Preliminary Results for In-Situ Alternative Propellants for Nuclear Thermal Propulsion

I.A. Water as a Propellant

Transforming water into hydrogen and oxygen will require the use of electrolyzers, which will require power. Therefore, from a lunar infrastructure standpoint, it may be more economical to use water directly in a propulsion system. There were two studies that analyzed this: (1) a Cornell University study looked at onboard electrolization of water for both rapid and slow electrolysis for use in chemical engines,^Citation6 and (2) Zuppero et al. looked at using water directly in a water alternative propellant nuclear thermal propulsion (A-NTP) engine.^{Citation7–11}

I.A.2. Using Water Directly as a Propellant

Nuclear thermal propulsion (NTP) produces heat by nuclear fission, allowing any fluid to function as a propellant if it does not degrade the reactor materials. The propellant of choice for these engines is hydrogen due to its low molecular weight resulting in an I_sp potential of up to 900 s with current materials. However, liquid hydrogen has a density that is only 7% that of water, which will result in high-volume tanks that will add to the dry mass of the vehicle. Furthermore, hydrogen storage is currently a challenge given that at a pressure of 1 atm, to keep hydrogen in the liquid state, a temperature of 20 K must be maintained.^{Citation12,Citation13} Therefore, it is desirable to explore other alternatives to using hydrogen-based nuclear thermal propulsion (H-NTP) engines.

A bleed cycle water A-NTP engine was proposed by Zuppero et al. with the engine schematic shown in Fig. 1. The I_sp of this engine was limited to 198 s due to the use of zirconium carbide as the fuel cladding material, which only had oxidation resistance at temperatures up to 1200 K sustainably at the water A-NTP reactor flow rates and pressures which resulted in a chamber temperature of 1100 K (CitationRefs. 7 through Citation11). Since then, advances in emergency reactor operation materials have been made with silicon carbide (SiC) being able to operate up to 5 h at temperatures up to 2400 K at the water A-NTP reactor flow rates and pressures according to analytical predictions.^Citation14 The limitation of this material is the recession rate due to oxidation, which becomes faster at higher pressures.^Citation15 Lower recession rates, and thus longer operation times, are possible at temperatures around 1400 K. However, no work has analyzed SiC between temperatures of 1400 and 2400 K and at flow rates and pressures characteristic of the A-NTP (CitationRefs. 16, Citation17, and Citation18).

Fig. 1. Bleed cycle water A-NTP engine schematic.Citation¹⁰

Display full size

The bleed cycle water A-NTP engine also considers boiling inside the reactor and can function only using the hot bleed flow to power the turbines since all flow upstream of the chamber will be liquid.^Citation8 The boiling inside the reactor will result in high errors in analytical predictions due to two-phase heat transfer correlations yielding results within 50% to 100% error even in laboratory conditions.^{Citation19,Citation20} Reliable operation of a rocket engine requires supercritical phase change to discourage boiling.^Citation21 The use of the hot bleed flow significantly limits the chamber temperature up to 1150 K according to current turbine material limitations.^{Citation22–24} Therefore, to produce higher I_sp from a water A-NTP engine, an expander cycle must be utilized.

I.B. Ammonia as a Propellant

Ammonia dissociation was analyzed in a solar thermal thruster. The study focused on the region of the engine from the chamber to the nozzle exit. It was determined that flow can be considered as frozen (no chemical variations) below temperatures of 2500 K due to the slow kinetics of ammonia. Significant effects of dissociation have shown to only be present at chamber temperatures above 3000 K. The largest of these effects was an increase in I_sp by tens of seconds. This study did not present any kind of engine cycle or consider turbomachinery.^Citation25

A mission analysis study was conducted that compared the use of ammonia A-NTP and H-NTP engines for the use of deep space science missions. This preliminary analysis assumed an ideal I_sp of 440 s for ammonia and 900 s for hydrogen for a chamber temperature of 2586 K. Current H-NTP considerations show that 2700 K is needed to produce the 900-s I_sp, suggesting gross overestimation of this parameter in this study.^{Citation26,Citation27} No ammonia A-NTP engine architecture or power balance model was found in literature.

II.A. Water Expander Cycle A-NTP

A preliminary expander cycle analysis on a water A-NTP engine was conducted with the goal of producing 25 000 lbf of thrust according to the H-NTP engine baseline provided by Aerojet Rocketdyne^Citation13 (AR). BWX Technology, Inc.’s (BWXT’s) reactor designed for NTP is considered, which features a moderator block with cylindrical fuel elements inside arranged in rings and a reflector on the outside of the moderator block with cylindrical control drums arranged in a ring inside. Further detail is export controlled. Note that this core was optimized for a gaseous hydrogen coolant/propellant. The consequences of using a different coolant, such as supercritical water at 30 MPa and supercritical ammonia at 20 MPa, need to be carefully evaluated and the core redesigned if necessary.^{Citation28,Citation29} It is assumed that water will be stored in tanks at 300 K with minimal heating required to keep it a liquid.

The schematic of the proposed engine is shown in Fig. 2 with the accompanying state point values shown in Table II. The required mass flow rate of water resulted in 35.37 kg/s. This is almost three times the mass flow rate of pure hydrogen, which is about 12.8 kg/s (CitationRef. 13). The selected reactor exit (chamber) temperature is 2400 K, which produces 320.4 s of I_sp and requires a total reactor power of 334 MW(thermal). As a comparison, H-NTP engines require 530 MW(thermal) to produce the same thrust. The total reactor power was iteratively modified until the 2400 K chamber temperature was met and was distributed to the control drums, moderator elements, and fuel elements according to BWXT’s proprietary heat fraction values.^{Citation13,Citation30} The outlet temperature of water was determined through fluid enthalpies based on the inlet conditions and the heat that was introduced into each element. To prevent boiling and work with supercritical water, the pump output pressure was set to be 304 atm and corresponds to State 2. The temperature of the water reservoir was assumed to be 300 K and further assumed to be stored as a liquid throughout the duration of the mission by the use of multi-layer insulation and strip heaters, which were previously considered for long-duration ammonia storage.^Citation27

Fig. 2. Expander cycle A-NTP engine schematic.

Display full size

To facilitate water’s transition into the supercritical phase, the regenerative cooling lines of the nozzle (State 2 to State 3) and control drums (State 3 to State 4) along with the moderator element cooling lines (State 4 to State 5) will not split the flow as they do in the AR H-NTP engine.^Citation13 Instead, the flow will cascade serially from the regenerative cooling to the moderator cooling. By assuming that the same fraction of heat will be taken from each of these lines as in the baseline H-NTP engine,^{Citation13,Citation30} the resulting outlet temperature from the cascaded lines was found to be 511.7 K with a pressure of 278 atm (pressure losses are assumed). These values correspond to State 5 in Fig. 2.

However, water is still a liquid at this pressure and temperature at State 5 since its critical pressure and critical temperature are 218 atm and 647 K, respectively. Therefore, a pass through 35% of the fuel elements is necessary to bring the water to a supercritical state to use in the turbines. This is analogous to a preburner in chemical engines to increase the enthalpy of the fluid to power the turbines. The resulting temperature coming out of the preheating fuel elements is 762.6 K at a pressure of 236 atm corresponding to State 6 in Fig. 2 and Table II. Although the heat deposition modeling allowed insight only into the inlet and outlet conditions of the preheating fuel elements, it is expected that at the transition point into supercritical water, roughly between temperatures of 600 and 750 K, a lower temperature gradient will be apparent within those temperature bounds corresponding to the extremely high heat capacity and high enthalpy gradient within those temperature bounds, according to Fig. 3, obtained by using the CoolProp thermophysical property library.^Citation31 However, the heat transfer mechanism will still only consider single phase since no boiling will occur. This will be examined in further detail in future work.

Fig. 3. Specific heat capacity and enthalpy of supercritical water.Citation³¹

Display full size

The supercritical water is transferred in between the reactor core components and to the turbines via lines made of an oxide dispersion–strengthened steel, specifically MA/ODS 12YWT alloy with nominal composition of Fe-12.3 wt% C-3% W-0.39% Ti-0.25% ${\mathrm{Y}}_2{\mathrm{O}}_3$. This material has shown oxidation and creep resistance up to 1100 K and radiation resistance with regard to swelling and embrittlement.^Citation32 Furthermore, the turbine material was assumed to be Nimonic 90 with creep resistance up to 1150 K (CitationRef. 24) and a metallic ceramic coating A08 from Sulzer to prevent oxidation.^Citation33 The coating inside the flow channels was assumed to be SiC as this material has oxidation resistance up to 1400 K (CitationRefs. 14, Citation16, and Citation17) and is also currently a considered material for the coating by BWXT and NASA for H-NTP engines.^Citation34 It can also offer oxidation resistance up to 2400 K (CitationRef. 35) at the expense of shortening the engine life to 5 h according to SiC’s recession rate at this temperature^{Citation15,Citation35} and the coating thickness.^{Citation28,Citation29} A more detailed analysis on the engine life is left for future work.

The supercritical water enters a turbine circuit (State 7 through State 10) with a bypass valve (State 9 and State 10). The turbine pressure ratio and efficiency were arbitrarily assumed to be 2 and 0.5, respectively, for conservative estimations based on prior H-NTP turbomachinery work.^{Citation13,Citation36} These values yield that 58.6% of the total flow will be bypassed. As a rule of thumb, about 10% should remain as bypassed to retain control of the turbine throttling, resulting in 48.6% of the flow being available to produce other work.^Citation37 The turbine circuit output temperature and pressure corresponding to State 11 in Fig. 2 and Table II are 657 K and 116 atm, respectively, and correspond to superheated steam.

After passing through the turbines, the flow passes through a final shut-off valve and the pressure is reduced by 10% to provide a pressure margin.^Citation37 Afterward, it enters the other 65% of the fuel elements and gets heated up to 2400 K with an outlet pressure of 83.4 atm. These values correspond to State 13 in Fig. 2 and Table II.

According to water fluid properties, at the inlet of the main heating fuel elements, the temperature gradient of the water will be smaller than in the rest of these elements due to the higher specific heat capacity as shown in Fig. 4, however, it will not be as small as in the preheating fuel elements. If the outlet temperature of the turbines is raised to around 800 K, then there will be no noticeable effect on the temperature gradient of the water. Furthermore, the heat transfer will also only consider single phase as the fluid is superheated steam throughout. Details of this heat transfer mechanism will be examined in detail in future work.

Fig. 4. Specific heat capacity and enthalpy of superheated steam.Citation³¹

Display full size

Finally, the flow enters the nozzle and expands, thus producing 25 klbf of thrust at 320.4 s of I_sp. The diameter of the throat is 0.091 m with an area ratio of 400 from the baseline H-NTP engine.^Citation13 Increasing the area ratio to 1000 will increase I_sp by 5.6 s, however, this difference may be marginal given the additional nozzle mass needed.

II.B. Ammonia Expander Cycle A-NTP

To analyze an ammonia A-NTP engine, the same approach for water was taken and the same engine flow schedule, as shown in Fig. 2, was used with the state values shown in Table III. This produced 25 klbf of thrust at 381.6 s of I_sp and a required total mass flow rate of 30.41 kg/s of ammonia with a total reactor power of 361.8 MW(thermal). The chamber temperature to produce this I_sp is 2650 K and is the same as the baseline H-NTP engine.^Citation13 However, since the critical pressure and critical temperature of ammonia are lower than water at 112 atm and 405 K, respectively, only 12% of the fuel elements were allocated for preheating, which resulted in the turbine inlet temperature to be at 538 K. This will also result in lower thermal strain on the turbines than the water A-NTP engine. Ammonia will also exhibit a small thermal gradient in a region inside the preheating fuel elements due to the transition from a liquid into a supercritical fluid corresponding to the high specific heat capacity between 400 and 500 K, as shown in Fig. 5. However, this will not be as extreme as in the case of water due to the specific heat capacity increasing by only half that of water.

Fig. 5. Specific heat capacity and enthalpy of supercritical ammonia.Citation³¹

Display full size

A key performance measure is the amount of dissociation by thermolysis, species present, and how the species concentration with respect to temperature curves change as the pressure is increased. Furthermore, metals such as tungsten can serve as catalysts for hydrogen production by absorbing the nitrogen in ammonia. This will result in nitride layers that penetrate more with increasing pressure.^Citation38 It is critical for the coating to not serve as a catalyst and to not react with either nitrogen or hydrogen at high pressures and temperatures. Baseline materials such as zirconium carbide and SiC for the AR H-NTP engine will not serve as catalysts for nitrogen absorption since no oxygen will be present.^{Citation3,Citation4,Citation34,Citation39} Since the ammonia solar thruster study determined that ammonia dissociation is slow and will not have time to occur at temperatures below 2500 K and at the high flow rate velocities inside the fuel elements,^Citation25 it is conservative to assume that no significant dissociation occurs inside the engine.

III.A. Constant Volume Analysis

A mission performance comparison was made that analyzes the impact to the mission time if the propellant volume were to stay the same as that of the baseline. The analysis was conducted with the MTV using H-NTP, H2+LOX chemical, water A-NTP, and ammonia A-NTP engines with the summarized results shown in Table IV. The hydrogen tank mass was subtracted from the dry mass in Fig. 7 and the mass of the tank for the propellant used was added. The relations for the tank masses used are shown in EquationEq. (1)(1) $m_{\mathrm{L}{\mathrm{H}}_{2}tank}=9.09\mathrm{\forall }+2.88{\left(\frac{\mathrm{\forall }}{4\pi /3}\right)}^{1/3},$(1) for hydrogen, EquationEq. (2)(2) $m_{\mathrm{L}\mathrm{O}\mathrm{X}tank}=12.16\mathrm{\forall }+1.123{\left(\frac{\mathrm{\forall }}{4\pi /3}\right)}^{1/3},$(2) for oxygen in the case of the H2+LOX chemical, and EquationEq. (3)(3) $m_{othertank}=12.16\mathrm{\forall }.$(3) for both water and ammonia^Citation43:

(1) $m_{\mathrm{L}{\mathrm{H}}_{2}tank}=9.09\mathrm{\forall }+2.88{\left(\frac{\mathrm{\forall }}{4\pi /3}\right)}^{1/3},$(1)

(2) $m_{\mathrm{L}\mathrm{O}\mathrm{X}tank}=12.16\mathrm{\forall }+1.123{\left(\frac{\mathrm{\forall }}{4\pi /3}\right)}^{1/3},$(2)

and

(3) $m_{othertank}=12.16\mathrm{\forall }.$(3)

The propellant volume and payload mass were kept constant in all cases, and it was assumed that the SLS will lift partially tanked stages to LDHEO where they will be transferred to NRHO and be fully tanked there prior to performing the mission. When looking at all these cases together, water provides the highest ΔV and the shortest travel times followed by H2+LOX chemical and ammonia. The H-NTP, on the other hand, provides the lowest ΔV due to the propellant’s low density. This confirms the fact that water is a viable solution given its superb performance in terms of ΔV and accessibility across the Solar System^Citation10 for the same propellant volume. However, H2+LOX chemical is a very close second choice.

III.B. Constant ΔV Analysis

The same analysis was performed but the ΔV was kept constant at 4222 m/s and the propellant volume was varied along with the number of inline stages. Table V shows the summary of the same considered propellants. Here H2+LOX chemical, water A-NTP, and ammonia A-NTP reduce the propellant volume significantly, resulting in only needing one inline tank. Furthermore, water outperforms both H2+LOX chemical and ammonia A-NTP in terms of decreasing the propellant volume. The vehicle that uses water results in having the highest initial gross mass, which is 83.45% greater than the initial gross mass of the H-NTP baseline. In contrast, H2+LOX chemical resulted in 21.75% greater initial gross mass and ammonia A-NTP resulted in 52.66% greater initial gross mass from the baseline. Although ammonia is present on the Moon and other bodies in the Solar System, it is still not as abundant as water and does not have the dual functionality as a propellant for the engine and a consumable for the life support system. However, ammonia will still be produced as it must be scrubbed from the mined lunar water. Efforts will need to be made to store the excess ammonia instead of venting it out.Citation¹ Coupling this with ammonia’s worse performance, the use of ammonia in an A-NTP engine should only be considered as a supplemental propulsion system if excess ammonia from water mining operations is available and not as a propulsion system of choice for the Mars conjunction mission.