Realisation of a Cold Gas Microthruster

Among the several micropropulsion systems a Cold-Gas based microthruster has been selected due to the reasonable maturity of the required technology and the possibility to realise a working prototype in the time frame of the present study. A

Cold-Gas system has been therefore designed, prototyped, tested and integrated in a nanosatellite. Further improvements on the system, namely the heating of the gas and the conception of a system to provide such heating in an efficient way, have been studied as basis for further research steps.

Fig. 9.14. Pulses vs. size and ACS accuracy

9.4.1 Gas- and Fluid Dynamics

The preliminary design of the micronozzle has been done using a simple one-dimensional model for the adiabatic flow of the compressible fluid with viscous friction (Oates 2004). With this model the first estimation of the nozzle thrust has been possible. The formulae used to take in to account of the losses are certainly very simplified and have been originally developed for macroscopic systems (Manzoni 2000). The coefficients of the formulae maintain their validity partially in the microscopic scale. In order to have a better idea of the flow in the designed nozzle a CFD simulation is performed. In order to use the Navier-Stokes equations the characteristics of the flow has to be checked (Gad-el-Hak 1999). In fact fluid flows in small devices differ from those in macroscopic applications because of:

• Higher surfaces effects

• The mean free path is close to the characteristic dimensions

• There could be a relative movement between the fluid and the walls (the no slip-condition looses availability)

As depicted in the Fig. 9.15 where the above-mentioned limits are traced and the working range of the microthruster is represented in the preliminary design, we observe that the Navier-Stokes equation and the continuum approach are still applicable.

The CFD simulation shows that a certain selection of nozzle length introduces losses higher than expected and the exit pressure is reached well before the end of the nozzle. With the help of such information, the length of the nozzle can be reduced in order to obtain an optimised design for better efficiency.

Flow character











Normalised gas density



Flow Character Limits

Flow Character Limits


molecular chaos molecular chaos


dilute gas ^thruster density p/p 0

dilute gas ^thruster density p/p 0

Navier-Stokes Validity area

Fig. 9.15. Flow Character

9.4.2 Prototyping

The manufacturing of the micro components has been based on silicon and glass assembled by means of anodic bonding and on some packaging techniques for the assembly with the other COTS, electronic and mechanical components. An investigation on the components for micropropulsion available on the market has been conducted and the results are represented in the Fig. 9.16.

We have observed that ancillary components like microvalves or pressure sensors are available on the market also in MEMS configuration while the heart of the propulsion system, the micronozzle, has been only demonstrated by some universities. The micro fabrication focus has been concentrated on the microthruster, while the other necessary MEMS components have been selected from devices available on the market. The preparation of the nanolithography mask and the lithography of the micronozzle have been performed initially from the Italian Institute for the Physics of the Matter in Trieste in cooperation with the company Microspace. Such process is an adaptation of the state of the art,

Reactive-Ion Etching (RIE) and has been optimised for the smoothness of the surface. Fig. 9.17 and Fig. 9.18 represents the recipe and sequence of the steps for the realisation of the microlitography mask and the lithography on silicon to realise the micronozzle (Santoni et al. 2002). Fig. 9.19 shows a view of the micronozzle in comparison with a human hair (about 100 micrometers). The Fig. 9.20 shows a detail of the throat area after a cut with microsaw; the round profile of the channel, due to the particular isotropic etching given by the RIE, can be noticed.


Chemical Micropropulsion Technology Readiness Evaluation


Miniaturised pressure reduction system Gas Generation System miniaturized NC valve not available concept and basic element available


under development under development commercially available pressure sensors microvalves microcombustion chamber micronozzle microturbme micropump microelectrical generator commercially available available but leaking and slow actuation not available demonstrated under development available for low pressure demonstrated commercially available commercially available under development available demonstrated available for medium pressure

Fig. 9.16. Technology readiness evaluation

500 |tm thick Si wafer substrate

2|tm thick, Si3N4 membrane deposition

Deposition of double layer of 5 nm of chromium and 10 nm of gold to be used as electrical contact for the electroplating

PMMA resist, spun on the substrate at 0.6 pmthikness

PMMA baking on a hot plate (7 minutes at 170-C) and annealing (10 minutes at 30°).

mask pattern using a JEOL 6400 lithographic system, working at 30 KeV, 1 nA beam current and 2.5 mm field size.

PMMA developing for 30 seconds at 20°C in 25% Methyl-isobutil-ketone (MIBK) and 75% of Isopropil alcohol (IPA)

Au electro-plating of 300 nm using a commercial gold cyanide electroplating apparatus, (current density 3 mA/cm 2 for 2 min)

Fig. 9.17. Mask nanolithography process

Fig. 9.21 shows a detail of the throat by focussing the microscope to the upper surface of the silicon and the bottom surface of the channel. In the first figure we notice the smoothness of the channel geometry. The narrowest part of the throat is 40 micrometers wide. The small irregularities in the round geometry are due to the errors in the optical transfer process from the optical mask to the RIE metallic temporary mask attached to the silicon surface. A geometry analysis can confirm the accuracy. The analysis shows differences of 1 or 5 micrometers from the several thrusters produced on the same wafer.

2 inch B-(lopcd 500 [mi thick Si wafer substrate

800 urn thick UVIII resist deposition soft-balcing (60 seconds at 150°C) and annealed (10 min at 30°C)

resist sincrotam X-ray exposure with 300mJ/cm 2 using a step-repeat procedure post baking (140°C for 75 s) and developing in CD26 (90s)

evaporate 20 nm nickel film

Unexposed resist (and nickel) removed in hot acetone and ultrasounds

ftunrocartxjn based RIE with -CF4 90%, 02 10% lulal fluí of •3Qsccm, jkesske of lOmTiif •Mai power 190 Wat and •DC bös of 100 Vota nickel removal with HN03 diluted solution

Fig. 9.18. Micronozzle nanolithography process

After the wafer with the thrusters has been produced it will be bonded to the support glass, previously prepared (with microholes) for each thruster. Fig. 9.23 shows the final result of the process. It is primarily a cube of 4x4x4 mm with one thruster, ready to be assembled in the micropropulsion system. An improved version of the micronozzle, shown in Fig. 9.24 is now under development at the National Institute for Advanced Industrial and Science Technology (AIST) in Japan in collaboration with the Microspace, Italy (Manzoni 2004). The improvements concern the use of a more repeatable process, better characterisation of the performances and realisation of a reliable packaging. The assembly of the main thruster's components comprehend essentially 4 elements.

• Micronozzle

• Microvalve

• Pressure sensor

• Mechanical structure

Several architectures are possible with respect to the interfaces of the single selected components. Fig. 9.25 shows a typical assembly structure to be used (Graziani et al. 2000) for commercial microvalve and pressure sensor by the company Redwood and Fraunhofer, respectively. Such assembly has also been mounted on the Italian microsatellite UNISAT-2 of the University of Roma "La Sapienza", launched in December 2002. Due to the big size of the packaging of the Redwood microvalves and the Fraunhofer pressure sensors, a very big volume is introduced between the valves and the thrusters. This kind of assembly does not allow a fast dynamic of the thrust variation. Furthermore the Redwood microvalves, being thermally actuated, are not functioning over 70 degrees Celsius.

Fig. 9.19. Micronozzle microscope photo (25x) with hair as comparison (100 p.m);


Fig. 9.20. Micronozzle section cut photo

Fig. 9.21. Micronozzle throat microscope photo
Fig. 9.24. AIST-Microspace Micronozzle
Fig. 9.25. Microthruster assembly

With this kind of assembly, a group of two thrusters occupies a volume of about 50 cc for a mass of about 100 g. The use of the commercial microvalves from the company Lee and the Intersema pressure sensor allows the conception of a much more compact thrusters group. In a volume of 5 cc and with a mass of less than 20 g it has been possible to assembly a group of 3 thrusters for 3 different thrust vectors. Such compact assembly allows faster thrust dynamic. The more compact design requires the use of 3 machined aluminium parts with grooves for the gas channels as shown in the Fig. 9.26 (Manzoni 2003).

Fig. 9.26. High dynamic microthruster assembly

The micronozzles have been mounted on a microbalance realised by the author, to measure |iN thrust in vacuum. A typical measurement of thrust is based on the detection of the acquisition. Averaging the balance oscillations carries out the calculation of the thrust. The thrust for different gas supply pressures and vacuum conditions is presented in the Fig. 9.27.

thruster n600

pressure (rel. bar)

thruster n600

pressure (rel. bar)

vacuum %

vacuum %

Fig. 9.27. Thrust experimental results

Was this article helpful?

0 0
Brain Blaster

Brain Blaster

Have you ever been envious of people who seem to have no end of clever ideas, who are able to think quickly in any situation, or who seem to have flawless memories? Could it be that they're just born smarter or quicker than the rest of us? Or are there some secrets that they might know that we don't?

Get My Free Ebook

Post a comment