Towards in-space crystal growth
Could the answer to how to produce the next-generation of reliable, high-quality semiconductor crystals lie in space?
BY ALASTAIR MCGIBBON AND NEIL MONTEIRO FROM SPACE FORGE
Compound semiconductors have successfully established themselves in market applications where their superior properties over silicon give them a clear advantage. However, for classes of material such as wide and ultra-wide bandgap semiconductors, bulk defect densities can limit the performance and reliability of the systems, which in turn can slow or limit the rate of adoption into markets.
Examples of this are seen in the wide- and ultra-wide bandgap semiconductors GaN, SiC, AlN and diamond. With these materials, high levels of bulk defectivity threaten to seriously curtail power performance, as well as the quality of the crystal, which is critical to high-coherence quantum applications. SiC and GaN are clearly making an impact in power RF and power electronics markets, but there remains the persistent challenge of crystal quality, reliability and yield, particularly for very high power-density and voltage applications.
Progress continues, with crystal quality slowly improving, particularly in epilayer and two-dimensional material development. But it’s still not possible to reliably grow inorganic crystal ingots with bulk defectivities on a par with silicon. This puts up barriers to market adoption.
Space Forge’s approach
At Space Forge, we are taking an entirely new approach in the search for reliable, high-quality semiconductor crystals, by posing this question: What if there could be a way to produce CVD seed wafers with a quality that’s at least two-to-three orders of magnitude better than what’s currently available?
A promising answer to this question lies in a broad, but disparate body of evidence that shows that the conditions found in low earth orbits (LEOs) and beyond have a reliable, positive effect on the quality of crystal growth. Broadly, these conditions are: microgravity, very low temperatures and a high-quality vacuum. Of these three, microgravity is a driving factor – it has been shown to greatly reduce convection, and hence nucleation points, in gas-phase crystallisation in orbit. Also aiding inorganic crystal growth is the high-quality vacuum found in a LEO, ensuring very low levels of nitrogen. At typical LEO altitudes, the density of N2 particles is less than 1 x 1012 m-3, and even lower in the wake of satellites, as seen in the epitaxial demonstrations of the Wake Shield Facility flights.
Our mission is to establish the quality of material that can be produced in an LEO, when optimising the design and operation of the growth tool, so that this makes the most out of the benefits of these conditions in space.
The concept of conducting CVD growth in space gives rise to a series of complex engineering, manufacturing and cost challenges – so it’s important to focus on a relatively simple strategy for proof-of-concept and early scale-up. Based on these considerations, we are prioritising the
in-space element of semiconductor growth to be as conservative as possible. Our motto is this: do only in space what only space can do. Consequently, our focus is on the quality of the crystal that can be grown in an LEO, and not the scale. Once that’s accomplished, efforts will only be directed at conventional terrestrial semiconductor scale-up, exploiting returned high-quality space seeds.
Pursuing this approach, we have already sent a pilot semiconductor growth tool into space, which is currently orbiting the Earth in a free-flying satellite. Our ForgeStar-1, launched last June, is now carrying out initial feasibility tests of the pilot plasma-enhanced CVD system in the payload. Although our tool is small scale, designed to grow only a few millimetres of material, it is a first significant step in space, proving that we are capable of using our semiconductor growth tool in orbit. Once we have completed this stage, we will embark on a series of missions that involve scaling up our tool and beginning to fully understand the physics of inorganic crystal nucleation and growth in orbit.
The successful integration of ForgeStar-1 onto the SpaceX Transporter 14 mission.
The next step
The challenges of in-space crystal growth and return are considerable – not least optimisation of the size, weight and power within the payload, so that it is practical to reliably and repeatably grow inorganic crystals in space. Energy requirements for most types of in-space manufacturing are very much the ‘elephant in the room’, providing a barrier for embarking upon such ventures. Tackling this issue head-on, we are developing one of the highest energy-density satellites in the market, to enable the growth of in-orbit seeds. Our next iteration of spacecraft will feature a magnitude of power-generation that’s normally reserved for telecommunication satellites, combined with the capability of returning from orbital velocities, something rarely seen outside crewed vehicles.
While working towards these goals, we will optimise the size, weight and power of our plasma-enhanced CVD tool. Our improvements in this regard are already leading to a significant innovation in energy efficiency that could even have an impact on purely terrestrial systems.
We are developing a tool capability that’s flexible, and does not have to be on our satellite to work – it can be hosted in future systems, and be adapted to meet material or experimental needs of other customers.
Design for space
Within the space industry, the International Space Station (ISS) provides the principal platform for developing in-space manufacturing. Unfortunately, as well as being a highly problematic venue for growing compound semiconductors, it is decommissioning. This means that industry is moving to a private model, involving Commercial LEO Destinations that are expected to replace much of the functionality of the ISS. Due to this, we are working with a number of partners to host our systems on their platforms – a mutually beneficial arrangement as commercial platforms need viable commercial routes to be able to function, as long as they have the power density capabilities to be able to do so.
After growing our seed crystals in a LEO, we must return them to Earth. That’s not necessarily as challenging as it may first appear, thanks to the highly competitive nature of space transport. Recently, a number of companies have emerged, such as Varda Space Industries in the US and Atmos Space Cargo in Germany. Both are individually developing ways to return satellites to Earth – although for these companies, the focus is on a low power density for pharmaceutical crystal growth. At present, we are working on an integrated approach with our own satellites, deploying a heat shield that acts as an aerostable means of return through the Earth’s atmosphere, with the ability to predict and track low-impact landing zones.
Scaling up
The crux of our proposition – and that of any inorganic-crystal in-space manufacturing – is that the quality of our crystals is high enough that, when returned to Earth, provide seeds for significant high-quality scale up into a wafer stock. How good does this stock have to be? Well, good enough to displace terrestrially-seeded material in the market.
Based on these factors, we are involved in a number of needle-moving activities related to in-space metrology and growth control, investigating some of the key questions we can already anticipate. These questions are: What techniques and data are needed to understand in-situ and in-space measurement and optimisation? How do you achieve controllable and repeatable manufacturing with no real-time intervention? How do you model in-space microgravity growth, in order to reduce development cost and increase reliability? All these questions are being addressed.
On the ground we are tackling all the technical and business challenges that every compound semiconductor supplier has to meet. It is pointless producing high-quality seeds in an LEO unless this is followed with a terrestrial growth process that maintains initial crystal quality for as long as possible. We are currently incubating with the Centre for Integrative Semiconductor Materials (CISM) in Swansea to develop that capability. In addition, we are directing efforts at post-processing grown material, which provides the heart of engineered substrates, with an appropriate surface treatment that enables competitive products in the marketplace.
Given the challenges of growing wide- and ultra-wide bandgap materials on Earth, that’s our initial target. Similar to the challenges faced by all materials providers in the compound semiconductor industry – both past and future – success is not as simple as just providing a high-quality solution with disruptive theoretical properties. What’s needed to generate significant sales, is the adoption of our material in the supply chain, so that it is eventually part of a co-designed end-user system that justifies the cost and effort involved. To this end, we are already actively involved as a conventional materials supplier into power, RF, photonic and quantum projects.
Space Forge is using the clean room at Centre for Integrative Semiconductor Materials at Swansea University.
Into the future
If our approach to in-space manufacturing succeeds, what comes next? It’s likely that our attention will turn to the development of new markets and applications that will be enabled by high-quality wide- and ultra-wide bandgap materials.
For example, once we have established ourselves in our initial target market of materials to assist with thermal management and integration at the package level, we could gradually move towards producing materials for active devices at the fab-level. Alternatively, we can begin to explore how our materials can support quantum applications.
A view from ForgeStar-1’s onboard camera in a low-Earth orbit.
We may also diversify our activities to other growth processes and materials, while maintaining our focus on systems where the quality of the bulk crystal drives end performance. Efforts on this front are already in their infancy, as we are involved in early discussions on the feasibility of non-linear optical crystal growth in orbit. Similarly, although alternatives to CdTe are being deployed as a material for solar cells, data on good growth of this material in orbit already exists and may be possible to expand once in-space manufacturing capability is in place.
However the future path of development unfolds, what’s clear is that as the barriers to in-space manufacturing are overcome, the benefits of the orbital environment – and the technological developments required to exploit it – promise to help to solve challenges in terrestrial inorganic crystal growth.
