Electro Optics: October/November 2010
Just to clarify
Greg Blackman on adaptive optics, devices originally developed in astronomy to improve the clarity of telescopes, but which are now finding uses in microscopy and optical tweezing.
The technology for deformable mirrors, such as this example from Edmund Optics, originated in the field of astronomy, but is now finding uses in microscopy and other biomedical imaging.
Whether or not we are we alone in the universe is still a key question for many astronomers. Telescopes have developed to such an extent as to map distant galaxies with great precision and clarity. The first full-sky image from the European Space Agency’s Planck telescope was revealed earlier in the summer, a map of primordial radiation emitted at the dawn of the universe. Other telescopes, including the James Web Space Telescope (JWST) and the European Extremely Large Telescope (E-ELT), are currently under construction to further space investigations. The E-ELT upon completion will have a primary mirror 42 metres in diameter, making it the largest ground-based optical telescope in existence.
Star gazing from Earth is made more difficult by the Earth’s atmosphere, which creates distortions in the light reaching us from space. It’s our atmosphere that causes the stars to twinkle, for example. Telescopes, both ground- and space-based, will use adaptive optics (AO), typically deformable mirrors, to correct for distortions in the light reaching the instrument and obtain a clearer image. ‘By removing those [atmospheric] distortions, you’re effectively taking the twinkle out of the star,’ comments Michael Feinberg, director of product marketing at Boston Micromachines Corporation (BMC), which manufactures deformable mirrors for adaptive optics systems.
A deformable mirror (DM) is one in which the mirror is deformed to achieve a surface profile that mimics the incoming wavefront, i.e. where there is a peak in the wave there is a trough in the surface of the mirror. The mirror will have a number of control points or actuators that pull and push on the surface of the mirror to distort it. In this way, the surface profile of the mirror can be shaped to alter the phase of light and, by integrating the mirror into an adaptive optics system, distortions in the image can be removed. With Boston Micromachines’ technology, the actuators pull on the mirror’s surface to alter it by up to 5.5μm.
The degree to which the surface of the mirror can be deformed is known as stroke, which determines how large an aberration can be corrected. Feinberg explains that smaller telescopes will typically use a single DM with lower resolution (determined by the number of actuators), but higher stroke. Larger telescopes, however, with apertures on the scale of 10m or more, would generally use two mirrors in a so-called woofer-tweeter design – similar in principle to audio speakers with a tweeter for high frequency and a woofer for low frequency, but applied to light instead of sound. A woofer-tweeter design would have one DM with high stroke, but low resolution and a second mirror with high resolution and low stroke.
The Gemini Planet Imager (GPI), an adaptive optics instrument for a ground-based telescope, uses deformable mirrors from Boston Micromachines. The imager will incorporate a BMC mirror in tandem with a second mirror in a woofer-tweeter design (the BMC mirror is the tweeter). Boston Micromachines’ high-resolution mirror is 25mm in diameter with 4,092 actuators.
An adaptive optics setup can be as simple as a deformable mirror acting on a light source with the end detector receiving the signal. This is an open-loop system. A closed-loop system requires a beam splitter that splits the incoming light into two paths. One of these beams would be directed to a wavefront sensor that profiles the incoming signal and communicates back to the deformable mirror. The other beam would be reflected off the mirror, which has been shaped according to the information provided by the wavefront sensor. The reflected light is then directed to an end detector or camera.
Boston Micromachines’ Multi-DM, suitable for a number of applications including microscopy, retinal imaging and laser beam shaping
‘Adaptive optics setups can be as simple or as complex as required for the application,’ comments David Henz, product line manager at Edmund Optics. ‘In a closed-loop system, the wavefront sensor is actively monitoring the light and the mirror is actively correcting for distortions.’
Currently, Edmund Optics supplies deformable mirrors from Italian company Adaptica, although it is in the process of expanding its adaptive optics product line to include another variety of mirror, as well as providing the wavefront sensor.
Tommaso Occhipinti, general manager at Adaptica – which, along with its distribution agreement with Edmund Optics manufactures other AO components and systems – notes that, traditionally, one of the issues with adaptive optics was the cost of the system. ‘You have to remember adaptive optics was born 20 years ago in the field of astronomy. It has quite a long history,’ he says. In addition to cost, Occhipinti says it was often quite complicated to apply AO technology to the final application.
‘A lot of adaptive optics systems can be simplified, including the production of the DM and the control electronics,’ Occhipinti says. ‘Simpler, more cost-effective systems allow AO to be used not only in the scientific field, but also in the industrial field. Adaptica was created in 2009 with the mission to bring adaptive optics from scientific applications to a more practical use in industry.’
Adaptica’s DMs can be produced using production methods originating from electronic chip production, which reduce production and system costs. ‘In our standard membrane DM there’s nothing innovative in the physical principle, but it is a big step forward in the simplification and engineering of the overall system,’ states Occhipinti.
Adaptica is developing other technologies specifically for the simplification of the deformation of the mirror. Currently, these mirrors are being tested in collaboration with the European Southern Observatory (ESO), but the technology has not yet been released.
Applying AO to microscopy
‘Imaging in astronomy was the main driver behind the creation of deformable mirrors,’ comments Henz, at Edmund Optics. ‘Now, however, adaptive optics are used to correct distortions in microscopy imaging, and deformable mirrors have been used to get down to very close to the diffraction limit of light. Essentially, whether it is astronomical or microscopic, these applications require similar properties from adaptive optics, but just on a different scale.’
Adaptive optics has been applied to ophthalmology in retinal imaging and optical coherence tomography (OCT). There are also some applications in environmental control, video surveillance and optical communications, in particular free-space optics.
Deformable mirrors are only one variety of adaptive optics; LCD spatial light modulators (SLMs) can also change the phase of light through altering the polarisation of the crystals. SLMs are used in biomedical applications, for instance, to correct images of the retina at the back of the eye. This is an important tool for studying eye diseases such as glaucoma and age-related macular degeneration (AMD). Ray Livingstone, sales engineer at Hamamatsu, comments: ‘To image the back of the eye requires a high spatial resolution and large phase depth, which SLMs provide due to the high diffractive efficiency and high phase modulation control. Using adaptive optics allows the blood vessels of the retina to be viewed, which can’t be seen with a standard imager.’
Hamamatsu’s X10468 SLM is a liquid crystal on silicon (LCOS) SLM – a programmable, holographic, active diffractive element, used to control the phase of the light. The response time of Hamamatsu’s SLM is now fast enough to enable the relevant components of the ocular wavefront dynamics to be studied and for an improvement to the aberration display rate. ‘With older technology, researchers couldn’t get conclusive results due to the slower response time,’ explains Livingstone, ‘but the new SLM is fast enough to do this. The response time has increased from around 4Hz to 60Hz, or even 120Hz in some cases.’
SLMs are also used in optical tweezing applications, in which the properties of light are used to trap and move around small particles, such as biological cells under a microscope. Professor Miles Padgett at the department of physics and astronomy, University of Glasgow is using SLMs, including those from Hamamatsu, in his work at the University of Glasgow’s Optics Group for a range of applications, optical tweezing being one of these.
‘Using the SLM allows one laser beam to be shaped into many spots in various different patterns,’ explains Padgett. ‘So, rather than trapping one particle with one laser, many particles are trapped. Then, because you can change the design of the distortion at 60 times per second, the particles can be moved around. The SLM gives a fully adaptable pattern that allows you to move small objects like cells around – bringing them together, pulling them apart – without physically touching them.’
An SLM can be used to correct the wavefront in order to see blood vessels in the retina (right), compared to a blurred image without adaptive optics (left). Image courtesy of Hamamatsu
The Optics Group also uses SLMs in reverse, with an optical detector in place of the laser to detect images in the order of single photons. ‘In the same way you can use an SLM to produce specific patterns of light, you can use an SLM to detect light patterns. This is difficult computationally, but possible optically,’ says Padgett.
According to Padgett, deformable mirrors aren’t suitable for these areas of application. ‘The trouble with a deformable mirror is that it’s not very deformable,’ he says. ‘A deformable mirror might have 32 actuators to change the surface profile, whereas a liquid crystal device, because it’s based on silicon, has a million pixels’ – Hamamatsu’s SLM has 800 x 600 pixels with a 20μm pixel pitch, and each pixel is individually modulated. ‘Having said that, mirrors are better suited to certain applications as they’re more reflective,’ Padgett continues. A DM might reflect 99 per cent, whereas SLMs typically only reflect 70 per cent of light.
‘SLMs are slower than deformable mirrors,’ says Padgett, pointing out that a typical DM will have relatively few actuators, but these can be moved at kilohertz speeds, whereas an SLM will only be able to operate at around 60Hz. However, SLMs provide far more detail. ‘You can’t do holographic tweezers with a deformable mirror; there just aren’t enough pistons to create the distortion you need,’ he says. ‘By the same token, a deformable mirror is probably exactly what you need if you’re building a telescope to look at the stars.’
Many biologists studying how muscles work would use optical tweezers, because it’s the main way of measuring the piconewton forces individual muscle fibres exert. On the measuring side, the Optics Group’s interest in optical tweezing is in secure communications. ‘Normally, a photon could be thought of as carrying a one or a zero. Because we can both shape the light and measure it in any shape, this provides potential for a lot more information to be carried per photon,’ says Padgett. A photon could be formed into any letter of the alphabet, which is 13 times more information per photon than a binary signal.
AO for industry
Adaptive optics were originally developed as a tool for astronomy and are still used heavily in the scientific community today. However, Occhipinti of Adaptica feels that simplifying AO and reducing the cost would open up its use in a number of industrial applications. AO can be used in laser optimisation for high-power lasers – it can be used to compensate for aberrations in laser light in cutting and engraving machines, for instance.
‘If you want to approach the industrial market, you need simpler systems,’ Occhipinti says. ‘In general, you don’t need a high order of aberration compensation. In addition, there are many applications in which the AO system would work without a wavefront sensor (known as sensorless adaptive optics). If you want to create an adaptive optics microscope with a low price, it is preferable not to use a wavefront sensor.’
Adaptica is involved in two areas: first are those for sophisticated, complex systems for scientific applications, such as those under development at European Southern Observatory (ESO). Adaptica is involved in a project with ESO to develop a large deformable mirror with a high spatial resolution.
The other trend is moving to simpler solutions by removing the wavefront sensor if it’s not required and using simpler electronics. Occhipinti feels that engineering an adaptive optics ‘black box’, so that users can turn on the system and not know in detail about its workings, would make these devices more accessible to industrial areas outside scientific laboratories.