for matter (0), radiation (1/3) and cosmological constant (−1) gives 1 q0 = Ωr,0 + Ωm,0 − ΩΛ,0 2 38 (6.27) Chapter 7 Recent Observations of the Universe at large 7.1 Testing the Benchmark model Over the past few lectures we have mainly been concerned with developing a model for understanding how the Universe expands or contracts based on the different components which contribute to its energy density at any given time. The natural question to ask is then if observations provide a handle to directly measure this, and if they support this model. We have seen in an earlier lecture that the benchmark Universe can be written on the basis of the contributions of the different components today • Ω0 = 1 • Ωm,0 = 0.27 • ΩΛ,0 = 0.73 • Ωr,0 ∼ 10−4 We are now interested in measuring how the Universe evolves from this position, and if the models we derived in the previous lectures are a good fit to the data. You can probably guess that they are, unless I have been maliciously wasting your time, but its still relevant to consider how we actually make those measurements. In previous lectures we have seen that different components of the density of the Universe have dominated at different times in the past. Radiation was dominant at early epochs, while matter dominated for the next 10 billion years. The cosmological constant (or whatever it is that is causing the Universe to accelerate) has only come to dominate in rather recent history. This means that the scale factor evolution has not always been the same, it has changed according to the dominant component of energy density in the Universe. We have also seen that we can estimate the point at which the scale factor will change by determining the tipping fro e.g. acceleration to deceleration. This final change happened at around z ∼ 0.7, and is quite abrupt. In comparison, the change from matter to radiation domination occurred very early in 39 the Universe, and was only associated with a small change in the way the scale factor evolves (t1/2 → t2/3 ). Given that this change occurred at an epoch we can never directly measure because it takes place before the formation of the CMB, we will have to content ourselves with measuring the change for matter to cosmological constant dominance. 7.2 Future tests of benchmark cosmologies The benchmark cosmology which we have described above represents a cornerstone of physics over the past century. However, it has by no means been a fixed model over that time. A hundred years ago the static model of the universe was the only game in town, until the discovery of Hubble’s law in the late 1920’s. Even after this we remained woefully short of the model we currently assume. Most notably because of the absence of a cosmological constant from observational models prior to the early-mid 1990’s. We are now in a period which some like to call the era of precision cosmology, but precision in cosmology is not necessarily what we would call precision in other fields. Indeed, the errors we currently have on cosmological parameters are broadly comparable to those which would be considered just about acceptable in a school laboratory were we talking about measuring electrical current etc. Therefore, despite our success in creating a self consistent picture of the universe there remains much to do if we are to truly understand it. Furthermore, despite having described the benchmark model in the previous chapters, we still know little about its actual constituents. In describing our benchmark model we make the assumption that the universe is 23% dark matter and 73% dark energy. That is to say that about 4% of the universe is made up of things that we broadly understand, the rest we do not. In fact, even most of the 4% which is made of baryons is actually invisible to us. We can understand how dark matter and dark energy impact the evolution of the universe since we can see this from observations, but we have only concepts of what they may be. This situation clearly needs to be resolved, and it is this which leads to the drive to improve our knowledge of cosmological models. Indeed, this is a cornerstone of NASA’s beyond Einstein programme (its current vogue), and will be a chief goal of at least one (and possibly) more future space observatories. The basic concept here is simple, to find a means of pinning down the cosmological parameters with greater accuracy than has been possible to date. We can potentially do this a number of ways, so I’ll discuss only a few possibilities here. 7.2.1 Large scale supernova searches One of the most successful tool for pinning down dark energy, and for general cosmographic use, has been type Ia supernovae. It has proved possible to calibrate these to reasonable accuracy (∼ 7% errors in distance) and this has made them the tool of choice for both current, and next generation endeavours. The aim of most supernova searches now is simply to locate large, homogeneously selected samples of type Ia supernovae which can be followed up. The work can be split into two main categories. The first is studies of relatively local supernovae. These studies aim to calibrate the standard candle relationships which are used to measure distances. There is a well known relation between the peak brightness of a SN Ia and its decay rate, known 40 as the Philips relation. In this the brighter the SN, the slower its decline. Correcting for this difference in lightcurve shape is what enables. SN Ia to be accurate distance estimators. Much work is continuing now using local samples of SN Ia to further refine this relationship, and also to test how well SNIa behave as standard candles at a variety of wavelengths. For example, when we look at optical wavelengths at z ∼ 1, we see light which has been redshifted by a factor of (1 + z), that is, visible light at z ∼ 1 actually originated in the ultraviolet in the rest frame of the supernovae, so we cannot simply assume that the behaviour observed at low redshift in the optical will be the same as at high redshift in the ultraviolet. 7.2.2 High resolution Cosmic Microwave Background maps Measuring the first peak of the Cosmic Microwave background provides one of the strongest constraints on the total density of the Universe, and in turn the fraction contributed by dark energy. There has been a significant investment in resources towards this aim over the past decade. The Wilkinson Microwave Anisotropy Probe (WMAP) is a NASA mission launched in 2002. Over the course of a 9 year mission it accrued detailed temperature and polarization data acros the CMB, pinning down the detailed cosmological parameters with great detail (the WMAP papers are amongst the most highly cited in all of science over the past decade). The cummulative 9 year data provides the tightest constraints on the total density parameter of Ω =. More recently, the European Space Agencies Planck satellite has provided an even more detailed and high resolution view, although only the initial temperature data has been provided to date. 7.2.3 Baryon Acoustic Oscillations Baryon acoustic oscillations are echoes of the cosmic microwave background imprinted on the large scale disribution of galaxies. Recall that the largest structures in the CMB have a size which is determined by the age of the Universe at the epoch of CMB formation, and the speed of sound in the plasma which dominated the early Universe. At decoupling the matter and radiation separate (i.e. the Universe ceases to be a plasma), so the photons free stream to us as the CMB, only impacted by the gravitational pull of the Universe as a whole. However, the baryons continues to evolve and interact, and proceeded to create large scale structure in the Univesrse. However, this large scale is frozen into the galaxy distribution. Large scale sky surveys such as the Sloan Digital Sky Survey have measured the galaxy distribution on these very large scales, and indeed seek a peak in the clustering power at large radii (much larger than typical clustering length of galaxies) – these are known as baryon acoustic oscillations (BAOs). Since this started as a standard ruler at the time of CMB formation, its measured size as a function of redshift provides us with a direct measurement of how the scale factor has evolved between the CMB and the epoch where the BAO is observed. Hence measuring the evolution of BAOs provides a powerful route to measuring the scale factor as a function of cosmic time, and hence of directly testing model cosmologies. However, it should be borne in mind that making such a measurement is extremely challenging, since it requires a detailed 3D map of the positions of galaxies in space. This is far from trivial even with modern technology. Secondly, it is limited by the number of oscillations that fit within our horizon at a given epoch, thus there 41 is a statistical noise limit which cannot be beaten even with perfect observations. 42 Chapter 8 Dark Matter As we have already discussed dynamical evidence for dark matter in the universe is widespread. The most notable evidence comes from the rotations curves of normal galaxies (like our own) which are flat at large radii, far beyond the extent of the visible disk. The expected orbital rotations scales as q vorb = GM (r)/r (8.1) and should fall off as r−1/2 beyond the visible light if M (r) is not continuing to increase. This indicates the presence of invisible matter beyond the radius of the majority of the stars, and a primary motivator for the presence of dark matter. However, this only works in systems which have ordered rotation, and this only comprises a few astrophysical objects (mostly spiral galaxies). Many others have stars on essentially randomised orbits, and so this approach cannot be used. Instead, for a system in equilibrium (which most can be assumed to be) we can make use of the Viral theorem. In doing this there are several observational considerations that should be considered (if doing the experiment properly), there are intentionally glossed over here for the sake of clarity. 2T + V = 0 (8.2) where T is the kinetic energy and V the gravitational potential energy. For a system of stars P1 2 1 2 T = 2 mv = 2 M hvi , where M is the total system mass and hvi is the velocity dispersion. The gravitational potential energy is given by 3GM 2 (8.3) 5R In practice this comes from the assumption of a spherically symmetric distribution of mass, which only feels the force internal to it (M < r), that in turn acts like a point source. In this system dm = ρ4πr2 dr and hence V =− dU = − GM (< r) ρ4πr2 dr r Since M (< r) = 34 πr3 ρ this becomes 43 (8.4) U =− 16 2 2 π ρ 3 Z R r4 dr = − 0 16 2 2 5 −3GM 2 π ρ R = 15 5R (8.5) Hence using the virial theorem gives M= 5 hvi2 R 3G (8.6) So, in principle if we measure the velocity dispersion (note that the measured and true velocity dispersions need some adjustment since we only measure the radial component), and the size of the object then we can determine its mass. In principle this should work for many systems e.g., star clusters, globular clusters, elliptical galaxies, galaxy clusters. Since we can also measure the luminosity of these objects we can then determine their mass to light ratio. This provides one of the few diagnostics we can make of the nature of dark matter, since we can determine how the dark matter distribution depends on the mass or size of a given object. As we saw in lecture 3, it is common in astronomy we define the Sun as our ultimate standard and define its mass to light ratio as unity M /L = 1. Table 8.1 shows the mass to light ratios seen in different types of astronomical systems. In general more massive systems (in terms of the mass of stars) exhibit relatively more dark matter, as evidenced by higher mass to light ratios. This suggests that the gravitational pull of larger systems is sufficient to ”trap” more dark matter. It also shows that dark matter is not uniformly distributed across space. An exception to this rule comes from the local population of dwarf spheroidal galaxies. These galaxies have stellar masses (i.e. the masses of the stars in the system) which are comparable to those of globular clusters, but are typically a factor of 10 larger. Like globular clusters they are mostly ancient systems, consisting of old stars. However, unlike globular clusters, which have no evidence for dark matter, the dwarf spheroidal galaxies are dominated by dark matter, with mass to light ratio’s of 100 or more at times. That systems of similar stellar mass but different size can have such different dark matter contents is suggestive that dark matter has some minimum scale size or tens of parsecs and masses of 107 M . Systems small than this do not have sufficient gravitational pull to hold onto dark matter. 8.0.4 Dark matter temperature This leads to the concept of a dark matter temperature, which really refers to the speed at which dark matter particles are travelling (in analogy to the speed that particles in a gas move as the gas is heated). The two types of dark matter that are generally discussed are hot dark matter (HDM) and cold dark matter CDM. In practice there can be a wide range of properties in between and warm dark matter, tepid dark matter and others have been proposed. Hot dark matter moves relativistically, and the classic example would be the massive neutrino. Because of its velocity it moves too rapidly to be caught up in collapsing peaks in the baryon density during structure formation. This results in hot dark matter particles being scattered broadly throughout the Universe, and not clustering in massive structures as is 44 Table 8.1: Masses an M/L ratios of common objects Object Globular Cluster Spiral (disc) Spiral (halo) Elliptical Galaxy Cluster Coma Cluster Dwarf Spheroidal M(stars)(M ) 106 1011 1011 + 109 − 1012 1010 − 1014 106 − 10 − 8 M/L 1 2-10 5-10 10-100 100+ 250 10-100 observed. Further, since the mass of haloes in galaxy formation models is strongly dependent on dark matter (since it is the dominant) source of mass. Small scale structure is effectively wiped out in hot dark matter models. In practice, a strong bound has been put on the contribution of hot dark matter, in particular the neutrino as Ων < 0.0072. In contrast cold dark matter readily falls into initially shallow potential wells and has a much larger influence on the process impacting galaxy formation. It can create the low mass haloes that are necessary for matching simulations to the observed Universe. However, the problem it faces is that it overpredicts the abundance of these halos, something known as the substructure crisis. There are numerous solutions to this problem that have been proposed, including altering the prescription for dark matter (e.g. self interacting dark matter), observational biases, or the presence of some haloes which simply contain few (or even no) stars. It is likely that one or more of these explanations are correct, and so the substructure crisis causes less concern that previously. Having accepted that cold dark matter is the likely explanation for the observed properties of the Universe at large scales, we are now faced with the question of what the dark matter actually is. There are two broad categories of models, those which arise from particle physics, and invoke Weakly Interacting Massive Particles (WIMPS) and those which come from astronomy, and use the (intentionally invented) name Massive Compact Halo Object (MACHO). 8.1 MACHOS MACHOS are essentially compact stars, which are sufficiently cool that their luminosity is insufficient to be observed, even with todays large ground based telescopes. The most likely MACHO candidates are: • Cool white dwarfs - ∼ 0.6 M • Brown dwarfs - ∼ 0.08 M • Neutron stars - ∼ 1.4 M • Black Holes - ∼ 3 − 104 M 45 • Other wierd objects, quark stars etc Although these objects are too faint for direct detection with telescopes it is possible to test what contribution they may make to dark matter by observing there effect on light from background stars as it passes close to the MACHO. This is a phenomena known as microlensing, which also has applications to extra solar planet research. The basic principle of microlensing is that as a MACHO moves across in front of a background star the light from the star is bent and magnified by the gravitational field. The light is deflected by an angle 4GM α= 2 (8.7) c b where b is the so called impact parameter and is the separation of the star and MACHO. This deflection can create a so called Einstein Ring, a perfect ring around the lensing object (although it is typically very small and not resolved for MACHO microlensing events). The size of this ring is 4GM 1 − x 1/2 θE = (8.8) c2 d x where x = dm /d, the ratio of the distance to the star d and the lens dm . θE is thus maximized for x = 0.5. This means that the maximum magnification and optical search is for MACHOs towards the Magellanic Clouds, where the background stars are roughly twice as far away as any candidate MACHOS. For typical parameters for the Magellanic Clouds −4 θE = 4 × 10 M arcsec M( ) !1/2 d 50kpc (8.9) In addition to the bending of the light into a Einstein Ring there is also amplification, while the background star is within the Einstein ring. Setting u = b/θE (8.10) the amplification is then A= u2 + 2 u(u2 + 4)1/2 (8.11) The total timescale of the lensing event (again for a typical event towards the Magellanic Cloud) is −1 dθE M 1/2 v ∆t = ≈ 90days (8.12) 2v M 200km/s where v is the velocity of the MACHO. Thus the length of the lensing event is dictated by the mass of the MACHO. The rate of MACHOS observed essentially sets their space density, and so the product of mass per MACHO and density yields the total mass in MACHOS, and can be compared with the measured dark matter density in galaxy halos. The punchline is that although some microlensing events have been observed, typically with lenses being low mass stars (brown dwarfs or white dwarfs) the total mass within is not enough to be responsible for the observed value of ΩM , but can probably make up only 20% of it. 46 8.2 WIMPS WIMPS are thought to be elementry particles are are predicted under certain variations of the standard model of particle physics. They must be electrically neutral in order to interact only via gravitational and potentially weak forces. They must be particles that are beyond the standard model since no standard model particle has the required properties. There are various attempts to extend the standard model which arise due to concerns within it. In particular about how one might include gravity, or about various fine tuning issues. In particular there is the Hierarchy problem – this refers to the fact that the Higgs vacuum expectation (vev - the average value of the field in a vaccum) lies at around 250 GeV, while the Planck scale, where quantum effects of gravity are needed is at 1019 GeV – it may well be expected that there is new physics in the massive energy range. The second problem is a fine tuning problem associated with the Higgs mass lying in the electroweak regime (125 GeV). Since it is expected that the electroweak and strong forces become a “Grand Unified Force” at energies of above 1016 GeV the cancellation to get a Higgs mass of only ∼ 100 GeV must be better than one part in 1014 , implausibly small. 8.2.1 Supersymmetry Supersymmetry is an appealing solution that allows the interchange of bosons and fermions. It means that each standard model particle has an supersymmetric counterpart. For example the sneutrino is a boson that is the superpartner of the neutrino. The most likely candidate for dark matter is the neutralino, which is a fermion. Rather complexly there are actually a family of neutralios which are superpositions of the superpartners of gauge bosons of the standard model (e.g. photinos, zibo (Z-boson super partners) with the super partner of the Higgs Boson (Higgsino). The lightest neutralino is expected to be stable, and may be present in large numbers within the Universe. Since it does not decay it is extremely difficult to detect. It is clear that the symmetry is not a perfect one. Superpartners do not have the same mass as their normal cousins (and only a difference in spin), but some mass hierarchy, otherwise they would have been discovered in large numbers via particle accelerators already. Their mass must therefore be in the range 100 - 1000 GeV. A final feature of supersymmetric partners is that they provide a new quantum number, R, that is conserved in multiplication, and that they a majorama particles, which means they can act as their own antiparticle. This is potentially valuable when it comes to searching for these objects, as we see below. 8.2.2 Axions Axions are created by extensions to the standard model containing the strong force. The physics behind the strong force actually allows for a strong electric dipole to be associated with the neutron. Clearly one has not been observed, and so a (broken) symmetry was developed that forced the neutron to be neutral (a similar symmetry forces the photon to be massless). If this symmetry is broken it allows for a low mass (µeV) particle to be produced, this is known as the axion. 47 8.2.3 Searches for particle dark matter Cryogenic detectors There are numerous ongoing searches for particle dark matter, which may manifest itself in many ways. The traditional route of undertaking such searches has been to use deep detectors, isolated from significant particle backgrounds on the Earth surface (e.g. cosmic rays etc). These detectors then search for dark matter, relying on the direct (weak) interaction of a dark matter particle with an atomic nucleus. This is possible since the normal interactions between atoms and incoming particles are mediated by electromagnetic interactions between the incoming particle and the electrons in the atom, such that nucleon interactions will be rare. When they do occur they can be detected via several general routes; • Phonon interactions – when a dark matter particle interacts with a nucleon in a crystal lattice, it rattles the lattice, creating vibrations which travel as phonons through the lattice and are potentially detectable. • Charge changes – The interaction that creates vibrations will also move charges away from their equilibrium positions since the nucleus will draw electrons with it. This means that dark matter created currents through the detector that can again be observed. • Temperature changes – The interactions within the sample necessitate the transfer of kinetic energy from the WIMP into the sample. This in turn will create a measurable change in temperature (for a sufficiently stable initial temperature, and accurate enough thermometer), and may be measured. • Cherenkov radiation – When nucleons are hit by the relatively heavy WIMP particles they receive a kick which imparts velocity to them. This velocity may be larger than the speed of light in the medium in which they reside. If this is the case as they slow they will give off Cherenkov radiation. Accelerator detectors In principle, particle accelerators whose collision energy exceeds the rest mass energy of a dark matter particle should be capable of producing particle dark matter. Once detected the detailed analysis of the resulting particle tracks should allow its existence to be inferred. In practice the weak interaction of dark matter particles means that they are unlikely to interact within the detector. However, they would still carry energy and momentum away from the instrument. Given this, a smoking gun for dark matter particles would be collision events which appear not to conserve energy and momentum. By reconstructing this missing energy and momentum it would be possible to constrain the properties of the particles themselves. Astrophysical detection In addition to Earth-bound searches there is also the possibility of conducting direct astrophysical searches for dark matter particles. While most naturally astrophysical observations are 48 sensitive to larger scale dark matter (such as MACHOs), various possibilities may allow the detection of dark matter association with astronomical sources. In particular, since neutralinos are majorana particles, interactions between neutralinos can cause annihilation and that the result must be a normal particle. e.g. χ+χ→γ+γ R = −1 × −1 = 1 × 1 (8.13) These particles may be photons or could equally be electron – positron pairs. If γ-rays are produced then their energy will be equal to Eγ = mχ c2 . If a γ-ray telescope observing the sky was sensitive at this energy then a line would be observed whose energy would directly provide the rest mass energy of the neutralino. There have been some suggestions by orbiting satellites (most notably PAMELA) that there is an excess of positrons (positrons are much rarer than electrons so easier to detect an excess or). This excess could be produced by dark matter annihilations, although other explanations such as position accelerations in pulsar winds also remain plausible. Similarly, axions may be produced by nuclear reactions inside the sun, and there is a recent suggestion that they may have been discovered based on a varying X-ray background observed by the XMM-Newton experiment, although such analysis is understandably complex, and awaits confirmation. 8.3 Modified Newtonian Dynamics F = mµ a a a0 (8.14) We do not know what µ is, but it is an empirical function that for low a becomes a/a0 and for high a tends to 1. a0 is a new fundamental constant of nature, the acceleration constant a0 = 1.2 × 10−10 m s−2 . This means in the high acceleration regime >> a0 , F = ma as usual, but at low accelerations F = ma2 a0 (8.15) With this functional form in place we can ask ourselves how the rotation curve at large radii in a galaxy may work. In this case a = v 2 /r and so ma2 GM m = a0 R2 and since a = v 2 /R in a circular orbit mv 4 GM m = 2 R a0 R2 49 (8.16) v 4 = GM a0 (8.17) In other words, the rotation at large radii will naturally result in a constant rotation velocity as observed. This does have problems though. Firstly, it is not a relativistic theory. The concept of an acceleration constant requires a frame against which the acceleration can be measured. It is therefore directly at odds with a relativistic theory. People have attempted to adapt this by creating TeVeS (Tensor Vector Scalar Gravity) 50 Chapter 9 Structure in the universe 9.1 The Jeans Mass At early times (i.e. viewing the Cosmic Microwave Background 300000 years after the big bang) the universe was very smooth, with variations of order 1 part in 105 , yet today the universe consists of individual stars, galaxies and planets. How is it possible to go between the two extremes? How does structure in the universe form? This is the question we want to address this lecture. In essence we are initially concerned with conditions to form gravitationally bound structures. Any overdensity in the Universe will see to collapse on a dynamical timescale. tdyn 1 = = (4πGρ)1/2 c2 4πG !1/2 (9.1) This is essentially a free-fall time scale, e.g. if you consider gravitational acceleration GM R2 a= (9.2) and a time to fall distance R = 1/2at2 , you get R∼ GM 2 t R2 (9.3) and for a spherical mass M = ρR3 , this equates to t∼ 1 (Gρ)1/2 (9.4) For the earth’s atmosphere, this dynamical time is approximately 9 hours. Clearly the atmosphere is not a puddle at our feet, and so something opposes this collapse. This is pressure. For a matter dominated Universe (matter is after all the thing collapsing) the equation of state parameter is given by w= kT µc2 51 (9.5) and signals will travel through the matter at the soundspeed (cs ). tpre = R cs where dP cs = c d (9.6) 1/2 (9.7) p i.e. since P = w this means the sound speed is cs = c (w). To avoid collapse a perturbation must respond to a gravitationally induced change more rapidly than the change can propagate. In other words tpre < tdyn . Setting these two equal gives a Jeans length (R = λJ ). c2 4πG λJ = cs !1/2 (9.8) This is a slightly simplified derivation. Gathering all the relevant factors of π etc. together yields the correct result. λJ = cs πc2 G !1/2 = 2πcs tdyn (9.9) Structure forms on scales that are larger than λJ , or its correspond Jeans Mass, MJ , 4 MJ = πρλ3J 3 Smaller structures are pressure supported. 9.2 (9.10) Formation in a radiation or matter dominated Universe Clearly, if we want to understand the evolution of the Universe we want to know the Jeans mass/length at any given time. Here we can note that the energy density (in a flat universe) is given by 3c2 H 2 8πG and that the age of the Universe t = 1/H, so = tH 1 = = H 3c2 8πG !1/2 (9.11) 1/2 3 2 = tdyn (9.12) substituting in we get λJ = 2πcs tdyn = 2π 52 1/2 3 cs 2 H (9.13) Now we can essentially work out the Jean’s mass or length as a function of the sound space √ c w. Radiation Dominated: w = 1/3, cs = 0.58c In this case λJ = 2π 1/2 1/2 3 1 c 2 3 H ≈ 3c H (9.14) Note, this is significantly larger than the horizon, (c/H). The corresponding mass is MJ ∼ 7 × 1018 kg, more massive than the largest galaxy clusters. In other words, structure doesn’t form in a radiation dominated Universe. This isn’t strictly true, since the crucial point isn’t the radiation domination, but the speed that the material is moving at. in practice this applies at any time while the matter and the radiation are coupled (i.e. before formation of the CMB). At the point when decoupling occurs, we move to a matter dominated Universe. Then cs (baryon) = kT µc2 1/2 c (9.15) We know kT = 0.26 eV directly from the CMB, so cs ∼ 10−5 c. At this point the density of baryons ρbaryon = 5 × 10−19 kg m−3 , and the Jeans Mass is MJ ∼ 105 M , roughly the mass of a dwarf galaxy. In other words, structure formation begins immediately after decoupling. In practice, the situation is slightly more complex than this. Most of the matter in the Universe is dark, it does not interact electromagnetically, and so isn’t coupled to the radiation in the early Universe, this material is extremely important, since it collapse from the beginning, and so the haloes that have become somewhat overdense in dark matter during the pre-CMB epoch go on to host galaxies and galaxy clusters. 53 Chapter 10 Inflation So far we have outlined a benchmark model for the universe, which provides a good description of the observations from the cosmic microwave background until the present day. However, while this description is powerful and very successful there remain some puzzles within it, and these require a novel explanation in the form of inflation. 10.1 The horizon problem When we view the cosmic microwave background we observe that it has essentially the same temperature across the entirety of the visible sky. That is, points opposite eachother on the sky have essentially the same temperature. The similarity of temperatures across the sky is indicative of thermal equilibrium, that is to say that the size of the universe at the time of the CMB shows the signs of being in thermal equilibrium. Yet we know that at this stage the hasn’t been time for signals to travel across the universe and thus for it to have come into thermal equilibrium. This is shown schematically in Figure ??. Our horizon is defined as the surface ct0 from our location in any direction, but two points opposite eachother on the horizon thus have a separation of 2ct0 , i.e. the light travel time from the two points is greater than the age of the universe. they have never been in causal contact, so how are they in thermal equilibrium? This is known as the horizon problem. 10.2 The flatness problem The flatness problem refers to the statement that the universe is flat. Again measurements from the CMB imply that Ω0 = 1.02 ± 0.02, but why is this the case? There is no special reason to believe that the universe should be flat, it could have taken any value, say 105 0, or 10−50 , unity is a very special value, and it is odd that the universe could (randomly) take that value. A common argument is that we don’t know for sure that the universe does have the critical density 1.02 ± 0.02 is not certainly 1. Perhaps it is just chance that the value is close to 1. However, this can be dismissed by considering how the density of the universe evolves with time if it is dominated by different components. We saw in Lecture 6 that the evolution of the density of the universe with time can be written as 54 1 − Ω(t) = −kc2 R02 a(t)2 H(t)2 (10.1) evaluated at the present moment this is simply 1 − Ω0 = −kc2 R02 H02 (10.2) where we have made the common assumption a0 = 1, diving the two equations gives 1 − Ω(t) H02 = 1 − Ω0 H(t)2 a(t)2 (10.3) We also know from earlier lectures that the Friedman equation for multicomponent universes can be expressed as H2 Ωm,0 Ωr,0 (10.4) = 3 + 4 2 a a H0 Ignoring the ΩΛ term since we are looking back towards the early universe where Λ is less important. Substituting this into 13.3 above gives 1 − Ω(t) = (1 − Ω0 )a2 Ωr,0 + aΩm,0 (10.5) where Ω0 = 1.02 ± 0.2, Ωm = 0.27, Ωr = 10−5 . We can now extrapolate this back in time. For example when a = 10−4 , and matter and radiation contributed equally then 1 − Ω(t) ∼ 10−5 . At the Planck time of 10−44 then 1 − Ω(t) ∼ 10−60 . In other words, any difference that we observe now get progressively small as time goes backward, so that if the universe is approximately flat now then immediately after the big bang it is flat to 1 part in 1060 . This means that it is extremely like universe is flat, not just nearly flat, since a universe which is flat remains so for ever. These two problems essentially related to apparent fine tuning of the universe under the standard model. Somehow it must have had the same temperature throughout, and have always been flat. This may be the case, but the probability of it is very low, and it can appear contrived if we cannot explain why the universe appears as it does. 10.3 Acceleration expansion at early times – Inflation The solution to these problems comes from a theory by Alan Guth, called inflation. It basically suggests that at very early times in the universe ∼ 10−36 s the universe underwent a brief period of exponential expansion, greatly enlarging our horizon and flattening the universe. Here is how it works. First, take the acceleration equation ä 4πG Λ = − 2 ( + 3P ) + a 3c 3 55 (10.6) and an equation of state P = we. Any value of w ¡ -1/3 gives an accelerating universe. So, if at early times for some reason the equation of state parameter of the universe became < −1/3, the universe would accelerate. For simplicity lets consider the case of a cosmological constant-like equation of state. The Friedman equation for a Λ dominated universe is 2 Λi 3 (10.7) a(t) ∝ eHi t (10.8) ȧ a = The solution to this is exponential expansion where Hi is the Hubble parameter at inflation, and is a constant for Λ dominance. So, if inflation kicks in at time ti and ends at time tf it dramatically impact the scale factor of the universe, so that a(t) = ai (t/ti )1/2 a(t) = ai exp (HI (t − ti )) t < ti (10.9) ti < t < t f (10.10) 1/2 a(t) = ai exp (Hi(tf − ti ))(t/tf ) t > tf (10.11) So the scale factor of the universe increases during inflation by a factor of a(tf ) = eN a(ti ) (10.12) If inflation begins at 10−36 s and ends at 1034 s then N = 100, an important point to note here is that while this is a very short time (10−34 s) it is actually 100 Hubble Times, so if the same were to happen today inflation would not end until the universe was 100 times its current age (roughly 1400 billion years old). The increase in scale factor is thus e100 ∼ 1043 , and so rapidly the scale factor has been stretched. This means that a region of space the size of our horizon prior to inflation becomes 1043 times the size of our horizon afterwards. In other words, regions which are in causal contact (and thus thermal equilibrium) are increased in size so much that they are now much larger than our horizon. Under this scenario it is entirely unsurprising that the universe has the same temperature. This solves the horizon problem. Now let us consider the flatness problem. As we had before 1 − Ω(t) = −kc2 R02 a(t)2 H(t)2 (10.13) During inflation H = Hi always, and so this can be simplified to 1 − Ω(t) ∝ 1 ∝ exp −2Hi t a(t)2 56 (10.14) Figure 10.1: A cartoon diagram of how inflation works. The universe may initially be significantly non-flat, and be evolving further away from flatness. When inflation kicks in the universe rapidly flattens and then remains flat for a long period (though it may not look so in log space!), eventually, depending on the components of the universe which dominate the universe may evolve away from flatness. so |1 − Ω(tf )| = exp (−2N )|1 − Ω(ti )| (10.15) If N = 100 then this means that during inflation the universe is flattened by a factor of 10−87 ! If the universe was anywhere within 87 orders of magnitude of being flat before inflation then it will be flat afterwards. This is not fine tuning, the universe really could have chosen any value to start with. This solves the flatness problem. There is another rather remarkable fact about inflation when it is considered. If the scale factor really increases by a factor of 1043 then tiny quantum fluctuations are enlarged phenomenally, these tiny variations suddenly become the size of galaxy clusters, and are, perhaps the source of the structure we see in the universe today. 57 Chapter 11 Baryogenesis and Nucleosynthesis Another key question is the origin of the particles (mainly atoms, electrons) that the universe is made up from. If we extrapolate the Universe back to early times, t ∼ 10−12 s, when the temperature was T ∼ 1 TeV (note that these energies are similar to those being created in the LHC, which is why people talk about it creating energies similar to those just after the big bang). In this scenario the Universe is radiation dominated and so t −1/2 T = 10 K 1s −1/2 t kT = 1MeV 1s 10 (11.1) 11.1 Baryogenesis At some (poorly constrained) point in this early Universe we set a first fundamental property of matter. Namely that the Universe is made of matter and not antimatter. We know this from observations today. Clearly the solar system, and indeed the Milky Way galaxy are made of matter. We have good reasons to believe that all of the rest of the visible Universe is also made of matter, since we do not see sufficiently large scale voids to explain regions of matter and antimatter, nor do we see obvious spatial clustering in photons which may originate from matter – antimatter annihilation. However, there is evidence that much of the Universe was at some point in the form of antimatter. This evidence primarily comes from the observed photon to baryon ratio. In the current Universe np /nb ∼ 109 (note that this is the number density of photons to baryons, not the energy density, these are low energy photons). If we go to a point in the early universe where kT > mb c2 , where mb is the mass of a baryon then the baryons are in equilibrium with the photons, and we might expect there to be somewhat similar numbers. γ+γ * ) b + b̄ 58 (11.2) This is a good time to set about making asymmetry, because all that you need to do is make a tiny fraction of additional baryons (1 part in 109 ). In other words we can define an efficiency η= nb − nb̄ nγ (11.3) However, the problem arises because we know very little about the physics that might introduce such a change, and very little of it has been physically tested (although this is one of the aims of the LHC). However, there are some conditions which can be imposed which may explain it. These are the so-called Sacharov conditions, which state that 1) Baryon number is not conserved. 2) CP violation 3) Interactions out of thermal equilibrium, probably because of the rapid expansion of the Universe. While the physics underlying these issues remains opaque, one thing that is clear is that from this early time onwards the Universe was dominated by matter rather than antimatter, and this is what we must consider when moving further. 11.2 Nucleosynthesis Most of the heavy elements in the universe have been formed in the core’s of stars, and subsequently ejected when the stars reach the end of their lives, either in the winds of red giants, or more explosively as supernovae ejecta. However, in the early Universe, the first generations of stars were clearly formed out of material that had not previously been processed by stellar nuclear fusion, and the bulk atomic properties of the Universe were actually set in the first few minutes. This is the subject of big bang nucleosynthesis, and is what we shall consider in this lecture. As the temperature passes through 1 MeV, the thermal energy of the Universe is rather comparable to nuclear binding energies, and so nucleosynthesis can be begin. At first, neutrons and protons are in equilibrium n + νe * ) p + e− n + e+ * ) p + ν̄e (11.4) Since p and n have different masses this can create/require energy Qn = mn c2 − mp c2 = 1.29Mev (11.5) n → p + e + ν¯e (11.6) Also, neutrons decay the halflife for this reaction is 900 seconds. 59 In equilibrium the number of protons and neutrons is given by the Maxwell Boltzman distribution nn = gn mn KT 2πh̄2 3/2 −mn c2 exp KT ! mp KT 2πh̄2 3/2 −mp c2 exp KT ! np = gp (11.7) where gn = gp = 2. We can now take the ratio of these to give nn = np mn mp !3/2 −(mn − mp )c2 exp KT ! (11.8) This can be simplified since mn /mp ∼ 1, to nn −Qn = exp np KT (11.9) This proves an exponential decline in the number of neutrons with decreasing temperature. If this equilibrium holds the after 5 minutes the nn /np ratio is 10−6 . Clearly this is not correct, since there are a reasonable number of neutrons present in the Universe today. The solution to this apparent contradiction is that the reactions which govern this equilibrium are weak interactions, the cross section of these is −47 σw = 10 2 m kT 1M eV 2 (11.10) This can be compared to the Thomson cross section of 7 × 10−29 m2 , which makes it clear that the weak interaction will only provide a high cross section for interactions at very high temperatures. The rate of interactions is then given by Γ = nσv Γ = nν σ w c (11.11) Since nν ∝ a−3 ∝ t−3/2 (11.12) σw ∝ T 2 ∝ a−2 ∝ t−1 (11.13) and The total rate of interactions is proportional to 60 Γ ∝ t−5/2 ∝ a−5 (11.14) In other words the reaction rate falls rapidly with time, and eventually Γ = H, so that there is less than one interaction per Hubble time. At this point we get something called Freezeout Tf reeze = 9 × 109 K tf reeze = 1 s At this time the neutron to proton ratio is approximately 0.2. At this point nucleosynthesis can proceed, and various reactions begin p + p → D + e+ + νe WEAK n + n → D + e− + ν¯e WEAK p + n → D + γSTRONG (11.15) This later reaction, as a strong reaction, dominates the observed rate. This builds up Deuterium, which then reacts further D + p →3 He + γ D + n →3 H + γ (11.16) and eventually Helium is made 3 3 He + n →4 He + γ 3 3 H + p →4 He + γ H + D →4 He + n He + D →4 He + p (11.17) Fusion continues until all of the neutrons are locked up in Helium. From the predicted ratio of nn /np we can then calculate the expected mass fraction in Helium. For each neutron we have 5 protons, and hence for each He nucleus we have two protons and two neutrons and hence an additional 8 unbound protons. This suggests that the mass fraction in He (Y) is equal to 4/12 = 1/3. In practice the actual number is Y=0.24. The reason for this is twofold. • Some neutrons are locked up in heavier elements • Some neutrons decay before they are locked up into nuclei 61 The heavier element reactions are 4 4 He + D →6 Li + γ He +3 He →7 Be + γ (11.18) One might wonder, given that the most stable nucleus is 56 F e, why does the reaction not proceed to give us a universe made of Fe and unbound protons. The reasons for this are twofold. Firstly, there are not enough neutrons to realistically do this, but more importantly there is not enough time to build these more complex elements. The reaction 4 He +4 He →8 Be + γ (11.19) leads to an unstable element, which decays in 10−16 s, it is very difficult to get beyond this in straight nucleosynthesis. In stars, these heavier elements are built up through chains of reactions known as the triple−α processes. However, this process is slow. Since the window for nucleosynthesis in the early Universe is short (15 minutes) there simply isn’t enough time to build up the heavier elements. Hence at the end of nucleosynthesis, the Universe is dominated by Hydrogen and Helium, with only traces of heavier elements. 62 Chapter 12 The Big Bang Throughout this course we have discussed cosmology entirely in the context of the big bang model for the universe. Although in many ways it was inflation which put the ”bang” in the beginning of the universe it is not inflation which brought the universe into existence. Despite the very strong evidence requiring a big bang, and the great success of the model there remains little insight into what happened at the time of the big bang itself. One of the big questions, both in physics and philosophy is therefore what caused the big bang? This is a question which is very difficult to address for a variety of reasons, nonetheless there do exist a number of possible mechanisms for creating the big bang, and we will briefly discuss them here. 12.1 The Planck scale An important scale in the early universe is known as the Planck scale, defined as Λp = Gh̄ c3 1/2 Gh̄ c5 1/2 ∼ 1.6 × 10−35 m (12.1) ∼ 5.4 × 10−44 s (12.2) and the associated Planck time tp = Once the universe shrinks below the Planck scale the particle wavefunctions overlap one another and it becomes impossible to consider gravity in a classical (or even general relativistic way). This means, that without a theory of quantum gravity (one of the great unsolved problems in physics) it is difficult, if not impossible to extrapolate beyond this time. In essence this means that models for the big bang are by necessity uncertain. 12.2 Vacuum fluctuations The most commonly discussed concept for the creation of the big bang is the Vacuum fluctuations. This basically states that it is possible to borrow energy from a vacuum to create a 63 particle – antiparticle pair, which they continues to create further particles, until ultimately it creates the universe we see today. This is allowed because of the uncertainty principle which states that h̄ ∆E∆t > (12.3) 2 and provides a viable (if far from certain) possibility for the big bang. 12.3 Branes and the Epkyrotic Universe Alternative models for the big bang come from recent attempts to derive so called Theories of Everything (TOEs). These include multidimensional theories such as String Theory and MTheory which envisage each elementary particle as a sting or membrane oscillating in 10 or 11 dimensional space. If the universe does contain a greater number of dimensions than the classical 4 (3 spatial plus one time) then it creates a new field of String Cosmology, and a the possibility of new explanations for the Big Bang. One of these models, postulated by Neil Turok and collaborations at Cambridge is that the Big Bang was caused by the collision of two 11 dimensional branes. The ripples in the space time spreading out in our 4-dimensional space time then create our universe. In this model the Friedman equation is modified to H2 = 3c2 1+ 8πG 2λ (12.4) where λ is the so called brane tension, which can provide a source on energy in the universe, but can also introduce degeneracy. One interesting consequence of this model is that the universe itself may be cyclic, with the big bang happening multiple times, with the clock restarting after every brane collision. This has also been coined as the ”Big Splat” scenario. These different models for the big bang remain very uncertain, although future space missions which might detect gravitational waves can distinguish between the two possible models briefly illustrated above. 64 Appendix A The universe on a side of A4 65 Distance The universe on a side of A4 MACHO and WIMP searches microlensing + gammaray D H0 v= Hubble’s law obeyed in local universe velocity a α eHt ω = -1 Figure A.1: The ΩΛ,0 = 0.73, Ωm,o = 0.27, Ωr,0 = 10-5 H0 = 72 km s-1 Mpc-1 Spatially flat Ω0 =1 Type Ia supernova searches Λ 66 Jeans mass ~105 Msol. Stars/Galaxies form Radial distance 10 ~Gyr - Matter- Cosmological constant equality Observational challenges/facts Key points/ideas Key periods in Universe history Rotation curves flat: Dark Matter Dominant: Probably Cold On large scales universe is homogeneous and isotropic (Cosmological Principle) Acceleration Horizon size Nucleosynthesis N+p => D D+p => 3He D+D => 4He First stars/galaxies Time (total age = 13.5 Gyr) Matter a α t2/3 ω=0 velocity Inflation ~300,000 years Cosmic Microwave background (e+p => H + γ) Fluid Peak of star formation Jeans mass >1018 Msol. Galaxies can’t collapse Big bang Vacuum fluctuations? Brane collisions? Friedman a α t1/2 ω = 1/3 10-36s - increase scale factor by ~1040 Solves flatness problem horizon problem a α eHt ω = -1 Radiation Energy densities εΛ = const εm α 1/a3 εr α 1/a4 Matter-radiation equality Appendix B Problems and Answers B.1 Problem Set 1a) A galaxy has a measured recession velocity of 5000 km s−1 , what is its; i) redshift ii) distance b) Show that, in any uniformly expanding space the Hubble constant can be attributed to the current rate of change of the scale factor ȧ/a. c) Explain what is meant by the terms standard candle and standard ruler. 2) a) Show that in a universe consisting purely of matter, that the universe will be flat if the matter density is c = 3c2 H02 8πG (B.1) b) Using the fluid equation show that the energy density of matter varies as 1/a3 (you may assume P = ω, where ω = 0). c) Hence, solve the Friedman equation, assuming that the scale factor has a power law dependence of a(t) ∝ tq . What is the value of q? d) How does H evolve? What does this mean for the evolution of the critical density of the universe? 3) The Friedman equation, in terms of multiple component universes can be re-written as 67 H2 Ωm,0 Ωr,0 (1 − Ω0 ) = 3 + 4 + ΩΛ + 2 a a a2 H0 (B.2) a) Explain the origin of each term. b) Given that the universe is spatially flat, with Ωm,0 = 0.27 and ΩΛ,0 = 0.73, at what scale factor were the two equal? c) What redshift does this correspond to? d) Perform the same exercise for matter and radiation, assuming Ωr,0 = 8.4 × 10−5 . e) At very early times the universe was dominated by radiation only. Using the equation above (or otherwise), show that in this phase its expansion was governed by a(t) ∝ t1/2 . 4) a) Explain what is meant my the terms horizon problem and flatness problem. b) Given that the Friedman equation can be written as 1 − Ω(t) = − κc2 R02 a(t)2 H(t)2 (B.3) Show how the density of the universe (1 − Ω(t)) evolves for universes dominated by matter, radiation and cosmological constant. c) How does a brief period of acceleration in the early universe solve the horizon and flatness problems? d) If the energy density and pressure of the inflaton field are given by 1 φ = φ̇2 + V (φ) 2 1 2 Pφ = φ̇ − V (φ) 2 (B.4) Under what conditions will the inflaton field mimic a cosmological constant with equation of state parameter ω ∼ −1. e) Sketch a potential which may satisfy this criteria. 5) a) For a given component of the universe, with equation of state parameter ω, show that the fluid equation can be re-written as d da = −3(1 + ω) , a 68 (B.5) and hence that = 0 a−3(1+ω) . (B.6) b) By solving the Friedman equation for this generic equation of state, assuming a power-law dependence (tq ), show that the scale factor evolves as a(t) ∝ t2/(3+3ω) , (B.7) assuming ω 6= −1. 6) a) Outline the evidence for dark matter in galaxies and clusters. b) Show that the rotation curve of a galaxy should follow the form s vorb = GM . R (B.8) c) Two possible origins of dark matter are WIMPS and MACHOS, what does each term mean? Give an example of a possible WIMP and MACHO candidate. d) Modified Newtonian Dynamics (MOND) states that Newton’s law should be modified to the form, a F = ma. (B.9) a0 in very low accelerations. Show that in this case MOND predicts flat rotation curves in galaxies. B.2 Problem Set with Answers 1a) A galaxy has a measured recession velocity of 5000 km s−1 , what is its; i) redshift Redshift is simply z = v/c = 0.017 ii) distance Hubbles law state v = H0 D, or D = v/H0 = 70 Mpc b) Show that, in any uniformly expanding space the Hubble constant can be attributed to the current rate of change of the scale factor ȧ/a. Suppose the triangle shown in Figure 1 expands uniformly. In this case the relative shape of the triangle is preserved in time. i.e. r12 (t) = a(t)r12 (t0 ) r23 (t) = a(t)r23 (t0 ) 69 Figure B.1: A triangle of galaxies, expanding uniformly. r31 (t) = a(t)r31 (t0 ) (B.10) Where a(t) is the so called scale factor which will be important later. Imagine an observer in galaxy one. The relative velocities observed for galaxies 2 and 3 are dr12 = ȧr12 (t0 ) = dt dr31 v31 (t) = = ȧr31 (t0 ) = dt v12 (t) = ȧ r12 (t) a ȧ r31 (t) a (B.11) Where we have made the obvious substitution that r12 (t0 ) = r12 (t)/a(t). The form of this equation is clearly that of Hubble’s law. In this case value of H0 can be equated with ȧ/a, at the time t = 0 (i.e. the current time). More generally the value H is known as the Hubble parameter, and is variable with time. c) Explain what is meant by the terms standard candle and standard ruler. A standard candle is simply an object of known luminosity, from its flux we can therefore measure its distance, or, more formally its luminosity distance. A standard ruler is an object with a known length, therefore by measuring its apparent dimensions its true distance can be calculated. Both of these are functions of the assumed cosmological model. 2) a) Show that in a universe consisting purely of matter, that the universe will be flat if the matter density is c = 3c2 H 2 8πG 70 (B.12) Take the Friedman equation with zero cosmological constant. Spatially flat implies that k = 0. So, if this is the case the Friedman equation reduces to 2 ȧ 8πG = a 3c2 Which can be re-arranged to give the required form above (B.13) b) Using the fluid equation show that the energy density of matter varies as 1/a3 (you may assume P = ω, where ω = 0). The fluid equation simply states ȧ ˙ + 3 ( + P ) = 0 a (B.14) ȧ ˙ + 3 = 0 a (B.15) 1 2 2 3a ȧ + a ˙ =0 a3 (B.16) since ω = 0, this reduces to This can be rewritten as which is equal to 1 d (a3 ) = 0 (B.17) a3 dt Since a > 0 then the solution for this equation is a3 =const, or ∝ 1/a3 . c) Hence, solve the Friedman equation, assuming that the scale factor has a power law dependence of a(t) ∝ tq . What is the value of q? The Friedman equation reduces to 2 ȧ a = 8πG 3c2 (B.18) If we assume that a(t) ∝ tq and ∝ 1/a3 , then we can equate each side of the equation in terms of dependence on q, so: 8πG 1 (ȧ)2 = (B.19) 3c2 a The LHS is proportional to t2q−2 and the RHS is proportional to t−q . This can be solved in q = 2/3. d) How does H evolve? What does this mean for the evolution of the critical density of the universe? 71 H = ȧ/a, = 2/3t 3) The Friedman equation, in terms of multiple component universes can be re-written as H2 (1 − Ω0 ) Ωm,0 Ωr,0 = 3 + 4 + ΩΛ + 2 a a a2 H0 (B.20) a) Explain the origin of each term. The first term refers to the expansion of the universe, and refers to the Hubble parameter at an arbitary time (t) compared to its value at the current time (t0 ). The second term describes the evolution of the component of the universe which consists of matter ∝ 1/a3 The third term describes the evolution due to radiation ( ∝ 1/a4 ) The fourth term describes the evolution due to the cosmological constant ( = const) The final terms describes the total density of the universe, or its curvature. b) Given that the universe is spatially flat, with Ωm,0 = 0.27 and ΩΛ,0 = 0.73, at what scale factor were the two equal? The two are equal at scale factor a = Ωm,0 /ΩLambda,0 )1/3 ∼ 0.7 c) What redshift does this correspond to? Since a = 1/(1 + z), z ∼ 0.4 d) Perform the same exercise for matter and radiation, assuming Ωr,0 = 8.4 × 10−5 . e) At very early times the universe was dominated by radiation only. Using the equation above (or otherwise), show that in this phase its expansion was governed by a(t) ∝ t1/2 . 4) a) Explain what is meant my the terms horizon problem and flatness problem. The horizon problem refers to the fact that all points along our horizon in the cosmic microwave background have very similar temperatures, indicating they were once in thermal equilibrium. However their distance from eachother on opposite sides of the sky is twice the horizon distance (2c/H0 ), and so they will never be in contact. The flatness problem refers to the measurement of the universe as flat. Since the Universe could have taken any value. Further, differences from universal flatness as a function of time mean that if the universe is flat now (ΩT = 1.02 ± 0.02) then it was even closer to being flat at early times. Why? b) Given that the Friedman equation can be written as 1 − Ω(t) = − κc2 R02 a(t)2 H(t)2 (B.21) Show how the density of the universe (1 − Ω(t)) evolves for universes dominated by matter, 72 radiation and cosmological constant. c) How does a brief period of acceleration in the early universe solve the horizon and flatness problems? d) If the energy density and pressure of the inflaton field are given by A brief period of superluminal expansion solves this problem since, as you have shown above for matter and radiation differences from flatness increase over time, but for the cosmological constant they decrease. Therefore, a brief phase where the cosmological constant was dominant flattens the universe. Further, since the expansion at this time was vastly superluminal is stretches regions which were in causal contact prior to the accelerated expansion, and leaves them out of causal contact at later times. 1 φ = φ̇2 + V (φ) 2 1 2 Pφ = φ̇ − V (φ) 2 (B.22) Under what conditions will the inflaton field mimic a cosmological constant with equation of state parameter w ∼ −1. Since P = w, and for a cosmological constant P = − (w = −1) this can be achieved if 2 ˙ phi << V (φ) e) Sketch a potential which may satisfy this criteria. See lecture notes for lecture 13 5) a) For a given component of the universe, with equation of state parameter ω, show that the fluid equation can be re-written as d da = −3(1 + ω) , a (B.23) = 0 a−3(1+ω) . (B.24) ȧ ˙ + 3 ( + 3P ) = 0 a (B.25) and hence that The Fluid equation states that 73 and can be rewritten using P = we to give ȧ d da 1 ˙ + 3 (1 + w) = + 3 . (1 + w) = 0 a dt dt a (B.26) d da = −3 (1 + w) a (B.27) which can be simplified to Integrating this gives da a (B.28) log = −3(1 + w) log a (B.29) = a−3(1+w) (B.30) Z d = −3(1 + w) e Z or which can be further simplified to b) By solving the Friedman equation for this generic equation of state, assuming a power-law dependence (tq ), show that the scale factor evolves as a(t) ∝ t2/(3+3ω) , (B.31) assuming ω 6= −1. The Friedman equation for a flat universe is 2 ȧ a = 8πG 3c2 (B.32) Substituting in from the above question 8πG (−1+3w) a (B.33) 3c2 Now if we assume a powerlaw dependence of a ∝ tq , we can get the value of q by equation the LH and RH sides, so: 2q − 2 = −q − 3wq (B.34) ȧ2 = Which can be trivially solved for q to give q= 2 3 + 3w (B.35) 6) a) Outline the evidence for dark matter in galaxies and clusters. Dark matter is primarily known to exist because of its dynamical (gravitational) impact on galaxies. Principally the rotation curves of galaxies at large radii are flat, white the velocity dispersions in clusters are much larger than expected based on the observed luminous matter, suggesting a large dark component 74 b) Show that the rotation curve of a galaxy should follow the form s vorb = GM . R (B.36) equating the centripetal acceleration with gravitational attraction GM v2 = 2 R R (B.37) which can be re-arranged s vorb = GM . R (B.38) c) Two possible origins of dark matter are WIMPS and MACHOS, what does each term mean? Give an example of a possible WIMP and MACHO candidate. WIMP stands for Weakly Interacting Massive Particle, and an example of this is the neutralino, MACHO stands for Massive Compact Halo Object, an example is a black hole. d) Modified Newtonian Dynamics (MOND) states that Newton’s law should be modified to the form, a F = ma. (B.39) a0 in very low accelerations. Show that in this case MOND predicts flat rotation curves in galaxies. Again equal centripetal acceleration, but this time allow for the additional term included via MOND. or a2 GM = 2 a0 R √ GM a0 a= R (B.40) (B.41) If this is centripetal acceleration = v 2 /r then v2 = r √ GM a0 R (B.42) v = (GM a0 )1/4 (B.43) or Which is a constant value. 75