Science and Technology

Science Case


The GMT Science Case has evolved over the course of the project. It has been influenced by the 2010 Decadal Survey’s report “New Worlds, New Horizons in Astronomy and Astrophysics” but has been updated to reflect new discoveries and scientific priorities. The 2018 version of the GMT Science Book is now available.

The GMT Science Book focuses on those areas of frontier science best explored with a large aperture ground-based telescope. The book describes the transformative impact that the GMT will have on areas spanning observational astrophysics—from exoplanets around neighboring stars to the formation of the first, most distant stars, galaxies, and black holes in the universe. The first chapter also describes the GMT itself, explaining its unique design and capabilities, including the first-generation instrument suite that has been chosen to maximize the GMT’s scientific impact during early operations. This book is accessible to a wide audience.

Read more and download the Science Book here.

Technical Overview & Requirements

The GMT Science Requirements for the telescope and associated instruments and facilities flow from the scientific priorities listed in the GMT Science Book. These requirements are used to optimize the telescope design and development process, and to define the goals and requirements for the GMT first generation instruments.

The Foundation Documents for the GMT are below:

  • Concept of Operations

    This document expresses the stakeholders’ and owners’ intention for the Observatory. Through high-level operational objectives and constraints, it describes what the observatory is expected to do.

  • Science Requirements

    This document quantifies the broad observational requirements needed to address the scientific goals of the Partnership, which are described in the GMT Science Book and the science cases for the first-generation instruments.

  • Observatory Requirements

    This document is the response of the GMT Project to the Science Requirements document. It contains the top-level engineering requirements for the Observatory that is to be built.

  • Observatory Architecture

    This document captures the top-level system design, consistent with the Observatory Requirements. It defines the subsystems and their interactions as they deliver the various System Configurations that enable the Observatory to implement the Observatory Performance Modes defined in the Operations Requirements document.

  • Observatory Operations Concepts

    This document describes how the Observatory design described in the Observatory Architecture document will be operated by the Stakeholders during operations to meet Concept of Operations objectives and Science Requirement/Observatory Requirements specifications.


The telescope structure is the large, yet precise, part of the telescope that holds the mirrors and directs them to targets on the sky. The telescope structure (known as the mount) is an alt-azimuth design and it will stand on a pier that is 22 meters in diameter.

Weighing 1,800 tons, the mount interfaces with most of the subsystems of the observatory. Several of the GMT’s scientific instruments will be housed inside the mount. The mount has been designed to be very stiff to minimize focal plane image motion under a wide range of observing conditions.

In 2019, GMTO announced that MT Mechatronics and Ingersoll Machine Tools will undertake the final design and build of the telescope structure.


Primary Mirror

MirrorLabVideoStill Inside Mirror Lab

The principal measure of the power of an optical/IR telescope is the diameter of its primary aperture.

Larger apertures translate into both greater light collecting area and, potentially, higher angular resolution. GMT’s primary mirror comprises seven segments that work together as a single mirror with the resolving power of a telescope 24.5 meters (over 80 feet) in diameter. Each of GMT’s seven mirror segments is 8.4 meters in diameter. The limitation of the size of a single mirror segment is related to the technology available to manufacture and transport such a mirror. GMT’s 8.4-meter mirror segments are being developed at the University of Arizona’s Richard F. Caris Mirror Laboratory (RFCML).

The mirrors are made from Ohara E6 low expansion glass molded into a light-weight honeycomb structure. The mirror segments are ground and polished to a precise optical prescription. The final polished surface departs from the desired shape by no more than 1/20 of a wavelength of green light or approximately 25 nanometers. After polishing, the surface is coated with a thin layer of aluminum to achieve maximum reflectivity.

One of the mirror segments is mounted at the center axis of the telescope. The other six mirrors are mounted surrounding the center mirror segment. Each mirror segment is mounted into its own “cell,” a complex active support system that keeps the mirror in the proper position relative to the other segments at all times and maintains its precise shape under varying temperatures within the enclosure.

The most challenging aspect of making the GMT’s mirror segments arises from the asymmetric shape of the six outer segments. These mirrors have a steeply curved shape similar to that of a potato chip. The peak departure from a symmetric mirror is 14 mm at the edge, amounting to 28,000 waves of green light. A new suite of test instruments and procedures were developed by the team at the Caris Mirror Lab to test these mirrors.


fab_01 GMT1 mold is prepared by installation of hexagonal silica fiber cores.
[Photo: RFCML]
fab_02 GMT1 mold is loaded with approximately 20 tons of glass in preparation for casting.
[Photo: RFCML]
fab_03 Glass is melted while entire furnace assembly is rotated.
[Photo: Pat McCarthy]
fab_04 The bottom side mirror blank is cleaned after successful casting.
[Photo: RFCML]
fab_05 “Load spreaders” are added on bottom side to distribute weight.
[Photo: RFCML]
fab_06 The top surface of the mirror is slowly polished into optical precision.
[Photo: RFCML]
fab_07 A computer controlled “stressed lap” tool is used to polish GMT’s aspheric surfaces. [Photo: RFCML]

The GMT mirrors are made at The Richard F. Caris Mirror Lab in Tucson, AZ.

Mirror fabrication occurs in 3 stages:

  1. Cast the mirror blank by melting glass in a rotating mold;
  2. Perform rough grinding of the front and back surfaces;
  3. Polish the front surface to optical tolerances.

In addition to these steps, the polished glass must be installed into its mirror cell. After it is transported to the mountain-top telescope site, the glass is given its reflective aluminum coating to make it a mirror. Finally, the mirror cell assembly is mounted on the telescope structure for alignment and testing.

Casting a mirror

The mirror “blank” is formed by melting glass in a mold formed by a high-temperature resistant refractory material.

The mold consists of a tub made of silicon carbide cement, lined with ceramic fiber. This tub will form the bottom half of a furnace to melt the glass. The mold is filled with about 1,700 alumina-silica fiber hexagonal boxes that form a honeycomb structure. The top of the boxes must follow the aspheric shape of the final mirror surface; no two are identical!

After the mold is prepared, the glass is placed inside the mold. Blocks of low expansion glass made by the Ohara Corporation of Japan are inspected and weighed. Approximately 20 tons of glass are loaded into the mold, one piece at a time. Finally, the furnace lid is placed on top of the mold.

The mirror is made using a unique “spin cast” process whereby the furnace is rotated as the glass melts. This gives the mirror surface a concave, or parabolic, shape. The mirror will still require additional shaping by grinding to achieve optical tolerances. However, this process saves several tons of glass and significantly shortens the annealing and grinding time because the glass is already in a parabolic shape. Heating coils in the walls and lid of the furnace raise the temperature to 1160°C (or 2120°F) as it spins at 6 rpm. The temperature is maintained for four hours to allow the glass to melt and fill the mold. The glass is then cooled rapidly to 900°C and then cooled more slowly for three months to avoid strains in the final mirror.

When the glass has cooled, the successfully cast mirror is lifted out of its mold. When tilted vertically, the mold’s attached “floor tiles” are visible on the rear surface of the glass. The tiles are removed and a high-pressure water spray is used to clean out the fiber boxes. This leaves behind a piece of glass with a honeycomb-like structure; the mirror is mostly empty space. This significantly reduces the mirror’s weight and enables its temperature to stabilize much more rapidly than a solid glass mirror.

Polishing a mirror

The mirror is inverted, and the rear surface and edges are lapped and polished. 165 “loadspreaders” are bonded to distribute the weight of the mirror and provide permanent attachment points to mount the mirror to its active support system in the telescope.

The mirror is turned face up, and the front surface is ground to its approximate final shape with a series of diamond grinding wheels. The mirror’s surface is then polished to precise specifications.

A variety of state-of-the-art polishing tools are used to shape the mirror’s highly aspheric surface. These are a “stressed-lap” polishing tool and conformal laps that use non-Newtonian fluids to achieve the desired shape.

As the mirror is polished, it is tested using multiple, redundant measurements to assure it is precisely figured. This “Optical Metrology” is the most challenging aspect of mirror fabrication.

More information: “How do you build a mirror for one of the world’s biggest telescopes”, The Conversation, January 15, 2016.

Optical Metrology

met_top An 82-foot high optical tower is
used to test each mirror to
extremely high precision.
met_top Technician inside optical tower
at Mirror Lab.

Due to the mirror arrangement in the GMT, each mirror has to be highly aspheric–that is, the mirror’s face is curved in a shape that differs from a sphere – the most symmetric geometric shape. One side of the mirror is higher than the other, differing by as much as 14 mm – an unprecedented asymmetry for a large mirror.

The difficulty of shaping each mirror segment is compounded by the fact that these large segments must have the same curvature to exacting standards in order to properly perform together. To achieve the stringent demand for accuracy, each mirror is periodically moved from the polishing machine to the test tower where its shape is carefully measured. The results of these measurements, in turn, guide the polishing program as it progresses.

The principal test to compensate for the unprecedented asymmetry of the mirror’s surface is conducted using a laser interferometer. Instead of measuring the mirror directly, a beam is first bounced off of two mirrors at oblique angles and then passed through a computer-generated hologram. With the aspheric departure removed, interferometric surface measurements may be taken.

In theory, this is complicated, and in practice, it is even more difficult. One of the bounce-off mirrors, called fold spheres, must be quite large: 3.8 meters in diameter. And, it must be positioned a considerable distance away from the primary mirror segment. The Test Tower protruding above the mirror lab had to be rebuilt to support this large fold sphere and accommodate the width of the interferometer beam.

Three additional methods are used to verify the mirror’s shape and direct the polishing activity:

  • The scanning pentaprism test is an independent measurement of low-order aberrations that guards against the possibility of a mistake in the implementation of the principal test.
  • The laser tracker measurement supports surface generation and loose-abrasive grinding processes by providing an independent measurement of radius of curvature and astigmatism.
  • The Software Configurable Optical Test System (SCOTS) provides an independent measurement of the mirror surface using relatively simple, and hence robust, technology. A computer monitor is used to project straight line grids onto the mirror. Distortions in the reflected light allow the team of optical scientists to reconstruct the shape of the mirror and determine how well it fits the desired shape.

Wavefront Control

The GMT will deliver excellent natural seeing images thanks to active optical control of the primary mirror segments and tip-tilt correction in the fast-steering secondary mirrors.

The primary mirror segments are housed inside a “cell” which protects the mirror and provides the interface to the telescope structure. Pneumatic actuators will push on the back of the primary mirrors to correct for the effects of gravity and temperature variations on the mirrors.

Another aspect of wavefront control is ensuring the 7 primary mirrors are in phase so they can act as one coherent surface. A dispersed fringe sensor measures the difference in phase of the light reflected off the mirror edges on either side of the gaps. Work on a prototype of this sensor is underway – read more in our August 2018 newsletter.

Adaptive Optics

The GMT’s Adaptive Optics (AO) system will be built into the secondary mirrors which will be deformable. The use of adaptive secondary mirrors allows us to implement adaptive optics without additional background or throughput losses. These Adaptive Secondary Mirrors (or ASMs) are comprised of an extremely thin sheet of glass that is bonded to more than 7000 independently controlled voice coil actuators. These actuators will be able to push and pull on the mirrors over 1000 times a second to correct for wavefront distortions introduced by turbulence in the Earth’s atmosphere.

The Gregorian optical prescription allows us to correct for atmospheric turbulence over a wide field of view and to utilize a variety of correction modes including diffraction-limited and ground layer correction. The segmentation of the adaptive secondary mirrors allows us to transfer large scale motions from the primary mirror segments to the much more agile secondary mirrors. Rather than moving 17 tons of glass to compensate for phase errors in each segment we only need to move roughly 2kg of glass at the secondary mirrors. This greatly simplifies the overall optical control challenge and provides more pathways towards realizing the exquisite imaging potential of the telescope.

The GMT will have several types of adaptive optics, including:

  • Ground layer AO: the GMT’s Gregorian optics allow efficient correction of image degradation by turbulent air flowing over the terrain near the telescope. This “ground layer” atmospheric turbulence to be corrected over a wide field of view ( ≥ 10 arcmin). Image size in the near infrared will be reduced by 50-200%.
  • Natural Guide Star AO: This mode uses a single bright star to determine image distortions and derive corrections needed to deliver diffraction-limited at wavelengths from the red end of the visible spectrum to the mid-infrared over a field of view of 20-30 arcseconds in diameter.
  • Laser Tomography AO: This mode uses six laser guide stars and a single, faint, natural guide star to extend diffraction-limited performance to nearly the full sky at IR wavelengths over a field of view similar to Natural Guide Star AO, but with reduced contrast.



The GMT’s enclosure is a 22 story building that tries to disappear at night. During the day, the enclosure protects the telescope from dust, the sun, and the weather. Once the sun sets, the enclosure’s vents open to allow airflow to equalize the inside temperature with the outside.

The enclosure will be approximately 56 meters in diameter and 65 meters tall. It will be able to complete an entire revolution in just over 3 minutes at maximum speed.



The GMT will be built on a peak in the Andes Mountains near several existing telescope facilities at Carnegie Institution for Science’s Las Campanas, Chile. The Las Campanas Observatory (LCO) location was selected for its high altitude, dry climate, dark skies, and unsurpassed seeing quality, as well as its access to the southern sky. Las Campanas Peak (“Cerro Las Campanas”), one of many peaks within LCO, has an altitude of over 2,550 meters (approximately 8,500 feet).

The GMT project is in the fortunate position of having clear access to an already developed site: road access, water, electrical power, and communications are already in place. The site has a long history of excellent performance. Light pollution is negligible and likely to remain so for decades to come. The weather pattern has been stable for more than 30 years. There are also many interesting objects that are primarily visible from the southern hemisphere such as the large and small Magellanic clouds, which are our closest neighboring galaxies, and our own galactic center.

Science Instruments

The GMTO Board of Directors has adopted an instrument development plan that follows the recommendations of the GMT Instrument Development Advisory Panel. Instrument development will be staged to match the technical development of the telescope and its adaptive optics system. Currently we are moving forward with four instruments and one facility fiber positioning system, summarized below. The summaries link to more information and related publications.


Visible Echelle Spectrograph – G-CLEF
A high resolution, highly stable, fiber-fed visible light Echelle spectrograph well suited to precision radial velocity observations, investigations in stellar astrophysics and studies of the intergalactic medium. G-CLEF will operate from 350nm to 950nm with spectral resolutions ranging from 25,000 to 120,000.


Visible Multi-Object Spectrograph – GMACS
A high throughput, general purpose multi-object spectrograph optimized for observations of very faint objects. GMACS will be used for studies of galaxy evolution, the evolution of the IGM and circumstellar matter, and studies of resolved stellar populations, among other applications.


Near-IR IFU and Adaptive Optics Imager – GMTIFS
A general purpose, AO-fed near-infrared (0.9 to 2.5 microns) integral field spectrograph and adaptive optics imager. The IFU mode will support multiple spaxel scales with spectral resolutions of 5,000 or 10,000.


IR Echelle Spectrograph – GMTNIRS
An AO-fed high-resolution, 1-5 micron narrow-field spectrograph aimed at studies of pre-main sequence objects, extrasolar planets, debris disks, and other mid-IR targets. The baseline configuration provides spectral resolutions ranging from 50,000 to 100,000.


Facility Fiber Optics Positioner – MANIFEST
A facility fiber positioning system that covers GMT’s full corrected 20 arcmin field of view. MANIFEST can feed G-CLEF and GMACS simultaneously with fiber bundles that may be configured to increase spectroscopic multiplexing, spectral resolution, and other scientific capabilities.


Commissioning Camera – ComCam
The commissioning camera will be used to validate the Ground Layer Adaptive Optics (GLAO) performance of the GMT facility Adaptive Optics System. It is also needed for the initial alignment of the telescope and for verifying the natural seeing optical performance in the Direct Gregorian Narrow Field (DGNF) mode.