There are number of potentially promising products to manufacture in space: protein crystal, semiconductor devices, contact lenses and other high quality optical instruments. Pharmaceutical companies are also doing research in space stations to search for the possibility of developing new drugs in space. In fact, any products that can fully utilize two advantages of space environment: clean and weightless, are under consideration.

The semiconductor devices are exceptionally susceptible to contamination from particles, such as Sodium and organic particles. The amount of salt on a finger tip is enough to destroy all the chips ever manufactured on earth. The space is an ultra-vacuum, nearly empty of all molecules. It is thus perfectly suited for manufacturing semiconductor devices which requires extraordinarily clean environment. In space, exceptionally pure and atomically ordered thin films of semiconductor compounds such as silicon and gallium arsenide can be developed, with qualities that are otherwise impossible on earth. This potential leads to the possibility of producing the next generation f semiconductor and superconductor thin film materials, and the devices they will make possible.

The market value of computer chips and the significant cost required to maintain a clean room on earth makes it also economically feasible to consider space as an alternative environment for manufacturing semiconductors. The clean room cost can be completed eliminated as space itself is a natural clean room. The typical computer chips that powers a home PC, valued at approximately $2.2 million/Kg, make the launch cost of material insignificant. The yield rate and quality of the chips will also be improved because of the reasons described above, further lowering the cost and increasing the profit.

However, there are obstacles in developing a space semiconductor factory - the initial startup cost and the fact that it has never been done before. Needless to say, the scale of the investment is too large for any single company to handle. And the risk involved in designing such a factory, making it competitive and profitable, is too great for a company to consider. Very few commercial firms are interested in spending their own money to put this idea into work. This wait and follow attitude makes the progress slower than what it should be. Thus, the success of this project requires the corporations between companies and governments.

In our project, we attempted to design the actual manufacturing facility in space and estimated the cost and feasibility of such investment. Following paragraphs are the summarization of our findings. The costs associated with the plants on earth and in space are quite different, they are presented in following tables.

The designs of the plant on earth and in space will be very different. In space, we need to consider the balancing of the system, the power supply, heat dissipation and various other unique situations that are not present on earth. (For details, refer to Design of space station) On earth, the design clean room constitutes large portion of the cost in the design of the plant, which is not present in space. However, due to the maturity of two technologies, we can almost be certain the initial design cost of plant in space will be larger than that on earth.

Although the manufacturing equipment used in both places will be different due to distinct natures of two environments, they are, in most cases, similar or identical to each other. Thus, we can safely assume that the cost of equipment and developing equipment are approximately same.

One factor in startup cost in space that is not present on earth is the launching cost of the material. The total weight of the equipment is estimated to be 34,187 kg. Adding the weight of the structure of the manufacturing facility, the launching cost could be quite significant. However, this is a one time investment, and assuming in the future a low cost reusable space vehicle will be developed, lowering the launch cost to $1,000, we can safely conclude that this cost can be brought under control.

Building costs mainly constitutes human labor costs and material costs. The space manufacturing facility will be built on the ground component by component, then shipped into space to be assembled by astronauts. The extra costs associated with astronauts and assembly equipment in space will make this process more costly in space.

As we can see, the initial startup cost of building a manufacturing facility in space will be significantly higher. However, it’s important to keep in mind that most of these investments are one time investments. And the higher costs associated with them are due to the innovative nature of this technology. Once it has been experimented, the process will be streamlined and cost will be significantly lower for future developments.

We can again assume that the equipment maintenance costs are approximately same in both places. (They will be using the similar equipment) What separates space and earth apart in on-going cost are the launching cost of material in space, and the clean room maintenance cost on earth. Semiconductor manufacturers spend a significant portion of their income to maintain the clean room each year. The cost runs into hundreds of millions range. On the contrary, the energy required to launch the material into low earth orbit is same as flying it across pacific. What keeps the launching costs high today is the non-reusable rockets, or semi-reusable shuttles (which are even more expensive to maintain). If we assume that in the future, a practical reusable space vehicle will be developed, then the launching cost will be negligible comparing to the maintenance cost of clean room on earth.

The other issue that we have to consider to make idea of space semiconductor factory feasible is the ability to mass produce chips. The plants on earth are able to produce chips at a much higher rate than what’s possible on the space station because they employ mass amount of equipment. Similarly, we can simply build more stations or add manufacturing components to the original facility to increase the productivity.(As the technology matures, the cost of building or adding additional components should be considerably lower that the first one)

According above analysis, we conclude that building a semiconductor manufacturing plant in space is economically feasible in the future. As we push the limit of semiconductor technologies, only space will provide the right environment for manufacturing smaller and faster semiconductor devices. Thus we urge that every semiconductor companies out there to jump on the boat. Stop waiting and start acting if you want to survive into the 21st century!!!.

We mentioned in the introduction the high resale value of computer chips as a motivating factor to their manufacture in space. We have not yet mentioned the other key motivator to developing this manufacturing station off earth. The significant cost required to maintain a clean room on earth makes is not matched in space, even when high launch costs are rolled into the maintenance bill. The clean room cost can be completely eliminated as space itself is almost a natural clean room. We would be setting up shop in low earth orbit, where there is still some atmosphere. Processes for which this is problematic can be executed behind a wake shield. The computer chips, valued at approximately $2.2 million/Kg, make the launch cost of material insignificant. In the environment we find in space, the yield rate and quality of the chips will also be improved once our manufacturing process matures, further lowering the cost and increasing the profit.

However, there are obstacles in developing a space semiconductor factory - the initial startup cost is considerable, and the fact that it has never been done before is food for thought. The risk involved in designing such a factory, making it competitive and profitable in the market it will enter, is high. In our project, we attempted to design the actual manufacturing facility in space and estimated the cost and feasibility of such investment. In the following paragraphs we give a summary of our findings on the costs associated with comparable development on earth and in space.

Start-up Costs

The designs of the plant on earth and in space will be very different. In space, we need to consider balancing the mass of the system, our rate of power consumption, heat sinks and various other unique situations that do not pose as much difficulty on earth. (For details, refer to TDesign of the space station.) On earth, the design clean room constitutes large portion of the cost in the design of the plant, which is not present in space. However, due to the maturity of two technologies, we can almost be certain the the design cost of plant in space will be larger than on earth.

One factor in startup cost in space that is not present on earth is the frequent launch cost of materials, replacement parts, and humans. We consider this part of our maintenance or operating costs. One launch currently costs on the order of $10000, and a launch can be made at most every six months. For development and production, we are assuming that launch costs are $1000 (which is what we think they will be within five years), and that we can make a weekly visit to the station.

Assembly costs mainly constitutes human labor costs. Although astronauts salary and living costs will be significantly higher than earth-bound labor, we also need to take account that less people will be required to assemble the space factory, and that they will be employed for a longer period than their earth counterparts.

On Going Costs

We can again assume that the equipment maintenance costs are approximately the same in both places. (since they will be mainly using the same equipment), and we take this to be 10% of initial equipment expenditure. Clean room maintenance costs runs to over hundreds of millions of dollars per year. The launch costs are seen to be negligible in comparison to the maintenance cost of the clean room on earth.

From above analysis, we hoped that building a semiconductor manufacturing plant in space would be economically feasible, based upon the reasonable assumption that in the future a reusable space vehicle will be invented to significantly lower the cost, and based on the assumption that we can get equal output in computer chips per unit time. While we maintain that both of these assumptions are reasonable, we've left it to another team to figure out how to make a reusable space vehicle (we'll need a small fleet of them), and in the text that follows we consider our production line:

The reason we have been hesitant to name facts and figures above is that it was about as hard to get company information on these costs as it is to get a dead horse to speak in tongues. But we do know the current assembly lines produce about 40 cassettes (of 25 wafers per cassette) per week. Given our initial estimate of being able to produce two cassettes per week, we had some catching up to do. We assume we have to match earthbound factories computer chip output numbers in order to be able to pay the high start up costs of the space-bound industrial park in a reasonable amount of time.

Constraints

At the end of the week, we need our matured process to produce the same number of chips as our competitors. For our calculations, we assume we are using 30 cm diameter silicon wafers manufactured on earth and transported to the station, and we assume that we are competing against facilities who also use 30 cm diameter wafers. We'll be submitting the details of our scheduling process to Dr. John Jones in hard copy form, since a mere six cassette run takes 16 to 20 pages to lay out. Bottle necks are created by long processing hours on few machines; the more time a cassette has to spend in a given machine on a run, the less cassettes we can process in a run. Developer and Photoresist are two of the four processes which consume 11 hours of each cassette's time per run. Thus in a week, we can push at most (7*24/11), or 15, cassettes through the manufacturing facility. But this is a theoretical limit, so of course we are not going to achieve it, and worse news is, it's a far cry from 40.

So it isn't a scheduling problem that threatens to put the space manufacturing site out of business. Unfortunately, earth manufacturing sites are, to us, black boxes; we don't know how many of each machine they have. Factors that might be holding our production levels down are:

• numbers of machines we have
• Power Consumption Rates
• Wafer Size

Using the advantages of being in space, we can make four large solar panels, instead of two, to double the amount of power available, and this expansion continues. We do not perceive our power consumption rate to be a limiting factor, after the capitol required to harness the sun's energy is located. So power is not what is really stopping us from competing. If we could cut silicon that is grown in space, then we could make larger wafers than our competitors use. This would inherently slow down some of our processing time, but not most of it, so we would be producing chips on average more quickly than our competitors. We don't want to double machinery if possible, because this plan changes the proposed station the most. Twice the mass would have to be balanced, about 1.5 times the initial mass would have to be launched, and the station's development would begin to cost significantly more than development costs of similar factories on earth. But even so, it would still pay off quickly enough if we could get a high enough yield on a good enough chip.

We think that any viable industrial park in space is going to use two to three of these solutions. Learning how to cut the space-grown large wafers is the most straightforward, since that technology makes our chip production capabilities theoretically limitless (when combined with the availability of free power generation), but there is a hidden cost. Designing and developing new processing machinery to handle the larger wafers could itself cost billions, and this value gets rolled into the development costs of the station. The most affordable competitive solution, currently, is to:

1. Use four solar panels (increased from two)
2. Nearly double all equipment (triple some)

So roughly, double the mass and equipment costs, and probably increase total development costs by up to 50%. With all other factors held to our initial assumptions (returning to a 30 cm diameter wafer launched from earth), we can in this scenario produce about 35 cassettes per week. This is nearing our mark. The schedule will be submitted to Dr. Jones and can be obtained from the group upon request.