A MEMS glass membrane igniter for improved ignition delay and reproducibility

A MEMS glass membrane igniter for improved ignition delay and reproducibility

Accepted Manuscript Title: A MEMS glass membrane igniter for improved ignition delay and reproducibility Authors: Daeban Seo, Juyoung Jeong, Taekyu Ki...

893KB Sizes 1 Downloads 12 Views

Accepted Manuscript Title: A MEMS glass membrane igniter for improved ignition delay and reproducibility Authors: Daeban Seo, Juyoung Jeong, Taekyu Kim, Jongkwang Lee PII: DOI: Reference:

S0924-4247(16)30453-8 http://dx.doi.org/doi:10.1016/j.sna.2017.02.011 SNA 9994

To appear in:

Sensors and Actuators A

Received date: Revised date: Accepted date:

12-9-2016 4-2-2017 10-2-2017

Please cite this article as: Daeban Seo, Juyoung Jeong, Taekyu Kim, Jongkwang Lee, A MEMS glass membrane igniter for improved ignition delay and reproducibility, Sensors and Actuators: A Physical http://dx.doi.org/10.1016/j.sna.2017.02.011 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

A MEMS glass membrane igniter for improved ignition delay and reproducibility

Daeban Seo1, Juyoung Jeong2, Taekyu Kim3, and Jongkwang Lee2†

1

Engine Test and Evaluation Team, Korea Aerospace Research Institute, 169-84 Gwahak-ro,

Yuseong-gu, Daejeon, South Korea, 34133 2

Department of Mechanical Engineering, Hanbat National University, 125 Dongsedae-ro,

Yuseong-gu, Daejeon, South Korea, 34158 3 Department of Aerospace Engineering, Chosun University, 309 Pilmun-daero, Dong-gu, Gwangju, South Korea, 61452



Corresponding author

Tel.: +82-42-821-1081 Fax: +82-42-821-1587 E-mail: [email protected]

Highlights

► This study describes a noble-glass-based MEMS igniter for improved reproducibility and ignition delay with high structural stability and a uniform membrane thickness ► A glass membrane which had a flat surface and uniform thickness was selected as the igniter material for high membrane structural stability. A heater is designed at the under surface of the membrane for direct contact with the propellant to improve ignition delay. ► Numerical simulations were performed to predict and compare ignition characteristics ► The designed glass membrane igniter was realized as an array-type using a MEMS fabrication process with a glass wafer. ► The measured average ignition delay and ignition energy were 17.08 ms and 25.6 mJ, respectively. The standard deviations of the ignition delay and ignition energy were calculated as 1.96 ms and 2.53 mJ, respectively.

ABSTRACT

A MEMS igniter with improved ignition characteristics and reproducibility is described. A glass wafer was selected as the igniter material for high membrane structural stability. To improve the reproducibility of the igniter, the membrane was designed to realize a flat surface and uniform thickness, which are essential factors for reproducibility. A heater is designed at the under surface of the membrane for direct contact with the propellant. It was expected that this would improve ignition delay compared with the previous glass-ceramic membrane igniter. Numerical simulations are performed to predict and compare ignition characteristics. The designed glass membrane igniter is realized as an array-type using a MEMS fabrication process with a glass wafer. Performance evaluation of the fabricated igniter is conducted through a

2

firing test. At the cubesat’s operational voltage of 15 V, the measured ignition delay was 17.1 ms, which is almost the same as the numerical simulation result. Additionally, this result is 34.45% shorter than the measured ignition delay of the previous glass ceramic membrane igniter. The reproducibility is evaluated by consecutively igniting five igniters at 15 V. The calculated average ignition delay and its coefficient of variation are 17.08 ms and 12%, respectively.

Keywords: MEMS igniter; MEMS solid propellant thruster; Glass membrane; Ignition characteristics

3

1. Introduction Satellites have become important to modern society, but their exorbitant launch cost is still a burden in developing or operating them. To overcome this point, CubeSats whose masses are less than 10 kg, have been developed in academic and industrial fields [1,2]. These satellites can significantly save launch costs while improving mission capabilities and success rates by using constellations rather than individual satellites, as many CubeSats can be launched simultaneously [3,4]. A propulsion system is recommended for CubeSats in order to form constellations and increase the lifespan of CubeSats by station-keeping [5]. Minimization of propulsion systems for CubeSats should be realized using MEMS technology. Unlike propulsion systems for micro-satellites or conventional satellites, feasibility of fabrication is the most important factor in achieving propulsion systems for CubeSats [6]. From this point of view, there are some difficulties in using electrical thrusters in terms of minimization and systematization owing to their complex structure, though their specific impulse is higher than other types of thrusters. Liquid propellant thrusters also have difficulties owing to their structural complexity and sealing problems in their components, such as tanks and valves [7]. Hence, MEMS solid propellant thrusters, fabricated by MEMS technology, can be considered as the most suitable thrusters for minimized propulsion systems owing to their structural simplicity and to the absence of sealing problems [8]. The one-shot characteristic of MEMS solid propellant thrusters can be compensated by the fabrication of an array-type system [9,10]. A MEMS solid propellant thruster consists of a nozzle, a chamber, and an igniter as in a conventional thruster. For precise and stable control of satellites, the thruster must have a short

1

ignition delay and good reproducibility among the unit thrusters in an array. These factors are determined by the igniter; hence, the MEMS igniter is the most important component [11]. The igniter typically consists of a heater, which generates heat applied to the propellant, and a membrane, which protects the propellant from the surroundings and structurally supports the heater. MEMS igniters have been developed in two types according to the igniter material: silicon or glass. Silicon-based igniters have a very thin dielectric membrane owing to their fabrication process [12-15]. This thin membrane is advantageous in terms of short ignition delay, but there are some limitations in protecting the propellant from external shocks or temperature changes owing to the low structural stability of the thin membrane. Additionally, particular methods or equipment are required to fill the propellant into the chamber. Because the MEMS thruster must operate in harsh space environments, structural reliability of the membrane is as important as short ignition delay. A glass-based igniter was developed to improve the structural reliability of the membrane; however, the ignition delay in these cases was deteriorated owing to the thicker glass-ceramic membrane [16,17]. Moreover, reproducibility could not be evaluated because of the unevenness of the membrane surface and thickness. In this study, a noble-glass-based MEMS igniter for improved reproducibility and ignition delay with high structural stability is described. The igniter is designed to realize a uniform membrane thickness and a flat membrane surface for improved reproducibility. Furthermore, the proposed heater is designed at the under surface of the membrane for short ignition delay. Numerical simulations are performed to predict the proposed design’s performance. The igniter is then realized by a MEMS fabrication process using a glass wafer. A

2

firing test of the fabricated igniter is performed to measure its ignition delay and to evaluate reproducibility. Reproducibility is successfully evaluated by igniting five consecutive unit igniters at the same voltage. This reproducibility evaluation result is the first to be reported among research on MEMS igniters.

2. Method

2.1 Design and Numerical Simulation 2.1.1 Igniter Design For high structural stability of the MEMS igniter, the igniter material is selected as glass in this study. This is because the thickness of the membrane is directly related to structural stability, and the thickness can be freely adjusted in the fabrication process of glass. In a previous glass-based igniter called the glass-ceramic membrane igniter, the membrane was formed by a wet-etching process of the crystallized glass section (glass-ceramic) [16,17]. Its thickness was then adjusted using a time-monitored etch-stop technique. However, it was impossible to make the membrane thickness uniform using the wet-etching process because the etchant concentration varied with time within the chamber area. Figure 1 shows the surface profile of the fabricated glass-ceramic membrane measured by an alpha-step profiler. The membrane thickness was not uniform even though all igniters were simultaneously fabricated on the same glass substrate. The uniformity of the membrane thickness directly affected the reproducibility of the MEMS igniter because the ignition delay was measured by the fracture time of the membrane. Therefore, a different fabrication process is needed to improve

3

reproducibility. Additionally, the flatness of the membrane surface was poor in the glass-ceramic membrane igniter owing to the etching properties of the glass-ceramic, as shown in Figure 1. Good adhesion between the membrane surface and the propellant is directly related to a higher ignition success rate, and this higher ignition success rate is achieved with improved reproducibility. Hence, to realize uniform membrane thickness and a flat membrane surface, a MEMS igniter design in which the membrane is fabricated using a chemical-mechanical polishing (CMP) process instead of the wet-etching process is proposed in this study. In both dielectric membrane igniters and glass-ceramic membrane igniters, a heater is placed on the membrane as shown in Figure 2(a). Ignition heat generated by the heater is transferred to the propellant through the membrane in this type of the igniter. By this point, the heat loss caused by the membrane and the heat propagation time to the propellant are inevitable in the glass-ceramic membrane igniter owing to its thick membrane. These are the major causes of degraded ignition characteristics. To minimize this, it is better to transfer ignition heat directly to the propellant. Hence, the heater is placed under the glass membrane in our design, as shown in Figure 2(b). The thickness of the membrane is determined to be 40 μm in the glassceramic membrane igniter for the purpose of comparing ignition delays. Besides the MEMS igniter, a chamber with propellant must be fabricated for performance evaluation in a firing test. The chamber is automatically fabricated during the fabrication of the igniter in the cases of the glass-ceramic membrane igniter and the dielectric membrane igniter. However, in our design, the chamber is separately designed and fabricated to realize the proposed MEMS igniter. The fabricated chamber is then integrated with the igniter

4

for the firing test. An additional layer over the igniter is also needed in our design because an electrode for applying a voltage from the power source is placed under the membrane. Therefore, the upper layer in this study is designed to structurally reinforce the electrode region. Finally, the bottom layer is designed under the chamber layer to seal the propellant. To fabricate the igniter in an array format, a 3 × 3 igniter array with an upper layer, chamber layer, and bottom layer is designed as shown in Figure 3. 2.1.2 Numerical Simulation Numerical simulations of the designed MEMS igniter are conducted to predict its ignition characteristics. A commercial code, Fluent, is used for the computations. Figure 4 shows the results of igniter modeling with the meshes used in the calculations. The number of meshes generated in modeling is about four million. The meshes near the contact surface between the membrane and the propellant are denser than in other areas in order to obtain accurate results. As a boundary condition, convection to the atmosphere is applied to the top surface of the membrane and conduction to the glass substrate is applied to the sides and the bottom of the propellant. The propellant used in this study is lead styphnate, which is one of the most widely used initiating compounds. Additionally, lead styphnate, which is used in a glassceramic membrane igniter, is selected to be a propellant in this study for the same purpose [17]. To compare the ignition delay of two types of glass-based MEMS igniters, a glass-ceramic membrane igniter is also modelled using the same boundary conditions. A voltage is applied to the heater, and the temperature change on the top surface of the propellant is then calculated with time. The ignition area on the top surface of the propellant is designated in this calculation

5

to determine whether or not the propellant ignites. The radius of the ignition area is selected to be 0.1 mm as the measured hot-spot size of the glass-ceramic membrane igniter [16]. The ignition temperature of lead styphnate is 260°C; hence, we consider that the propellant ignites when the average temperature of the ignition area exceeds 260°C . In general, 15 V is used as the operating voltage on CubeSats. Because MEMS solid propellant thrusters are equipped on CubeSats, the calculation is performed at this voltage. Figure 5 shows the temperature contours on the top surface of the propellant with time at the input voltage of 15 V. The ignition temperature in the ignition area is reached after between 15 ms and 20 ms. Figure 6 shows temperature contours of the previous glass-ceramic membrane igniter at the input voltage of 15 V. The ignition temperature in the ignition area is reached at 17.9 ms. Fig. 7 shows a comparison of the cross-sectional temperature contours between the glass membrane igniter and the glass-ceramic igniter with time. It is verified that the temperature increase rate on the top surface of the propellant in the glass membrane igniter is higher than the glass-ceramic membrane igniter owing to the position of the heater. Figure 8 shows the ignition delays for both igniters as functions of the input voltage. The ignition delay is calculated from the time between applying the input voltage to the time when the propellant ignition temperature is exceeded in the ignition area. The ignition delay exponentially decreases in these terms, as in the results of previous MEMS igniters. The ignition delays of the glass membrane igniter and glass-ceramic membrane igniter at the input voltage of 10.6 V were 46.6 ms and 56 ms, respectively. The ignition delay of the glass membrane igniter is improved by 17% compared with that of the glass-ceramic membrane igniter. The improvement of the

6

ignition delay in the glass membrane igniter decreases as the input voltage increases. At 15 V, the ignition delays of the glass membrane igniter and the glass-ceramic membrane igniter were 16.8 ms and 17.9 ms, respectively. The ignition delay of the glass membrane igniter is improved by 6.2% compared with that of the glass-ceramic membrane igniter. As the input voltage increases, the heat generation rate of the heater increases according to Joule’s law. Therefore, the heat flow rate through the membrane in the glass-ceramic membrane igniter also increases. Because the ignition temperature of the propellant is constant, an increase of the heat flow rate reduces the delay effect of the membrane during ignition. As a result, the ignition delay difference between the glass membrane igniter and the glass-ceramic membrane igniter decreases at higher input voltages.

2.2 Fabrication of the MEMS Igniter 2.2.1 Fabrication of Each Layer The fabrication of the MEMS igniter is divided into two parts: the fabrication of each layer and the integration of each layer, including propellant filling. All components of the igniter are fabricated using a photosensitive glass wafer, which allows for the fabrication of microsystems with high aspect ratios [18]. First, the upper layer, chamber layer, and bottom layer are fabricated by anisotropic etching of photosensitive glass(Hoya optics). This process is shown in Figure 9(a). The photosensitive glass is diced into 30 × 30 × 1 mm parts. The glass is then selectively exposed to ultraviolet (UV) light with a wavelength of approximately 310 nm. Then, the exposed glass wafer is inserted into a programmable furnace for heat treatment. During heat treatment, the properties of the glass change owing to re-crystallization. After heat treatment, the

7

wafer is soaked in a 10% hydrofluoric acid (HF) solution and prepared for the etching process. The re-crystallized area is etched 20 times faster than the area unexposed to the solution. Because the unexposed area is partially etched, the surface of the wafer roughens. Consequently, the wafer is chemically and mechanically polished in order to make its surface flat. The thickness of the chamber finally reaches 0.5 mm. Second, the membrane layer together with the heater is fabricated through a lift-off process of the metal layer. This process is shown in Figure 9(b). A photoresist is spin-coated onto the glass wafer with dimensions 30 × 30 × 1 mm. A quartz mask is prepared on which a chromium absorber is patterned and selectively exposed to UV light with a wavelength of approximately 365 nm. The photoresist is then developed. A layer of titanium 200 Å thick is deposited with the use of a sputter in order to improve the adhesion between the platinum and the glass. A layer of Pt 2000 Å thick is deposited by the sputter onto the Ti layer. The wafer on which the photoresist and the Ti/Pt layer are deposited is soaked in acetone in order to remove the photoresist. Through this process, only the Ti/Pt pattern of the heater remains and, finally, the heater is formed. 2.2.2 Integration of Each Layer Before the membrane layer is polished to form the thick glass membrane, the membrane layer and the chamber layer are joined together using UV-curable glue, as shown in Figure 10(a). The membrane layer and the chamber layer are then aligned and exposed to UV light. Because the glue exposed to the air in the chamber area is not pressed during UV exposure, only the glue coated between two layers solidifies. Therefore, the glue in the chamber area is

8

dissolved when the two layers are soaked in acetone after UV exposure. The thickness of the membrane layer is then adjusted to be 40 μm by CMP as shown in Figure 10(b). The next step is to bond the upper layer using UV glue as shown in Figure 10(c). After the UV bonding process, the propellant is loaded into the chamber as shown in Figure 10(d). The lead styphnate used here is originally a powder. The propellant is dissolved in a solution that contains the solvent and binder before it is loaded into the chamber. The solution is injected into the chamber and dried to remove the solvent. The next step is to bond the bottom layer using the UV glue as shown in Figure 10(e). After exposure to UV light, the fabrication of the MEMS igniter array is completed. Figure 11 shows the fabricated MEMS igniter array before propellant loading.

3. Results 3.1 Experimental Setup The experimental setup used for the performance evaluation of the MEMS igniter is illustrated in Figure 12. The input voltage is applied by a power supply, and this voltage is measured by an oscilloscope. A dummy resistance is connected to the igniter, and the voltage across this resistance is also measured by the oscilloscope. The voltage applied to the igniter is calculated by subtracting the dummy voltage from the input voltage. The electric current flowing through the igniter is calculated by dividing the igniter voltage by the value of dummy resistance. The resistance of the igniter is calculated by dividing the igniter voltage by the current. A micro-positioner is used to accurately apply the voltage to the igniter, and an ohm meter is used to verify an electrical contact between the micro-positioner and the igniter. A high-

9

speed camera is used to capture the ignition of the igniter. All measured data and images are processed using a personal computer (PC).

3.2 Ignition Characteristics Results First, the ignition delay of the MEMS igniter was measured by a firing test at room temperature. Figure 13 shows the firing test of the igniter array. A plume generated by the combustion of the propellant was observed during the firing test. The ignition delay was measured as the input voltage increased from 10.59 V to 18.02 V. The ignition delay was measured from the time of applying the input voltage to the time when the voltage signal in the oscilloscope was blocked by the rupture of the membrane. The measured ignition delay of the fabricated MEMS igniter is shown in Figure 14. The ignition delay decreased exponentially as the input voltage increased. The measured minimum ignition delay was 15.4 ms at the input voltage of 18.02 V, and the maximum ignition delay was 30.46 ms at the input voltage of 10.59 V. Figure 14 also shows a comparison of the experimental results and the numerical simulation results. To determine whether or not the igniter ignited, the ignition area was set to be 0.0314 mm2 in the numerical simulation. This value was selected from the measured hot-spot size of the glass-ceramic igniter at the input voltage of about 15 V [16].

4. Discussion 4.1 Ignition Characteristics At the input voltage of 15.37 V, the ignition delays in the experiment and the

10

simulations were almost the same. This is because the real ignition area was the same as the ignition area used in the numerical simulations. Hence, the numerical simulation results in this study matched well with the experimental results when the exact ignition area was used in the calculations. The ignition delay in the experiment was 34.6% shorter than in the numerical simulation at the input voltage of 10.59 V. At the input voltage of 12.9 V, the ignition delay in the experiment was 18.4% shorter than in the numerical simulation results. This means that the real ignition area was smaller than the ignition area used in the numerical simulations at voltages lower than 15 V. At the input voltage of 18 V, the ignition delay in the experiment was 22.7% longer than in the numerical simulations. This also means that the real ignition area was wider than that in the numerical simulations at higher voltages than 15 V. Based on these results, it was verified that the ignition area of the igniter varies with the input voltage. A further analysis of the change of the ignition area according to the input voltage is needed to predict more accurate ignition characteristics of the MEMS igniter. Figure 15 shows the measured ignition delay of the glass-ceramic membrane igniter and the glass membrane igniter. At the input voltage of 10.83 V, the ignition delay of the glass-ceramic membrane igniter was 60.6 ms [17]. The ignition delay of the glass membrane igniter was 49.74% shorter than for the glass-ceramic membrane igniter at the input voltage of about 10.7 V. At the input voltage of about 15 V, which is the working voltage of CubeSats, the ignition delay of the glass ceramic membrane igniter was 27.5 ms [17]. The ignition delay of the glass membrane igniter was 18.02 ms at the input voltage of 15.372 V. The ignition delay of the glass membrane igniter was 34.45% shorter than for the glass ceramic membrane igniter. In the

11

numerical simulation results, the ignition delay of the glass membrane igniter was 6.2% shorter than for the glass ceramic membrane igniter. This is because the measured ignition delay of the glass ceramic membrane igniter from the test was much longer than the calculated ignition delay. This means that the non-uniform membrane thickness and uneven membrane surface of the glass ceramic membrane igniter caused deterioration of the ignition delay. Hence, the improvement of the ignition characteristics of the glass membrane igniter with a uniform membrane thickness and flat surface was verified from this result.

4.2 Reproducibility To evaluate reproducibility, five unit igniters were consecutively ignited at the input voltage of 15.37 V. Figure 16 shows the results of the reproducibility evaluation. All of the five unit igniters successfully ignited, and their ignition delay and ignition energy values were successfully measured. The ignition delay was measured by integrating the applied voltage to the igniter with respect to time. The measured average ignition delay and ignition energy were 17.08 ms and 25.6 mJ, respectively. The standard deviations of the ignition delay and ignition energy were calculated as 1.96 ms and 2.53 mJ, respectively. The coefficient of variation of the ignition characteristics was also calculated to be within 12%. In previous research on MEMS igniters, including glass-ceramic membrane igniters, the standard deviation and the coefficient of variation of the ignition characteristics were not reported owing to low reproducibility or low ignition success rates. The present results verified that the developed igniter achieves improved reproducibility.

12

5. Conclusion The design, performance calculation, fabrication, and test results of a glass membrane MEMS igniter were described in this study. A photosensitive glass wafer was used as the substrate for high structural stability. The igniter was designed to improve reproducibility and ignition delay compared with previous glass-ceramic membrane igniters. The main features of the designed glass membrane igniter are uniformity of membrane characteristics and heater position. Numerical simulations were performed to predict the ignition delay of the igniter. A performance comparison of the glass membrane igniter with a glass-ceramic membrane igniter was also conducted to evaluate the performance improvement of the glass membrane igniter. The designed igniter was then realized by a MEMS fabrication process using a glass wafer. After each layer was separately fabricated, the igniter was formed by integration of membrane layers and chamber layers. Lead styphnate was used as the solid propellant. In the firing test, the measured ignition delay of the glass membrane igniter was 30.46 ms at the input voltage of 10.59 V. The minimum ignition delay was 15.4 ms at the input voltage of 18.02 V. The measured ignition delay at the input voltage of 15.37 V was almost same as in the numerical simulations. However, the measured ignition delays at other input voltages did not match with the numerical simulation results because the ignition area used for the calculations was different from the real ignition area. The measured ignition delay was 34.45% shorter than the measured ignition delay of the glass ceramic membrane igniter at an input voltage of about 15 V. Considering that the operating voltage of CubeSats is 15 V, improvement of ignition delay for the glass membrane igniter was verified from the firing test. Reproducibility was also evaluated by consecutively igniting five igniters at the input voltage of 15 V. All igniters successfully

13

ignited, and the calculated coefficient of variation of the ignition characteristics was 12%. This value has not previously been reported in research on micro-igniters; hence, the improvement of reproducibility in the proposed glass membrane igniter was verified.

Acknowledgments

This research was supported by Space Core Technology Development Program through the National Research Foundation of Korea (NRF) funded by the Ministry of Science, ICT and Future Planning (2014M1A3A3A02034813).

14

References

[1] Heidt, H., Puig-Suari, J., Moore, A. S., Nakasuka, S., Twiggs, R. J. 2001 CubeSat: A New generation of picosatellite for education and industry low-cost space experimentation AIAA Paper SSC00-V-5 [2] Toorian, A., Blundell, B., Suari, J. P., Twiggs, R. 2005 CubeSats as responsive satellites AIAA Paper 2005-6828 [3] Herrero, F.A., DiJoseph, M., Moore, T. E., Slavin, J.A., Vondrak, R., 2000 NanoSat constellations for geospace science AIAA Paper 2000-5174 [4] Esper, J., Neeck, S., Slavin, J.A., Leitner, J., Wiscombe, W., Bauer, F.H. 2003 Nano/micro satellite constellations for earth and space science Acta Astronautica 52 785–791 [5] Cass, S. 2001 MEMS in space IEEE Spectrum 38 56–61 [6] de Groot, W.A. 1998 Propulsion options for primary thrust and attitude control of microspacecraft NASA/CR-1998-206608 [7] Rossi, C. 2002 Micropropulsion for space - A survey of MEMS-based micro thrusters and their solid propellant technology Sensors Updates 10 257–292 [8] Zhang, K.L., Chou, S.K., Ang, S.S., Tang, X.S. 2005 A MEMS-based solid propellant microthruster with Au/Ti igniter Sensors and Actuators A 122 113–123 [9] Larangot, B., Rossi, C., Camps, R., Berthold, A., Pham, P.Q., Briand, D., de Rooij, N.F.,

15

Puig-Vidal, M., Miribel, P., Montane, E., Lopez, E., Samitier, J. 2002 Solid propellant microrockets—towards a new type of power MEMS NanoTech 2002—At the Edge of Revolution, Houston, Texas, USA, September 9–12 [10] Seo, D., Lee, J., Kwon, S. 2012 The development of the micro-solid propellant thruster array with improved repeatability Journal of Micromechanics and Microengineering, 22 094004 [11] Briand, D., Pham, P.Q., de Rooij, N.F. 2007 Reliability of freestanding polysilicon microheaters to be used as igniters in solid propellant microthrusters Sensors and Actuators A 135 329–336 [12] Rossi, C., Temple-Boyer, P., Esteve, D. 1998 Realization and performance of thin SiO2/SiNx membrane for microheater applications Sensors and Actuators A 64 241–245 [13] Rossi, C., Larangot, C., Lagrange, D., Chaalane, A. 2005 Final characterization of MEMSbased pyrotechnical microthrusters Sensors and Actuators A 121 508–514 [14] Rossi, C., Briand, D., Dumonteuil, M., Camps, T., Pham, P.Q., de Rooij, N.F. 2006 Matrix of 10x10 addressed solid propellant microthursters: review of the technologies Sensor and Actuators A 126 241–252 [15] Tanaka, S., Kondo, K., Habu, H., Itoh, A., Watanabe, M., Hori, K., Esashi, M. 2008 B/Ti multilayer reactive igniters for a micro solid propellant array thruster Sensors and Actuators A 144 361–366

16

[16] Lee, J., Kim, K., Kwon, S. 2010 Design, fabrication, and testing of MEMS solid propellant thruster array chip on glass wafer Sensor and Actuators A 157 126–134 [17] Lee, J., Kim, T. 2013 MEMS solid propellant thruster array with micro membrane igniter Sensors and Actuators A 190 52–60 [18] Dietrich, T.R., Ehrfeld, W., Lacher, M., Krämer, M., Speit, B. 1996 Fabrication technologies for Microsystems utilizing photoetchable glass Microelectronic Engineering 30 497–504

Photographs and Biographies 1. Daeban Seo

Daeban Seo received the B.S. degree from Pusan National University, Busan, Korean, in 2009 and the M.S. and Ph.D. degrees in Aerospace Engineering from Korea Advanced Institute of Science and Technology (KAIST), Daejeon, Korea, in 2011 and 2014, respectively. In 2013, he joined the Korea Aerospace Research Institute, Daejeon, Korea, where he is currently a senior researcher. His research interest include a propulsion system such as liquid rocket engine, solid rocket engine and micro thruster.

17

3. Juyoung Jeong

Juyeong Jeong received the B.S. and M.S. degree from Hanbat National University, Daejeon, Korea, in 2014 and 2016, respectively. His research interests include micro thruster and CFD. In 2016, he joined the Korea Railroad Research Institute, Gyeonggido, Korea, where he is currently a post-master. 3. Tae-Gyu Kim

Taegyu Kim received the B.S. degree from Korea Aerospace University (KAU), Goyang, Korea, in 2003 and the M.S. and Ph.D. degrees in Aerospace Engineering from Korea Advanced Institute of Science and Technology (KAIST), Daejeon, Korea, in 2005 and 2008, respectively. In 2009, he joined the Department of Aerospace Engineering at Chosun University, Gwangju, Korea, where he is currently an Associate Professor. His research interests include various kinds of micro energy conversion

18

systems, such as micro fuel cell, micro reformer and micro plasma reactor.

4. Jongkwang Lee

Jongkwang Lee received the B.S. degree from Korea Aerospace University (KAU), Goyang, Korea, in 2002 and the M.S. and Ph.D. degrees in Aerospace Engineering from Korea Advanced Institute of Science and Technology (KAIST), Daejeon, Korea, in 2004 and 2009, respectively. From 2009 to 2012, he worked in the LCD R&D Center, Samsung Electronics, Yongin, Korea, as a Senior Researcher. In 2012, he joined the Department of Mechanical Engineering at Hanbat National University, Daejeon, Korea, where he is currently an Associate Professor. His current research focuses micropropulsion devices, micro-fluidics, and CFD.

19

Membrane area

500

Top

Thickness (m)

400

300

Chamber layer 200

Propellant area

100

Bottom 0

0

1000

2000

3000

4000

5000

6000

Length (m) Figure 1 Surface profile of a glass-ceramic membrane

20

Propellant chamber (thickness : 0.5 mm)

Glass-ceramic membrane (thickness : 35 μm)

Pt/Ti heater (2200 Å)

1 mm

Propellant (Lead styphnate)

(a) Glass membrane (thickness : 40 μm)

Propellant chamber Pt/Ti heater

Propellant

(b)

Figure 2 (a) Glass-ceramic membrane igniter; (b) glass membrane igniter

21

3 x 3 unit igniters

30 mm

Upper layer Membrane layer

Heaters and electrodes

Chamber layer

Bottom layer 30 mm

Figure 3 Schematic of the designed 3 × 3 MEMS igniter array

22

Convection (h = 5 W/m2K)

Conduction

Conduction Figure 4 Modeling of the MEMS igniter for simulation

23

Temperature (K) 600

500

5 ms

10 ms

15 ms

20 ms

400

300

Figure 5 Temperature contours of the propellant surface with time at 15 V (glass membrane igniter)

24

Temperature (K) 600

500

5 ms

10 ms

15 ms

20 ms

400

300

Figure 6 Temperature contours of the propellant surface with time at 15 V (glass-ceramic membrane igniter)

25

Temperature (K) 600 Glass membrane

Glass-ceramic membrane

Glass membrane

Glass-ceramic membrane

500 5 ms

10 ms

400 Glass membrane

300

Glass-ceramic membrane

15 ms

Glass membrane

Glass-ceramic membrane

20 ms

Figure 7 Cross-sectional temperature of the glass membrane igniter and glass-ceramic membrane igniter with time

26

60

Ignition delay (ms)

Glass memb. (CFD) Glass-ceramic memb. (CFD) 50

40

30

20

10

12

14

16

18

Voltage (V) Figure 8 Comparison of ignition delay between the glass membrane igniter and glass-ceramic membrane igniter

27

Qz wafer Cr Preparation of Qz mask wafer

Starting wafer

UV

PR patterning 310 nm UV exposure

Heat treatment & polishing

Pt/Ti deposition (Pt : 0.2micro) Micro heater

Lift off

HF etching

(a)

(b)

Figure 9 Fabrication process of (a) upper chamber and bottom layer; (b) membrane layer with a Pt/Ti heater

28

(d) Propellant loading

(a) UV bonding

(b) CMP process (e) Bottom layer bonding

(c) Upper layer bonding

(f) Micro igniter

Figure 10 Integration process of the MEMS igniter

29

30 mm

1 mm

Pt/Ti heater (2200 Å)

Propellant chamber Glass membrane (thickness : 0.5 mm) (thickness : 40 μm)

Figure 11 Fabricated MEMS igniter

30

Data processing

Oscilloscope

Power supply Voltage probe

Voltage probe

Dummy resistor Micropositioner

MEMS igniter Ohm meter

Figure 12 Experimental setup

31

Micro-positioner

Igniter

Figure 13 Firing test of the glass membrane igniter

32

Ignition delay (ms)

50 Experiment Numerical simulation

40

30

20

10 10

12

14

16

Voltage (V) Figure 14 Ignition delay with input voltage

33

18

Ignition delay (ms)

100 Glass-ceramic memb. (EXP.) Glass memb. (EXP.)

80

60

40

20 10

12

14

16

18

Voltage (V) Figure 15 Comparison of ignition delay between the glass-ceramic membrane igniter and glass membrane igniter

34

40 Ignition delay Ignition energy

30

30

20

20

10

10

0

1

2

3

4

5

Igniter sample number Figure 16 Results of reproducibility evaluation

35

0

Ignition energy (mJ)

Ignition delay (ms)

40