Week 1 HW: Final Project Ideas

HTBAA Homework #1 Automated Biomanufacturing of Iron-Rich Feedstock from Martian Regolith Student: Aswaajieet J. M. Course: HTBAA Homework: #1 Due: September 22, 2026

1. What I Would Like to Biomanufacture and Why 1.1 The idea I want to build an automated biological system that turns Martian regolith into an iron-rich feedstock for 3D printing.

I’m not aiming to print a finished tool or spare part in this project. The first goal is narrower: take Martian soil, use microbes to enrich the iron-bearing fraction, pull that fraction out, and end up with a material that’s consistent enough in composition and quality to feed a printer later on.

Longer term, I’d like to connect this feedstock step directly to additive manufacturing, so that regolith could go in one end and tools, replacement parts, and eventually habitat components could come out the other.

This isn’t coming out of nowhere. Castelein et al. already showed that Shewanella oneidensis can process Martian regolith simulant, after which the iron-rich material can be pulled out magnetically. They 3D printed the treated material, and the printed parts had much better compressive strength than those made from untreated material. One caveat worth stating up front: their work was a proof of concept on simulants, not on real Martian soil. [1]

1.2 Why this matters One of the biggest constraints on going to Mars is that everything has to be launched from Earth. With conventional conjunction-class mission profiles, a crew could be away for roughly 900–1100 days. NASA describes these round trips as about 180–300 days of travel each way, with a long surface stay in between. [2] Over that kind of timeline, being able to make or recover materials locally starts to look less like a nice-to-have and more like a necessity.

Carrying every spare part, tool, and bit of construction material you might need is a huge logistical burden. What makes local manufacturing more interesting to me is that it keeps paying off after the first mission. A base that can produce its own materials can slowly grow its capabilities instead of waiting on the next resupply.

So the system is really trying to do three things:

Cut down on how much material has to be launched from Earth. Use Martian resources as manufacturing feedstock. Start as something that supports a small exploration mission, and later scale toward a permanent base. This lines up with the framework in Averesch et al., who argue that in-space biomanufacturing becomes more valuable the longer and farther a mission goes. They highlight materials as a particularly important target, since structural metals and plastics make up a large share of mission mass, and they specifically mention microbial biomining of iron-bearing regolith as a possible route to construction materials and replacement parts. [3]

1.3 Why use biology at all? Conventional regolith processing tends to need heavy mechanical equipment, a lot of energy, and chemical inputs. Biology offers a different route: microbes can change the chemical state of minerals and make certain fractions easier to recover.

I want to be clear that the bacterium doesn’t “make” iron. The iron is already in the regolith, mostly as iron oxides. What the microbe does is change the oxidation state and mineral chemistry of that material so that an iron-rich, magnetically recoverable fraction can be separated out.

That’s why I’d describe this as an ISRU (in-situ resource utilization) bioprocess with a manufacturing endpoint, not biological synthesis of iron from scratch.

2. What Living System Would I Use? 2.1 Chassis I’d use Shewanella oneidensis MR-1.

It’s a natural fit because it can reduce metal oxides, which is exactly what matters for iron recovery. Castelein et al. showed that S. oneidensis interacts with several lunar and Martian regolith simulants and increases the amount of magnetically extractable material. With JSC-Mars1, the Martian simulant in their study, bacterial treatment followed by magnetic extraction gave more extracted material than untreated controls, and that material had a higher iron concentration. [1]

So I don’t need to invent a hypothetical organism. There’s already an experimentally backed starting point that can be studied and optimized for a space manufacturing setting.

2.2 The biological process At a high level:

Martian regolith → preparation → microbial treatment → iron reduction → magnetic separation → iron-rich feedstock

Martian regolith has a lot of iron, much of it oxidized. S. oneidensis can reduce Fe(III) to Fe(II), and that reduction encourages the formation of more magnetic iron phases, including magnetite, which can then be pulled out with a magnet. [1]

Step by step:

Regolith preparation – dig up the regolith and grind or sieve it to a suitable particle size. Biological treatment – load it into an anaerobic bioreactor with S. oneidensis. Reduction – the bacteria transfer electrons to iron-bearing minerals, pushing Fe(III) toward Fe(II). Mineral transformation – the reduced iron helps form magnetic iron-containing material. Magnetic separation – separate the magnetic fraction from the rest of the regolith. Feedstock conditioning – dry, characterize, and process the recovered fraction into something reproducible enough for additive manufacturing. Future integration – eventually hook this directly into an automated 3D printer. The end product, then, isn’t pure elemental iron. It’s an iron-enriched, magnetically recoverable feedstock whose composition and handling properties are controlled well enough to print with.

2.3 What the simulant experiments show The most direct evidence comes from Castelein et al., who used JSC-Mars1 (along with other simulants). In their reported comparison, bacterial treatment plus magnetic extraction yielded 5.8 times more extracted material than magnetic extraction of untreated JSC-Mars1. The extracted material had an 18.2% higher iron concentration than the starting material, and it was successfully used in lithography-based ceramic 3D printing. [1]

Their data also give a rough starting point for processing time. In anaerobic experiments with JSC-Mars1 at 10 g/L, magnetic extraction peaked at around 72 hours in one condition. In a defined minimal medium, about 3.1 mg of material was extracted at 72 hours and about 4.4 mg at 168 hours from a 10 g/L suspension. [1]

These numbers are useful for a first engineering estimate, but I don’t want to treat them as a mission-ready production rate.

2.4 A major limitation None of this has been done on actual Martian soil. For now, all experiments have to rely on simulants.

That’s a real problem, because simulants don’t perfectly match Martian mineralogy. Castelein et al. point out differences between JSC-Mars1 and measured Martian soil, and note that JSC-Mars1 may be more magnetic than the real thing, which could inflate magnetic extraction results. [1]

So in this project, I treat S. oneidensis + Martian regolith simulant as the development platform, without assuming that the numbers measured on Earth will carry over directly to Mars.

3. How Would I Use Automation? 3.1 Starting point I don’t want to try automating the whole mine-to-print pipeline right away. The first step is to automate the biological processing and the measurements around it.

A prototype would include:

an automated regolith feeder a controlled anaerobic bioreactor automated nutrient and water dosing sensors for pH, temperature, dissolved oxygen/redox state, and biomass automated sampling automated dissolved-iron measurement automated magnetic separation weighing and compositional analysis of the recovered material With that setup, I could systematically vary:

regolith concentration bacterial inoculum incubation time oxygen availability nutrient concentration temperature pH electron-donor availability 3.2 Why automate? The process has a fairly large optimization space. Extraction performance probably depends on regolith loading, cell density, incubation time, electron donor, medium composition, and how these interact.

Testing all of that by hand would severely limit how many conditions I could actually run. Automated workflows can run many conditions in parallel and produce standardized, comparable data.

Petzold and Mukhopadhyay make this general point for biomanufacturing: automated, high-throughput workflows let you explore optimization spaces that would be impractical to cover manually, and they produce structured datasets that work well for computational optimization. [4]

Here, automation would serve two roles:

Development – find the conditions that maximize recovery of iron-rich material. Operation – keep the process stable once those conditions are known. 3.3 Development workflow In the lab, I’d run this as a design–build–test–learn loop:

Design processing conditions ↓ Automated preparation ↓ Parallel microbial treatments ↓ Automated sampling ↓ Iron / composition measurements ↓ Automated analysis ↓ Select improved conditions ↓ Repeat Once a good process is established, the same setup could be converted into a continuous or semi-continuous production unit.

3.4 A future autonomous system on Mars Longer term, automation would extend beyond the bioreactor:

Martian regolith ↓ Excavation ↓ Particle-size preparation ↓ Automated biological processing ↓ Magnetic separation ↓ Feedstock characterization ↓ Feedstock conditioning ↓ 3D printing ↓ Tool / replacement part I’m deliberately treating this as future scope, not as the minimum needed for this project. A first exploration mission might only carry the extraction and feedstock-production module. Once that’s reliable, it could be paired with a printer.

4. How Much Would I Need to Make, and How Long Would It Take? 4.1 Setting a target For a first mission-scale demonstration, I’m defining one useful unit of product as:

1 kg of iron-rich feedstock, enough to go on to make a small tool or replacement part.

I’m not claiming any particular mission needs exactly 1 kg. It’s just a reference point that lets me turn the biological performance into a concrete mass and time requirement. From there the system could scale as needed:

1 kg → 10 kg → 100 kg → tonne scale

4.2 Yield from the published experiment Castelein et al. reported about 4.4 mg of magnetically extracted material after 168 hours of anaerobic treatment in defined minimal medium, from a 10 g/L JSC-Mars1 suspension. That works out to roughly:

0.44 mg/g or 0.044 % 0.44 mg/gor0.044% of the treated regolith mass.

That’s very low, and it’s one of the main reasons the published experiment can’t just be scaled up directly. I’m using it as a conservative baseline, not as a prediction of what a mature process would achieve. [1]

4.3 How much regolith for 1 kg? Using that baseline yield:

Y

0.00044 kg feedstock per kg regolith Y=0.00044 kg feedstock per kg regolith For 1 kg of feedstock:

M regolith

1 kg 0.00044 ≈ 2,273 kg M regolith ​

0.00044 1 kg ​ ≈2,273 kg So if the small-scale yield scaled linearly, I’d need to process about 2.3 tonnes of regolith to get 1 kg of magnetically extracted material.

Honestly, this is the most revealing number in the whole exercise. It shows that the current process can’t just be scaled up linearly and flown as-is.

4.4 What reactor volume does that imply? The experiment ran at about 10 g/L regolith. [1] At the same solids loading:

V=2,273,000g10 g/L ≈ 227,300 L V= 10 g/L 2,273,000 g ​ ≈227,300 L That’s roughly 227,000 L of slurry just to get 1 kg of feedstock, which is obviously not something you’d send on an early Mars mission.

So the engineering goal shouldn’t be “build a 227,000 L bioreactor.” The research should focus on improving:

regolith solids loading the fraction of iron that becomes magnetically recoverable extraction rate reactor productivity magnetic recovery efficiency how many reactors or cycles can run in parallel water and nutrient recycling In other words, this calculation doesn’t show the system is mission-ready. What it does is pinpoint the central scale-up problem.

4.5 Processing time The published experiments showed substantial changes within 72 hours. In one anaerobic JSC-Mars1 condition, magnetic extraction peaked around 72 hours, while the defined-medium experiment ran to 168 hours. [1]

For a first estimate, I’ll assume a 7-day processing cycle.

If a single reactor could eventually produce 1 kg of feedstock per cycle:

1 kg/cycle×1 cycle/week≈4.3 kg/month

4.3 kg/month×12≈52 kg/year To be clear, that’s an illustrative assumption, not demonstrated productivity. Under it:

Target Reactors Approx. time 1 kg 1 ~1 week 10 kg 1 ~10 weeks 100 kg 1 ~100 weeks 100 kg 10 in parallel ~10 weeks 100 kg 100 in parallel ~1 week This is where automation and parallelization really start to matter as the production target grows.

4.6 Why this isn’t the final design Everything above is a straight extrapolation from a lab experiment, and industrial bioprocesses almost never run at the same concentration, yield, and configuration as small lab tests.

A real scale-up would focus on improving volumetric productivity:

Volumetric productivity = kg recovered feedstock x L reactor × time ​

The plan would be to experimentally find the highest practical regolith loading and recovery rate before settling on reactor size. A possible path:

10 g/L lab experiment ↓ 100 g/L optimized process ↓ Higher-solids pilot reactor ↓ Parallelized production reactors ↓ Continuous / semi-continuous Mars processing unit The intermediate concentrations here are development targets I’m proposing, not established values.

5. Scaling from an Exploration Mission to a Mars Base The starting target is an exploration mission, not a permanent settlement. For that, the system only has to show that local production can supplement a limited stock of tools and parts brought from Earth.

Moving to a base would require a few big changes.

5.1 More throughput A base will need far more material than a small crew. That means:

larger reactors higher solids loading faster biological processing many reactors running in parallel continuous or semi-continuous operation 5.2 More consistent feedstock If you’re printing replacement parts, the material properties have to be reproducible. Every batch would need to be checked for:

iron concentration silicon and other major components particle size magnetic properties density mechanical properties after processing suitability for the chosen printing method 5.3 Recycling resources A base can’t keep importing growth media and consumables forever, so water, nutrients, and other inputs would need to be recovered and reused as much as possible.

Averesch et al. stress that Mars biomanufacturing needs to combine ISRU with loop closure rather than treating biological production as a standalone process. [3]

5.4 Connecting to additive manufacturing The end goal is to link feedstock processing directly to automated manufacturing:

Martian regolith→biological processing→iron-rich feedstock→3D printing→replacement part At that point the system becomes a genuine mine-to-manufacturing platform rather than just an extraction step. Averesch et al. specifically discuss coupling biomining and bioleaching with additive manufacturing for future in-space manufacturing. [3]

6. What Don’t I Know Yet? 6.1 Can this reach useful productivity at scale? This is the biggest open question. The published work shows biological extraction at lab scale, but the yield is far too low to simply scale up to mission needs.

What I really need to find out is:

Can the process run at much higher regolith loading and volumetric productivity without losing bacterial activity or extraction efficiency?

Answering that means mapping out how these relate to each other:

solids concentration bacterial concentration residence time degree of iron reduction recovered mass reactor volume 6.2 Will it work on real Martian regolith? We use simulants because real Martian soil isn’t available in the amounts needed for lab work. But simulants aren’t perfect copies, and Castelein et al. note differences between JSC-Mars1 and actual regolith in both magnetic properties and mineralogy. [1]

So the process would need to be tested on several increasingly realistic simulants before anything gets deployed. A key question is whether the magnetic enrichment seen with JSC-Mars1 is representative of what would actually be available at a future landing site.

6.3 Will the recovered material meet manufacturing requirements? The published study did print the extracted material, which is a strong proof of concept. But a practical manufacturing system needs much tighter control over composition and batch-to-batch consistency.

So the real question isn’t just “Can it be printed?” It’s:

Can the biological process reliably produce enough material, with predictable enough properties, to make functional parts?

That requires characterizing mechanical properties, composition, particle size, processing behavior, and batch-to-batch variability.

7. Conclusion In this proposal I’ve outlined an automated biological system for Mars that uses Shewanella oneidensis MR-1 to turn Martian regolith into an iron-rich feedstock for additive manufacturing.

The concept builds on existing experimental work: S. oneidensis has already been shown to process Martian regolith simulant, increase both the amount and the iron concentration of magnetically recoverable material, and yield material that can be 3D printed. [1]

What I’m trying to do is shift the question from

Can microbes extract iron from Martian regolith?

to

Can this process be automated, optimized, and scaled enough to become a useful manufacturing subsystem on a Mars mission?

For a first exploration mission, the focus would be on reliable biological processing and producing printable feedstock. Later, the system could combine excavation, biological processing, magnetic separation, feedstock conditioning, and printing into a largely autonomous mine-to-manufacturing platform.

In the end, the point isn’t to make one replacement part. It’s to build the capability to produce things locally from local resources. As missions get longer and farther from Earth, relying on Martian materials instead of shipping everything becomes more and more important. Averesch et al. make a similar argument, pointing to ISRU, loop closure, materials, and biomining as key parts of future off-world manufacturing. [3]

The scale calculation also makes it clear how much work is still ahead. Extrapolating the published lab yield to 1 kg of feedstock means processing about 2.3 tonnes of regolith. That tells me the real research challenge isn’t proving that biological extraction works; it’s making it productive and scalable enough to matter.

References Castelein, S. M., Aarts, T. F., Schleppi, J., Hendrikx, R., Böttger, A. J., Benz, D., Marechal, M., Makaya, A., Brouns, S. J. J., Schwentenwein, M., Meyer, A. S., & Lehner, B. A. E. (2021). Iron can be microbially extracted from Lunar and Martian regolith simulants and 3D printed into tough structural materials. PLOS ONE, 16(4), e0249962. https://doi.org/10.1371/journal.pone.0249962 NASA. Rendezvous with Mars. Houston We Have a Podcast. https://www.nasa.gov/podcasts/houston-we-have-a-podcast/rendezvous-with-mars/ Averesch, N. J. H., Berliner, A. J., Nangle, S. N., Zezulka, S., Vengerova, G. L., Ho, D., Casale, C. A., Lehner, B. A. E., Snyder, J. E., Clark, K. B., Dartnell, L. R., Criddle, C. S., & Arkin, A. P. (2023). Microbial biomanufacturing for space-exploration—what to take and when to make. Nature Communications, 14, 2311. https://doi.org/10.1038/s41467-023-37910-1 Petzold, C. J., & Mukhopadhyay, A. (2025). From bench to biofactory: High-throughput technologies and automated workflows to accelerate biomanufacturing. Current Opinion in Biotechnology, 94, 103320. https://doi.org/10.1016/j.copbio.2025.103320 Asin-Garcia, E., Fawcett, J. D., Batianis, C., & Martins dos Santos, V. A. P. (2025). A snapshot of biomanufacturing and the need for enabling research infrastructure. Trends in Biotechnology, 43(5), 1000–1014. https://doi.org/10.1016/j.tibtech.2024.10.014 NASA Science. Inspecting Soils Across Mars. NASA/JPL-Caltech/University of Guelph. https://science.nasa.gov/resource/inspecting-soils-across-mars/ AI Use Disclosure I used AI to help brainstorm the concept, work through the relevant scientific considerations, structure the assignment, and refine the engineering calculations.

Tools used:

ChatGPT (OpenAI), GPT-5.6 Luna – brainstorming, structuring, and drafting Claude (Anthropic), Opus 5.5 – language editing and Markdown formatting of the final draft Prompts used (ChatGPT):

“Okay space biomanufacturing sounds good. we could produce necessary materials that are needed in deep space travel. Or we can focus on bacteria that help us produce the ore processing. Like Li ore processing might be difficult…Suggest previous studies in this topic” Follow-up prompts to refine the product, the Mars mission scenario, the chassis choice, the automation scope, the scale assumptions, and the motivation. Prompt used (Claude): “Can you write this content in a markdown format.”

I used AI as a research and drafting aid, not as a replacement for reading the sources. I’ve tried to check the scientific claims and numbers against the cited papers, and any remaining errors are my own.