Measurements of recent changes in the expansion rate via supernovae, published by myself in 1998with the High-z Team (Riess et al. 1998) and the Supernova Cosmology Project in 1999, first revealed that the expansion of the Universe is now accelerating.  At the time, the High-z Team began (Schmidt et al. 1998) the Universe was expected to be decelerating under the gravitational braking applied by its own mass, so this finding was an enormous surprise.  The extra expansion of the Universe in recent times is revealed by the excessive faintness of distant type Ia supernovae, whose brightness calibrates their distances.  Type Ia supernovae are identified as the end result of a carbon-oxygen white dwarf star which, as the victim of generosity by its binary companion, grows in mass until it reaches the Chandrasekhar limit.  At this critical mass, the white dwarf is engulfed by a runaway thermonuclear explosion, which reaches a peak luminosity of about 4 billion times the Sun’s.  These supernovae are so bright they can be seen with the Hubble Space Telescope 75% of the way across the Universe.  Their uniform brightness makes them excellent tools for measuring astronomical distances.

Cosmic Growth Spurts

When a distant SN Ia erupts, its light begins a long, isotropic journey to our telescope. The ongoing expansion of the Universe stretches the wavelengths of its propagating light to the redder part of the electromagnetic spectrum. The degree of this `redshift’ provides a direct measure of the change in the size scale of the Universe during the journey of the light. The distance to a supernova is measured by comparing its apparent and intrinsic brightness and reveals the time over which that signal has traveled at the speed of light. Thus, every supernova’s measured redshift and distance records the past change in scale over the inferred time interval, or taken together, the expansion rate. A large number of these supernovae record the expansion history for the Universe just as a series of marks on a doorframe are used to record the growth spurts of a child in their youth. In this analogy, the supernovae provide the marks and telescopes the means to see them.

Ground-based telescopes (with a little help from the Hubble) were responsible for the measurements of the recent growth history that revealed the surprising accelerated growth rate for the Universe.

The reason this was so surprising was because the self-attraction of all of the matter in the Universe should act as a natural retardant to the expansion. Thus, the growth rate of the Universe was expected to be decelerating. Imagine how surprised you would be if you tossed a ball in the air and it went up, not falling back on Earth again! Your expectation that the gravity of the Earth should pull the ball back to Earth is the same cosmologists had, in predicting the slowing expansion. You might think that gravity cannot cause objects to repel each other, but Einstein’s theory of gravity, called General Relativity (GR), actually can! In fact, shortly after developing his theory in 1916, Einstein himself was not bashful in invoking this exotic, repulsive mode of GR to explain his `incorrect’ belief that the Universe was holding back its own crushing gravity to remain static in size. He called this repulsive characteristic the cosmological constant. Quantum Mechanics has a more physical name for this feature, vacuum energy, and predicts it should exist. Vacuum energy has negative pressure (you must do work to expand the Universe’s inventory of the vacuum), and it is this property which gives rise to repulsive gravity. Einstein, upon learning from Edwin Hubble that the Universe was not static, but rather was expanding, discarded this notion as unnecessary. However, it is back with a vengeance today as this, or something alike, is needed to explain the observed data and the cosmic acceleration. Indeed, all such incarnations of energy with negative pressure are called dark energy.

The Deceleration test

But, what if we were being fooled by the supernovae into only thinking there were dark energy? How could this occur? In the late 1990’s it was suggested that (large grain) grey dust could have been ejected from galaxies at early times and now scatter SN light during its passage through intergalactic space, thus causing the SNe to appear too faint for their distance. Another concern involved an evolution of the SN luminosity towards fainter events with time in the past. Fortunately, it is possible to test these astrophysical hypotheses for the dimming of distant SNe Ia against the cosmological hypothesis of dark energy with negative pressure. Because dark energy is so weak, its repulsive gravity only dominates in recent times when the attraction of dark matter is diluted by expansion. But at z>1, when the Universe was more than 9 billion years old years old, the cosmological constant’s influence would be at least 8 times weaker than at present (the ratio scaling approximately as (1+z)3 and we expect attractive gravity to dominate). Thus, the Universe must decelerate at higher-redshifts (younger ages) and this should cause a relative brightening of the supernovae. The apparent brightness of ever higher redshift SNe Ia can distinguish between boring dust or evolution and mind-boggling dark energy. This is a test worth doing!

This measurement is easier considered than done. Only Hubble with ACS and NICMOS can routinely find and deliver measurements of SNe Ia at z>1 (despite valiant but unsuccessful attempts from the ground). Our team, the Higher-z Team, has led the effort to do just that, beginning shortly after the installation of ACS in 2002 and halting the day it failed in January 2007. Our basic strategy has been to repeatedly image 30 pointings spread over 2 fields (GOODS South and North) in the z-band (red filter) and one or two bluer colors. Using a combination of photometric selection (SN IA are UV faint) and host photometric redshifts, we were able to identify reliable candidates for photometric and spectroscopic follow-up. On average, we find one useful SN Ia at z>1 in every 10 ACS fields searched.

After 3 years we have measured 135 SNe of all types, 50 of these are SN Ia and about 25 are at z>1. These SNe Ia approximately double the redshift range over which the expansion history has now been mapped (Riess et al. 2004 and 2007)

As shown in Figure 1 inset, the data are consistent with an epoch of decelerating expansion preceding the recent accelerating expansion and inconsistent with the aforementioned, potential systematic errors.

Figure 1

In addition, close scrutiny of these earliest SNe Ia show they behave like their recent brethren, further validating their use across the cosmic time (see Fig 2). Thus, our confidence that we are on the right track has grown substantially.

Figure 2

There are further reasons for optimism, coming from recent, non-supernova measurements which confirm the reality of dark energy from fluctuations in the CMB, combined with any one of the following: LSS, BAO, H0, and weak lensing.

Long-lived Dark Energy

However, we have really just literally and figuratively reached the end of the beginning for acceleration and dark energy. Our quest now is to understand the nature of dark energy, i.e., the physical principles behind its existence.

To do this, we need either a complete and satisfying explanation or in the absence of that, a comprehensive set of candidate explanations and the experimental means to distinguish between them. Until a truly great theory comes along, we are forced to follow the latter course.

Initially we are attempting to compare the most apparent characteristics of dark energy, its repulsive strength and its longevity. These properties are contained the ratio of the (negative) pressure to energy density (called the equation of state, w) and its time evolution. The value of w at any time determines the size of the boost in the expansion rate we can expect to see at that time and as a direct consequence, the change in the dark energy density. If w=-1, at all times, then we can see the Universe always increases in energy by the same amount of work it does against the negative vacuum pressure to increase its volume. This is the physical signature of a constant (zero-point) energy in the vacuum (e.g., the quantum mechanical oscillator) and mathematically represents the same cosmological constant Einstein invoked.

A topological defect in the fabric Universe like flat membranes of mass-energy would only change its potential energy in response to the increase in the area from expansion (i.e., w=-2/3).

Alternatively, dark energy may be a transient phenomenon resulting from a low-energy potential field, where such fields are invoked (and observed) in particle physics to explain the way a force interacts with a particle. The added potential energy from the field created as space expands has the same effect as vacuum energy, except with an equation of state which varies (hence not always -1) due to the changing amount of work done by the Universe on space as the field grows or decays. This explanation is not without precedent, as a similar (though much higher energy) field, the inflation field, is invoked in the Inflation Theory to produce an accelerating, super-luminal expansion of the Universe which flattens the geometry of space and drives causally connected regions beyond the horizon, solving the flatness and horizon problem together.

If there is an analogous “acceleration” field, dark energy itself may be changing its strength, with its longevity limited. Some theories suggest that this field may be coupled to the Universe in such a way as to mirror the dominant form of mass-energy at the time. Thus, dark energy would behave like dark matter (w=0) in the past and like radiation (w=1/3) before that. Distinguishing between these scenarios (and others too numerous to discuss here) would provide an invaluable clue to theorists who are trying to understand how gravity operates on the largest scales and perhaps even how to unify quantum mechanics with General Relativity (i.e., Quantum Gravity), the holy grail of theoretical physics, which eluded Einstein for 30 years.

The new sample of SNe Ia at z>1 from HST provides one of the few new clues about dark energy, crude though it is. By using other techniques like the CMB and BAO to account for the expected degree of deceleration, the SNe Ia at z>1 reveal a nascent dark energy, just beginning to make its presence felt with w(z>1)<0. Thus, dark energy appears to have been in place with its defining characteristic (negative pressure) at least 9 billion years ago and was neither missing (as would be expected for a rapidly changing field) nor mimicking the gravitational behavior of matter or radiation. Vacuum energy or Einstein’s cosmological constant passes this test, a test which they could have failed. This is exactly the kind of test we need to perform, only with much enhanced precision in the future. A plethora of new tests and enhancements of older ones are being sharpened to pin down the nature of the ghostly dark energy. As Hubble wrote of the initial quest to measure cosmic expansion, “The search will continue. Not until the empirical resources are exhausted, need we pass on to the dreamy realms of speculation.”