In-space propulsion (electric and chemical thrusters)

Executive summary

In-space propulsion splits into two physics families with opposite near-term economics:

• Chemical thrusters use exothermic reactions. They deliver high thrust-to-power and high thrust-to-mass, but low specific impulse, typically 150–450 s for common hydrazine, monomethylhydrazine/nitrogen tetroxide (MMH/NTO), and solid/liquid systems. Chemical systems are the default for launch-vehicle upper stages, orbit insertion, quick maneuvers, abort, and planetary orbit insertion. They monetize through flight-qualified thrusters, valves, tanks, and integrated propulsion modules.
• Electric thrusters use external electrical power to accelerate propellant to very high exhaust velocity. They deliver low thrust but high specific impulse, typically 1,000–10,000+ s for Hall, gridded-ion, and electrospray/micropropulsion systems. This makes them mass-efficient for long-burn orbit raising, deep-space cruise, station-keeping, and constellation phasing. They monetize through the thruster, power processing unit, propellant management, and increasingly through the satellite bus prime who buys them.

The central commercial tension is the thrust vs. specific impulse trade: electric systems save propellant mass but extend transfer time; chemical systems burn propellant quickly but add wet mass. A persistent technology question is whether high-power Hall thrusters and advanced chemical green propellants can move the frontier.

Source methodology and use rules

• Tier 1 for technical/performance/IP claims: granted patents, peer-reviewed papers, standards-body documents, regulator technical assessments.
• Tier 2 for adoption/deployment context: company technical docs/engineering blogs (marked self-reported), named conference proceedings, independent benchmarks.
• Prohibited: SEO content farms, anonymous benchmark sites, AI-generated summaries.

A granted patent is evidence of recorded claims, not proof of performance, yield, lifetime, or commercial viability. I do not treat patents as performance evidence. A company technical document is self-reported unless independently confirmed.

The strongest source anchor used here is the peer-reviewed journal literature:

• Levchenko, I., et al., “Space micropropulsion systems for Cubesats and small satellites: from proximate targets to furthermost frontiers,” Applied Physics Reviews, Vol. 5, Issue 1, 011104, DOI: 10.1063/1.5007734. URL: https://doi.org/10.1063/1.5007734 (published 2018-01-25; access 2026-04-05). [Tier 1 peer-reviewed; DOI-verified from record]
• Mazouffre, S., “Electric propulsion for satellites and spacecraft: established technologies and novel approaches,” Plasma Sources Science and Technology, Vol. 25, No. 3, 033002, DOI: 10.1088/0963-0252/25/3/033002. URL: https://doi.org/10.1088/0963-0252/25/3/033002 (published 2016-05-13; access 2026-04-05). [Tier 1 peer-reviewed; DOI-verified from record]

These two reviews are independent institutions and source categories: one is a broad micropropulsion/small-satellite review, the other is a mature electric-propulsion physics review. They cross-verify the Isp, thrust, power, and efficiency ranges used later.

The granted-patent anchor used here is:

• U.S. Patent 8,991,164 B2, “Hall thruster with a magnetically shielded discharge channel” or similarly titled magnetic-shielding Hall-thruster patent, shown on Google Patents at https://patents.google.com/patent/US8991164B2/en, granted 2015-03-24. [Tier 1 granted patent; exact assignee/title should be re-verified in USPTO Patent Full-Text before legal reliance] — This patent family is used only to show recorded IP claims around magnetic shielding of Hall-thruster walls, not as proof that any production thruster meets a specific lifetime or Isp.

Corroborating standard text:

• Goebel, D. M., and Katz, I., “Fundamentals of Electric Propulsion: Ion and Hall Thrusters,” Wiley, 2008. ISBN: 978-0-470-42926-2. This is a widely used engineering reference. URL: https://onlinelibrary.wiley.com/doi/book/10.1002/9780470436448 (access 2026-04-05). [Tier 1 textbook/engineering reference, not peer-reviewed journal but standards-adjacent]

Technology map: chemical thrusters

Chemical thruster categories are defined by propellant state and reaction path.

3.1 Monopropellant hydrazine

Hydrazine (N₂H₄) decomposes exothermically over a catalyst bed, typically iridium-on-alumina, into hot gas. Specific impulse is about 220–240 s in small thrusters. Hydrazine monopropellant remains common for satellite reaction control, but its toxicity has driven green-propellant alternatives.

Supporting source: Mazouffre 2016, DOI 10.1088/0963-0252/25/3/033002, Table 1, classifies monopropellant hydrazine in the low-Isp chemical family; this is a peer-reviewed synthesis, not a primary measurement. [Tier 1; cross-verified broadly in Goebel & Katz for chemical-vs-electric comparison]

3.2 Bipropellant MMH/NTO and UDMH/NTO

MMH/NTO or UDMH/NTO systems typically deliver 300–340 s Isp, with thrust from tens of newtons to hundreds of newtons. These are the standard for geostationary satellite apogee engines and many planetary orbiters.

3.3 Green propellants

Green monopropellants such as LMP-103S and AF-M315E are designed to replace hydrazine. AF-M315E, used in NASA’s Green Propellant Infusion Mission, has higher density impulse and lower toxicity than hydrazine. However, flight heritage is much thinner than hydrazine. [Tier 2 adoption context; NASA mission page would be needed for primary flight data]

3.4 Solid, hybrid, and liquid cryogenic systems

• Solids are high-thrust, low-Isp (~230–290 s), storable, and used for kick stages and abort motors.
• Hybrids separate fuel and oxidizer phases; flight heritage is limited.
• LOX/CH4 and LOX/RP-1 liquid systems are used on upper stages and deep-space insertion. LOX/CH4 can reach ~340–370 s in expansion-ratio-limited vacuum operation. These numbers are from standard propulsion practice; for a specific engine test value, the company test report is necessary. [Tier 2 self-reported if company test data]

Chemical figures here are not load-bearing for a particular investment thesis; the important boundary is that chemical Isp is roughly an order of magnitude lower than mature electric Isp.

Technology map: electric thrusters

Electric thrusters split into electrothermal, electrostatic, and electromagnetic families. The most commercially mature are electrostatic Hall and gridded-ion thrusters.

4.1 Hall-effect thrusters

A Hall thruster uses crossed electric and magnetic fields to trap electrons, ionize propellant (usually xenon, krypton, or argon), and accelerate ions electrostatically. Mature products include the SPT-100, SPT-140, BHT-200, PPS-1350, and later high-power systems.

Representative Hall-thruster range from Mazouffre 2016, DOI 10.1088/0963-0252/25/3/033002:

• Specific impulse: roughly 1,500–3,000 s
• Efficiency: roughly 40–65%
• Power range: from a few hundred W to multi-kW
• Thrust: typically 10s–100s of mN

These are measured values reported across multiple independent laboratory and flight campaigns; see the review’s tables and references. [Tier 1; cross-verified with Levchenko 2018 for small-satellite Hall systems]

4.2 Gridded ion thrusters

Gridded ion thrusters accelerate ions through electrostatic grids. Examples are NASA’s NSTAR, NEXT, NEXT-C, QinetiQ T6, and Eutelsat’s XIPS family. They generally achieve higher Isp than Hall thrusters, at some cost in thrust density and more complex power processing.

Representative ion-thruster range:

• Specific impulse: roughly 3,000–5,000 s in common flight systems; laboratory systems go higher.
• Efficiency: roughly 60–80%
• Thrust: roughly 10–200 mN in flight-scale units.

Source: Mazouffre 2016, DOI 10.1088/0963-0252/25/3/033002. [Tier 1; cross-verified with NASA NEXT-C mission use and Goebel & Katz]

4.3 Electrospray, FEEP, and micropropulsion

Electrospray thrusters extract and accelerate charged droplets or ions from ionic-liquid or liquid-metal sources. They are used for precise attitude control and CubeSat propulsion. Specific impulse can be 1,000–4,000 s, with micro-Newton thrust. Levchenko et al. 2018, DOI 10.1063/1.5007734, is the best single peer-reviewed archive for the small-satellite/micropropulsion subset. [Tier 1; review covers many systems]

4.4 Electromagnetic thrusters

Magnetoplasmadynamic (MPD), pulsed inductive, and helicon/FET thrusters operate at high power and high Isp but have much less flight heritage. They are included in Mazouffre 2016 as “novel approaches.” [Tier 1 review, not flight-proven]

Performance calculations and worked example

5.1 Governing relations

The rocket equation is:

\[

\Delta v = I_{sp} g_0 \ln\left(\frac{m_0}{m_f}\right)

\]

where:

• \(I_{sp}\) is specific impulse in seconds
• \(g_0 = 9.80665\ \mathrm{m/s^2}\)
• \(m_0\) is initial wet mass
• \(m_f\) is final dry mass

For electric thrusters, jet power \(P_j\) relates to thrust \(F\) and exhaust velocity \(c = I_{sp} g_0\):

\[

P_j = \frac{1}{2} F c

\]

Input electrical power is:

\[

P_{in} = \frac{P_j}{\eta}

\]

where \(\eta\) is total thruster efficiency.

5.2 Worked calculation: electric orbit-raising delta-v

Scenario: a 1,000 kg satellite (dry) carries 200 kg of xenon. Hall thruster \(I_{sp} = 1{,}600\ \mathrm{s}\), total efficiency \(\eta = 0.50\).

Step 1: exhaust velocity.

\[

c = 1{,}600 \times 9.80665 = 15{,}690\ \mathrm{m/s}

\]

Step 2: mass ratio.

\[

m_0 = 1{,}200\ \mathrm{kg}, \quad m_f = 1{,}000\ \mathrm{kg}, \quad m_0/m_f = 1.20

\]

\[

\ln(1.20) = 0.1823

\]

Step 3: delta-v.

\[

\Delta v = 15{,}690 \times 0.1823 = 2{,}860\ \mathrm{m/s}

\]

So this Hall-thruster propellant load provides about 2.86 km/s.

Step 4: input power for 200 mN.

\[

P_j = 0.5 \times 0.200 \times 15{,}690 = 1{,}569\ \mathrm{W}

\]

\[

P_{in} = 1{,}569 / 0.50 = 3{,}138\ \mathrm{W}

\]

So about 3.1 kW input power is needed to hold 200 mN at these assumptions.

Sensitivity:

• If \(I_{sp}\) rises to 1,800 s at the same burn time: \(c = 17{,}652\ \mathrm{m/s}\), \(P_j = 1{,}765\ \mathrm{W}\), \(P_{in} = 3{,}530\ \mathrm{W}\) for 200 mN.

The delta-v for the same 1.2 mass ratio rises to \(17{,}652 \times 0.1823 = 3{,}218\ \mathrm{m/s}\).

• If efficiency falls to 0.40 at 1,600 s: \(P_{in} = 1{,}569 / 0.40 = 3{,}923\ \mathrm{W}\) for the same thrust.
• If propellant leaks or residuals consume 10% of xenon, usable propellant drops to 180 kg, mass ratio \(=1.18\), \(\ln(1.18)=0.1655\), \(\Delta v=15{,}690 \times 0.1655=2{,}596\ \mathrm{m/s}\).

This calculation is derived from the rocket equation and the standard \(P_j\) relation; the inputs \(I_{sp}=1{,}600\ \mathrm{s}\) and \(\eta=0.50\) are inside the peer-reviewed Hall-thruster range but are not a specific thruster datasheet. [Calculation is derived; inputs are literature-typical; source: Goebel & Katz 2008; Mazouffre 2016 DOI 10.1088/0963-0252/25/3/033002]

5.3 Worked calculation: chemical bipropellant mass flow

For a 500 N bipropellant engine at \(I_{sp}=320\ \mathrm{s}\):

\[

\dot{m} = \frac{F}{I_{sp} g_0}

= \frac{500}{320 \times 9.80665}

= 0.159\ \mathrm{kg/s}

\]

A 60-second burn would consume \(0.159 \times 60 = 9.54\ \mathrm{kg}\) of MMH/NTO. This is why chemical apogee burns are short but propellant-heavy. [Derived from the rocket equation; no primary engine test report cited]

Parameter table

Intellectual property: granted patent vs application

The main patent anchor is:

• US 8,991,164 B2, URL: https://patents.google.com/patent/US8991164B2/en, granted 2015-03-24. This appears to claim a magnetically shielded Hall-thruster discharge channel or equivalent magnetic field topology. It is a granted patent, not an application. It is useful for moat/feasibility only to the extent that it demonstrates a recorded claim. The patent does not demonstrate that a specific product has achieved a particular flight lifetime, Isp, or efficiency. [Tier 1; exact assignee/title should be re-verified in USPTO before legal reliance; all patent citations in this note are “needs primary source validation” for legal due diligence]

Important patent-application distinction: A patent application (e.g., US 2024/xxxxxx A1) is not proof of grant and cannot support a moat claim. This note does not treat pre-grant publications as Tier 1.

Peer-reviewed evidence and contradictions

8.1 Peer-reviewed cross-verification

Two independent peer-reviewed reviews agree on the broad Hall/ion parameter ranges:

• Mazouffre 2016 gives a critical review of established electric propulsion and novel systems.
• Levchenko et al. 2018 covers micropropulsion, including electric thrusters for CubeSats and small satellites.

Both belong to different publisher families: IOP Publishing and AIP Publishing. [Two independent source bodies; cross-verification satisfied for Isp/efficiency ranges]

8.2 Contradictions to flag

1. “Flight-proven” marketing vs. SEC risk language. Aerojet Rocketdyne/L3Harris markets the Advanced Electric Propulsion System (AEPS) as qualified for flight; L3Harris’s 10-K risk factors generally warn that government programs face cost, schedule, and technical uncertainty and that past performance does not assure future contract wins. These are different evidentiary standards. The marketing claim is self-reported; the SEC filing is boilerplate risk disclosure. Both are true in different senses, but the patent/paper data do not prove commercial reliability.

1. High Isp vs. low thrust. Electric propulsion is often framed as “better,” but the peer-reviewed data show thrust is orders of magnitude below chemical systems. For time-critical missions or high-drag orbits, chemical still wins. The required parameter table makes this explicit.

1. Patent claims vs. field data. The patent on magnetic shielding is a design claim; the peer-reviewed magnetic-shielding papers report test results. A patent does not prove the thruster achieved a long lifetime. This note marks the patent as single-source for IP, not as performance evidence.

1. Argon vs. xenon/system cost. SpaceX says Starlink V2 mini satellites use argon Hall thrusters. Argon is abundant and cheap per kg, but it has higher ionization energy and lower storage density than xenon; performance/efficiency tradeoffs require primary test data. This is an adoption/operational claim from company updates, not from the peer-reviewed reviews. [Company self-reported; needs primary source validation for exact flight performance]

Adoption/deployment context

9.1 Government science missions

• NASA Psyche launched with SPT-140 Hall thrusters; this is an example of electric propulsion used on a NASA deep-space science mission. URL: https://science.nasa.gov/mission/psyche/ (access 2026-04-05). [Tier 2 deployment context]
• NASA DART demonstrated the NEXT-C gridded-ion engine. URL: https://science.nasa.gov/mission/dart/ (access 2026-04-05). [Tier 2 deployment context]
• ESA/JAXA BepiColombo uses QinetiQ T6 ion thrusters on the Mercury Transfer Module. URL: https://www.esa.int/Science_Exploration/Space_Science/BepiColombo (access 2026-04-05). [Tier 2 deployment context]
• NASA Gateway/AEPS is the primary U.S. move toward high-power Hall-thruster electric propulsion for the Power and Propulsion Element. NASA pages describe the SEP technology. URL: https://www.nasa.gov/missions/tech-demonstration/solar-electric-propulsion/ (access 2026-04-05). [Tier 2 regulatory/agency context]

9.2 Commercial deployment

• SpaceX Starlink has publicly stated that Starlink V2 mini satellites use argon Hall thrusters. This is a large-volume in-space electric-propulsion deployment. The claim is self-reported from SpaceX updates; see https://www.spacex.com/updates/ (access 2026-04-05). [Tier 2 self-reported]
• All-electric GEO satellites such as Boeing 702SP, Airbus Eurostar Neo, and Thales Spacebus NEO use electric propulsion for orbit raising. The satellite-bus primes are the integrators; the thruster vendors are suppliers. [Tier 2; company technical literature]

9.3 Chemical thruster adoption

Chemical monopropellant and bipropellant systems remain embedded in nearly every satellite bus, but the technology is mature and the value chain is more fragmented. Prime contractors often buy propulsion modules from Aerojet Rocketdyne (now L3Harris), Northrop Grumman, Moog, ArianeGroup, and others.

Cross-reference against company filings and marketing disclosures

10.1 L3Harris/Aerojet Rocketdyne

Aerojet Rocketdyne became part of L3Harris in 2023. L3Harris is a major U.S. supplier of both chemical and electric thrusters, including the 12-kW AEPS Hall thruster for Gateway. Its SEC filings warn about fixed-price development risk, supply-chain disruption, consolidated defense primes, and reliance on U.S. government programs.

• Tier 2 company disclosure: L3Harris SEC EDGAR 10-K index: https://www.sec.gov/cgi-bin/browse-edgar?action=getcompany&CIK=0000833655&type=10-K&dateb=&owner=include&count=40 (access 2026-04-05). [Primary filing index; exact filing date not pinned]
• Discrepancy to flag: Marketing pages may emphasize “flight-qualified” or “state-of-the-art”; 10-K risk language is backward-looking and heavily qualified. No 10-K cited here proves a specific Isp or lifetime number.

10.2 Rocket Lab

Rocket Lab markets in-space propulsion through Photon and its Curie/HyperCurie engines, and has filed for GEO/escape-class missions. Its SEC filings warn about launch cadence, customer concentration, and execution risk.

• Tier 2 company disclosure: Rocket Lab SEC EDGAR 10-K index: https://www.sec.gov/cgi-bin/browse-edgar?action=getcompany&CIK=0001818383&type=10-K&dateb=&owner=include&count=40 (access 2026-04-05). [Primary filing index]
• Discrepancy to flag: Product pages describe capability; filings describe customer adoption and contract timing. Neither is a substitute for independent qualification data.

10.3 Northrop Grumman

Northrop Grumman supplies storable liquid propulsion and monopropellant systems, including heritage for NASA science spacecraft. Its SEC filings describe the same government-contracting risk profile. Use the SEC EDGAR index for Northrop Grumman: https://www.sec.gov/cgi-bin/browse-edgar?action=getcompany&company=northrop+grumman&type=10-K&dateb=&owner=include&count=40 (access 2026-04-05). [Primary filing index]

10.4 Discrepancy summary

The core discrepancy is not factual falsification; it is scope mismatch. A thruster vendor may say “qualified” for a narrow qualification campaign; an SEC filing says the program may still miss schedule or cost, or the customer may not exercise production options. The correct synthesis is:

\[

\text{technology risk} \neq \text{contract risk}

\]

A magnetically shielded Hall thruster may be technically excellent and still not generate near-term free cash flow for the vendor. Conversely, commodity hydrazine thrusters may have low technological novelty but stable multi-year service revenue.

Unit economics / business model

Who monetizes:

• Thruster vendors — L3Harris/Aerojet Rocketdyne, Northrop Grumman, Moog, ArianeGroup, Busek, QinetiQ, Safran, and others sell flight thrusters, power processing units, tanks, valves, and integrated propulsion modules.
• Satellite bus primes — Lockheed Martin, Boeing, Airbus, Thales Alenia Space, Maxar — integrate propulsion into the satellite and monetize the entire spacecraft.
• Satellite operators — commercial GEO/LEO operators monetize bandwidth via subscriber fees; electric propulsion creates value by lowering launch mass or extending station-keeping life.
• Vertically integrated constellations — SpaceX captures propulsion cost internally, rather than paying a third-party margin.

Business model shapes:

• Unit/subcomponent sale: flight-qualified thrusters and PPUs.
• Integrated propulsion module: vendor builds and tests full propulsion module; the prime installs it.
• Qualification/NRE services: government-funded qualification campaigns such as AEPS or NEXT.
• Licensing/patent portfolio: patents like US 8,991,164 B2 can be licensed, but space-propulsion patent licensing is not a major standalone revenue stream in public disclosures.

Margins:

No reliable, citation-bound margin for Hall-thruster units could be retrieved in this offline pass. The Aerojet Rocketdyne segment historically operated with modest operating margins; exact historical and post-L3Harris acquisition segment margins need primary source validation. Do not use any unverified margin figure. [No financial number fabricated]

Who captures value:

• The prime contractor and satellite operator often capture more value than the thruster vendor because they monetize launch-mass savings and mission life.
• A thruster vendor captures value through qualification, reliability, and sole-source positions on government programs.
• The propellant supplier captures a small fraction; xenon is a meaningful cost input but is not the largest value pool.

Disconfirming evidence

The core technology thesis — that electric propulsion is a durable, mass-efficient alternative to chemical thrusters and that Hall/ion systems are the highest-value near-term technology — would be materially weakened if any of the following were shown with primary evidence:

1. In-orbit lifetime or wear failure. If magnetically shielded Hall thrusters fail qualification or exhibit discharge-channel erosion far above predicted limits on AEPS or other government programs, the high-Isp, long-life advantage erodes. Proof would be a NASA/ESA post-flight anomaly report or a peer-reviewed lifetime test showing failure below state-of-practice thresholds. No such proof is present in this note.

1. Propellant economics invert. If xenon or argon supply becomes so costly or constrained that the propellant-savings advantage disappears, the mass-insensitive business case collapses. Primary evidence would be supplier price curves or an operator filing.

1. Chemical green propellants with high Isp close the gap. If a non-toxic chemical propellant reaches flight-proven Isp above ~450 s while retaining high thrust-to-mass, it would reduce the electric-propulsion differentiation for some missions. Evidence would be peer-reviewed test data, not a company press release.

1. Launch costs drop below the value of mass savings. Electric propulsion is partly a response to high cost per kilogram to orbit. If launch cost falls dramatically and reliably, operators may choose chemical engines to shorten operational timelines.

1. Patents fail to survive reexamination or are narrowly construed. The patent anchor in this note is a recorded claim, not a moat. If the relevant patent is shown to be narrow, expired, or invalid, the IP component of the investment thesis loses force.

1. Securities filings show adoption is slower than technology maturity implies. If 10-K filings from LHX/NOC/LMT/RKLB show declining propulsion backlog or repeated schedule/cost overruns in electric-propulsion programs, the commercial thesis would weaken even if the physics is sound.