Cryogenic performance evaluation of commercial SP4T microelectromechanical switch for quantum computing applications - Nature

IntroductionQuantum computers have garnered significant attention in recent years due to their potential to solve certain problems with remarkable speed and efficiency by leveraging quantum phenomena such as entanglement, superposition, and interference1. Unlike classical computers, which rely on binary bits as the fundamental unit of computation, quantum computers utilize quantum bits, or qubits, capable of representing multiple states simultaneously. The implementation of qubits can be realized through various approaches, including superconducting circuits2, trapped ions3, photon polarization4, and electron spins5. Among these, superconducting quantum computers have emerged as the leading platform due to their high scalability, rapid computational speed, and compatibility with semiconductor technologies6. Superconducting quantum computers generally consist of a quantum processor and control/read-out electronics as shown in Fig. 1a. Since quantum processors are highly sensitive to external environmental factors such as heat and electromagnetic fields, they are typically located at the 10 millikelvin (mK) base temperature stage of a dilution refrigerator, and connected to room-temperature control and read-out electronics through wide-band cables7. While this architecture is effective for experimental verification with a small number of qubits, scaling to the millions of qubits required for practical quantum applications8,9 presents significant interconnect challenges due to the limited mK temperature stage area and cooling power of the dilution refrigerator. To address these challenges, new architectures utilizing cryogenic multiplexers have been actively researched to prepare the large-scale quantum computing system10,11 (Fig. 1b). In this architecture, DC and RF signals from the room-temperature electronics are transmitted to the mK temperature stage, where the quantum processor is located, using only a small number of cables. At the mK temperature stage, these signals are distributed to multiple qubits or quantum information is collected from multiple qubits through the multiplexers. Since it enables effective communication between the room-temperature electronics and the quantum processor at 10 mK using a minimal number of wires, it is considered a next-generation interconnect architecture for realizing large-scale quantum computers with more than 1 million qubits. However, in order to implement reliable large-scale quantum computing with multiplexers, the switching device for the multiplexer must satisfy several quantitative requirements10. First, they should operate stably without performance degradation even at cryogenic temperatures, since the multiplexer is placed on the 10 mK stage of a dilution refrigerator. Second, their power consumption must remain below the cooling budget of the dilution refrigerator (approximately 20 μW), as superconducting qubits are highly sensitive to thermal electromagnetic radiation. Third, the devices should exhibit low insertion loss ( 30 dB) to suppress inter-channel leakage, thereby ensuring crosstalk-induced infidelity below 0.1% within the 4–8 GHz qubit frequency range. Fourth, the devices should achieve fast switching times shorter than 2 μs to enable control and readout time-division multiplexing. Various research groups have proposed cryogenic switching devices for implementing cryogenic multiplexers in scalable quantum computing systems. Superconducting nanowire-based switch12 has demonstrated ultra-fast rise and fall times along with extremely low power dissipation (~ 15 nW). However, the structure of the device was single-pole single-throw (SPST) configuration, which offers poor scalability, and exhibits low isolation performance (~ 10 dB), making them less suitable for high-fidelity quantum applications. Another study based on high-electron-mobility transistors (HEMTs)13 has demonstrated the single-pole four-throw (SP4T) configuration and 3D stackable integration architectures. Nevertheless, this approach exhibited high insertion loss (approximately 5 dB) and utilizes non-CMOS compatible materials such as InGaAs, which hinders large-scale manufacturability. Josephson junction (JJ)-based switches14 have also been explored, offering the key advantage of zero internal power dissipation. However, they also rely on SPST configurations and exhibit limited on/off ratios ( ~ 20 dB). Recently, new strategies for cryogenic multiplexer and de-multiplexer circuits have been proposed using InAs nanowires15. However, they still rely on non-CMOS-compatible materials, including InAs and GaAs substrates. Therefore, in order to realize large-scale quantum computing, many recent studies have focused on developing Cryo-CMOS based cryogenic multiplexers10,11,16 using scalable and mature semiconductor fabrication processes. These works have predominantly adopted a single-pole four-throw (SP4T) architecture. However, current cryo-CMOS multiplexers still exhibited relatively high insertion losses in the range of 1.6–2.5 dB. Furthermore, there have been no reports on the long-term reliability results under cryogenic conditions. Microelectromechanical system (MEMS) switches have recently emerged as a promising alternative for implementing large-scale quantum computing. Similar to semiconductor devices, MEMS switches consist of source, drain, gate, and ground electrodes17. When a voltage exceeding the operating threshold is applied between the gate and ground electrodes, the source beam is mechanically deflected to contact the drain electrode, allowing current to flow. This operating principle is known as electrostatic actuation. Thanks to the mechanical movement with electrostatic actuation, MEMS switches can operate reliably at cryogenic temperatures without dopant-related challenges18,19. In addition, The MEMS switch provides excellent port-to-port isolation20 and near-zero static power consumption21 in the off-state, as the top and bottom electrodes are physically separated by an air gap. In the on- state, metallic contact yields a very low on-state resistance, resulting in low insertion loss. These features make MEMS switches a promising alternative for quantum applications. However, large-scale quantum computing systems with more than 1 million qubits would require a substantial number of reliable MEMS switches. Therefore, the use of commercially manufactured MEMS switches, which ensure high yield and consistent quality, would be an ideal solution for future quantum computing systems. In this paper, the DC and RF characteristics of commercial RF MEMS switches are evaluated at sub-10 K temperatures, and their applicability to quantum computing systems is explored. To investigate the cryogenic characteristics of the commercial MEMS switches, finite element method (FEM) simulations and experimental measurements of their DC and RF performance at cryogenic temperature are conducted. The results revealed that the MEMS switches exhibit a lower operating voltage, lower on-resistance, and improved RF performance compared to their room temperature results. Additionally, the WLCSP-packaged MEMS switches exhibited a bouncing phenomenon due to a quasi-vacuum condition inside the package at the cryogenic temperature. This issue was effectively mitigated by employing an engineered pulse waveform that minimized the cantilever’s contact velocity. Consequently, the MEMS switches exhibit high reliability over 100 million cycles and stable SP4T operation characteristics, showcasing their potential role in quantum computing systems. Furthermore, it is demonstrated that logical operations, such as NAND and NOR gates can be performed using the MEMS switches by leveraging its unique operating mechanisms.Fig. 1: Commercial microelectromechanical system (MEMS) switches for large-scale quantum computing systems.a Layered architectures of current superconducting quantum computing systems. b Future superconducting quantum computing systems with cryogenic multiplexers. c Optical photograph of the SP4T device comprising MEMS switches developed by Menlo Microsystems, Inc. d Block diagram of the SP4T MEMS deviceFull size imageResultsCommercial RF MEMS switch for large-scale quantum computing systemsFigure 1c, d presents an optical image and block diagram of the single-pole four-throw (SP4T) MEMS device developed by Menlo Microsystems, Inc. The device is configured as a SP4T device with center input, which can be routed to one of the four outputs by applying a voltage to the corresponding gate electrode. Each individual MEMS switch features a cantilever structure22. In the absence of an operating voltage at the gate, the cantilever beam remains physically separated from the output signal line. When a gate voltage is applied, the cantilever beam is mechanically deflected and contacts the output signal line, enabling signal transmission. This mechanical operation results in ultra-low insertion loss, high isolation, and superior linearity. Furthermore, the MEMS switch is hermetically sealed through wafer level chip scale package (WLCSP) technology, enhancing its operational reliability. Therefore, the developed MEMS switch is widely utilized in various applications, including wireless communication, defense/aerospace, and automatic test equipment (ATE) and test instrumentation. In this study, the MEMS switch’s properties at cryogenic temperatures are characterized to evaluate its potential applicability to quantum computing systems.DC and RF performance evaluation at cryogenic temperatureFirst, the electrical characteristics of the MEMS switch, a fundamental component of the SP4T device shown in Fig. 1d, were assessed through FEM simulation. Air gap changes due to deflection significantly impact the performance of the MEMS switch23, so deflection as a function of temperature was also simulated (Fig. 2a). The cantilever switch exhibited minimal z-direction displacement with temperature changes, as it can freely expand in length, unlike the fixed-fixed beam structure. The maximum displacement of approximately 60 nm occurred when the MEMS switch was exposed to the cryogenic temperatures. This result indicates that the MEMS switch can maintain consistent operating characteristics even at cryogenic temperatures since the change in air gap was insignificant. Indeed, as shown in Fig. 2b, the operating voltage decreased slightly by approximately 3.5% at 0 K compared to the room-temperature result. The robust cryogenic performance of the MEMS switch, predicted through FEM simulations, was further confirmed experimentally. All cryogenic experiments were conducted at approximately 5.8 K. Figure 2c presents the pull-in voltages of 16 MEMS switches measured at room temperature (left), at cryogenic temperature (middle), and after returning to room temperature (right). In alignment with the simulation results, the operating voltage at 5.8 K exhibited a slight reduction of approximately 3.1% due to the small decrease in the air gap distance. Notably, this result is significant compared to many conventional MEMS switches, which experienced substantial shifts in pull-in voltage at cryogenic temperatures24,25,26,27,28. Furthermore, the pull-in voltages returned to their original value after the MEMS devices were brought back to room temperature. Figure 2d shows the measured on-resistance (the resistance between the two interfaces during mechanical contact) at both room temperature and cryogenic temperature. The on-resistance decreased by approximately 15.3% at cryogenic temperatures. The decreased on-resistance is attributed to the diminished phonon scattering in metals at lower temperatures29. Furthermore, the decreased on-resistances were fully recovered to their original values after returning to room temperature. The RF characteristics, including insertion loss, isolation, and return loss were also evaluated at the cryogenic temperature. The RF performance of the SP4T MEMS switch can be estimated using the transmission-line model provided by Menlo Microsystems, inc. provided in Supplementary Fig. 1. We experimentally measured the insertion loss at both room and cryogenic temperatures, as shown in Fig. 2e. The RF measurement was conducted between output_1 and output_3 (or output_2 and output_4) of the SP4T device as shown in Fig. 1d. At both temperatures, the insertion loss remains below 0.5 dB within the qubit frequency range of 4–8 GHz owing to the low on-resistance of the MEMS switch. Notably, the values measured at cryogenic temperature were slightly improved compared to those at room temperature due to the effect of lower resistivity at low temperatures. Meanwhile, in both cases, sharp dips are observed between 3 and 4 GHz. These notches are artifacts of the cryogenic probe station which could not be completely calibrated out and are not related to the performance of the MEMS switch itself, as such features were not present in other room-temperature probe station measurements. The isolation characteristics were also measured as shown in Fig. 2f. The MEMS switch exhibited excellent isolation performance, exceeding 35 dB, due to the presence of an air gap in the off- state. Importantly, the MEMS switch maintained exceptional isolation performance at cryogenic temperatures, as the airgap distance remained consistent even at the cryogenic temperature. The return loss was also measured. Supplementary Fig. 2a and 2b show the return loss when the gate voltage is applied (on-state) and not applied (off-state), respectively. In the on-state, the switch shows excellent return loss better than 20 dB at both room temperature and cryogenic temperature. Additionally, the switch presents a nearly ideal reflection in the off-state with a return loss below 0.5 dB. In conclusion, the simulation and measurement results support that the MEMS switch maintains its outstanding performance without degradation at the cryogenic temperatures, making it a promising candidate for quantum computing applications.Fig. 2: Performance evaluation of MEMS switches at cryogenic temperature.a Simulated deflection as a function of the temperature. b Simulated pull-in voltage as a function of the temperature. c Pull-in voltage measured sequentially at room temperature (left), at cryogenic temperature (middle), and room temperature again (right). d On-resistance measured sequentially at room temperature (left), at cryogenic temperature (middle), and room temperature again (right). e Measured insertion loss at both room and cryogenic temperatures. The insertion loss remains below 0.5 dB within the highlighted qubit frequency range of 4–8 GHz. f Measured isolation at both room and cryogenic temperaturesFull size imageDynamic Response of the MEMS SwitchThe transient response of the MEMS switch was also measured through a voltage divider circuit to evaluate the switching speed and dynamic characteristics as shown in Fig. 3. The voltage divider circuit is shown in Supplementary Fig. 3a. This circuit consists of a MEMS switch connected in series with a load resistor. When the MEMS switch is in the off-state, the voltage from the power supply is entirely applied across the MEMS switch due to its high off-resistance. In contrast, when the MEMS switch is turned-on by applying a gate pulse, the voltage from the power supply is distributed to the load resistor. Figure 3a, b shows the transient response when 10 kHz gate pulses were applied to the MEMS switch. Due to mechanical movement of the cantilever, contact occurred approximately 2.7 μs after the gate pulse was applied to the device, which is the switching speed. However, at cryogenic temperature, the transient response exhibits a different behavior, as shown in Fig. 3c, d. While the gate voltage is off, severe oscillations in the output voltage were observed. This behavior can be attributed to the WLCSP packaging of the MEMS switch. At cryogenic temperatures, the gas inside the package undergoes a phase transition, creating a quasi-vacuum environment. The absence of air damping in the cryogenic environment can lead to a pronounced bouncing phenomenon. Supplementary Fig. 3b shows the dynamic response over an extended period, illustrating that the bouncing persisted for approximately 150 μs. Meanwhile, the measured transient response reveals that the output voltage briefly exceeds the 1 V supply level when the gate voltage is turned-off. This overshoot is primarily attributed to the parasitic inductance inherently present in the system. Because all physical conductors in the circuit exhibit inductance, the abrupt interruption of the current upon switch turn-off releases the energy stored in these inductive paths, generating a back electromotive force (EMF). This back-EMF temporarily opposes the rapid decrease in current and results in a transient voltage overshoot at the output node, momentarily raising it above the supply voltage. To determine whether the bouncing phenomenon is temperature-dependent, the dynamic response was measured as the temperature decreased from 100 K to 10 K. As shown in Supplementary Fig. 4, the bouncing phenomenon began near the boiling points of oxygen and nitrogen. These measurement results provide clear evidence that the bouncing effect was caused by a phase transition of the gases inside the packaging. This phenomenon at the cryogenic temperature can present a significant challenge to the stable dynamic operation of the MEMS switch in the cryogenic application. To address this issue, an engineered waveform employing a dual-pulse approach was introduced, as shown in Fig. 3e–f, to ensure stable operation even at cryogenic temperatures. This method tailors the actuation waveforms to minimize the velocity of the cantilever upon contacting the bottom electrode, thereby effectively suppressing bouncing30,31,32. The introduced engineered waveform consists of four distinct regions, departing from the typical square pulse shown in Fig. 3b. A detailed design process is illustrated in Supplementary Fig. 5. The key concept of the engineered waveform is to ensure that the switch contacts the bottom electrode with near-zero velocity. As shown in Supplementary Fig. 5a, a voltage higher than the pull-in voltage (\({V}_{a}\)) is applied for a duration \({t}_{1}\) to provide sufficient momentum for the switch to reach the bottom electrode. This is immediately followed by a voltage lower than the pull-in voltage (\({V}_{l})\) applied for a duration \({t}_{2}\), which allows the cantilever to approach the bottom electrode with near-zero velocity. In our implementation, \({V}_{a}=\) 90 V and \({V}_{l}=\) 55 V were used, and the durations \({t}_{1}\approx\) 2.1 μs and \({t}_{2}\approx\) 1.1 μs were calculated using the equations (1)–(4) in Supplementary Fig. 5. A detailed illustration of the waveform width and period is presented in Supplementary Fig. 5c. Consequently, in the red region of Fig. 3f, a voltage of 90 V (higher than the pull-in voltage) is applied for 2 μs to provide sufficient momentum for the cantilever to contact the bottom electrode. In the subsequent orange region, a voltage of 55 V (lower than the pull-in voltage) is applied for 1 μs, enabling contact with near-zero velocity. Immediately after contact, the holding voltage of 90 V is applied to reduce oscillation during the contact phase. The same strategy is employed during the release process to suppress bouncing when the cantilever detaches. In the blue region, a voltage of 80 V (higher than the pull-in voltage) is applied for a duration of 2 μs, immediately following the green region, where 0 V is applied for 1 μs. This sequence can bring the velocity of the cantilever to zero, enabling smooth detachment without the occurrence of bouncing. While this engineered waveform slightly increased the switching speed (approximately 3.3 μs), it can ensure stable dynamic operation even at the cryogenic temperature. The total power consumption of a single MEMS switch consists of both static and dynamic components. The static power consumption refers to the power consumed when the gate voltage is 0 V. In the case of MEMS switches, the static power consumption is effectively zero due to their electrical isolation of the gate electrode. The dynamic switching energy of a single MEMS switch is calculated as follows: \({E}=\frac{1}{2}C{V}^{2}=\frac{1}{2}\times 15{\rm{fF}}\times {(90{\rm{V}})}^{2}=\,60.7{\rm{pJ}}\). Notably, this represents the energy to charge and discharge the gate electrode, and is predominantly dissipated in the driver and wiring located outside the 10 mK stage. Therefore, it does not significantly impact the cooling budget of the 10 mK substrate. Consequently, MEMS multiplexers offer significant advantages for quantum computing applications, primarily due to 1) zero static power consumption, 2) reliable operation under cryogenic temperature, and 3) excellent RF characteristics.Fig. 3: Dynamic response of MEMS switches.a, b Measured dynamic response and magnified view at room temperature. c, d Measured dynamic response and magnified view at cryogenic temperature, showing a severe bouncing phenomenon. e, f Measured dynamic response and magnified view at cryogenic temperature when the engineered gate waveform was introducedFull size imageLifetime and logical operation of the MEMS switchUsing the engineered waveform, reliability tests are conducted at cryogenic temperature. Figure 4 presents the measurement results of the dynamic response after 103, 106, 107, and 108 cycles. This is consistent with a previous study showing that cryogenic temperatures under the boiling temperature of the oxygen enhance the reliability of the MEMS switch by suppressing native oxide formation at the contact interface18. Furthermore, the cantilever structure of the MEMS switch experienced minimal stress, as it can freely deform as the temperature decreases. This structural advantage indicates that the MEMS switch can maintain operational stability even at cryogenic temperatures. Consequently, the MEMS device demonstrated stable operation with no observable performance degradation after exceeding 100 million cycles. The overlapping graphs corresponding to these measurements are provided in Supplementary Fig. 6a. The on-resistance values were also extracted from the dynamic response measurements (Supplementary Fig. 6b). The on-resistance remained consistent throughout repeated operation, indicating excellent operational reliability. These results show that the operating characteristics of the MEMS switch remain constant for 100 million cycles. Next, the operational stability of the SP4T device composed of MEMS switches under cryogenic temperature is examined. The SP4T MEMS device comprises one input and four outputs, allowing the input signal to be routed to a desired output by applying a gate voltage to the corresponding gate electrode. Figure 5a–c show the measurement results for routing the input signal to output_1 when the operating voltage is applied to the gate_1 electrode. Without the gate voltage, the input signal was not transmitted to the output_1 electrode. However, upon applying the gate voltage, the input signal was successfully measured at the output_1 electrode. Similarly, the input signal was routed to output_2 when the operating voltage was applied to the gate_2 electrode (Fig. 5d–f). The measurement results with the signal routed to output_3 and output_4 are also provided in Supplementary Fig. 7. These results indicate that the SP4T MEMS device reliably routes the input signals to the desired output electrode even at cryogenic temperatures, demonstrating its potential for quantum applications. Additionally, NAND and NOR logic operations using the MEMS switches are demonstrated at the cryogenic temperature. Figure 6a–c show the measurement results of the NAND operation at approximately 5.8 K utilizing the interconnect structure of the SP4T device. The NAND operation was implemented by configuring a circuit where one load resistor and two switches are connected in series, as illustrated in Fig. 6a. In this setup, the two inputs correspond to the gate voltages of the two switches, and the output is the voltage across the MEMS switches connected in series. Utilizing the design of the SP4T MEMS device, which features four ports connected to a central input, the NAND operation can be performed as shown in Fig. 6b. Figure 6c shows the measured output signal of the NAND operation when 10 kHz input signals were applied to both gate electrodes. When neither Vinput_1 nor Vinput_2 was applied, or when only one input was applied, Voutput was measured as 1 because the resistance of the MEMS switches connected in series was significantly higher than the load resistor. Conversely, when both Vinput_1 and Vinput_2 were applied, Voutput was measured as 0 because the resistance of the MEMS switches connected in series became small compared to the load resistor. For the NOR operation, the circuit was configured by connecting a load resistor in series with MEMS switches arranged in parallel, as shown in Fig. 6d. To implement the NOR operation using the structure of the SP4T MEMS device, a circuit was configured as shown in Fig. 6e. Figure 6f shows the measured output signal of the NOR operation when 10 kHz input signals were applied to both gate electrodes at 5.8 K. When neither Vinput_1 nor Vinput_2 was applied, Voutput was measured as 1 because the resistance of the MEMS switches connected in parallel was significantly higher than the load resistor. However, if either Vinput_1 or Vinput_2 was applied, Voutput was measured as 0 because the resistance of the MEMS switches connected in parallel became negligible compared to the load resistor (Fig. 6f). These results indicate that the MEMS switch can successfully perform logical operations even at extremely low temperatures, highlighting its potential for advanced cryogenic applications. While the MEMS switches show excellent performance at cryogenic temperatures through FEM simulations and various measurements, certain challenges remain to be addressed. Our MEMS switch showed an on-switching time of 3.3 μs. In general, time-division multiplexing (TDM) in quantum systems can be categorized into two types—readout multiplexing and control multiplexing—each with distinct switching speed requirements. Readout multiplexing typically requires switching rates in the range of 0.25–1.00 MHz, corresponding to pulse durations of 0.5–2.0 μs, whereas control multiplexing demands much higher switching rates (10–50 MHz) with shorter pulses (10–50 ns)10. Based on these requirements, our SP4T MEMS multiplexer is well suited for readout multiplexing in cryogenic quantum systems. However, for potential use in control multiplexing, further improvements are required. Another critical issue is dielectric charging. Since the MEMS switch operates through the electrostatic actuation, dielectric charging can occur with repeated operation. This issue is particularly pronounced at cryogenic temperatures, where charge carriers lack sufficient thermal energy to escape the traps within the dielectric33. As a result, the dielectric charging phenomenon becomes more severe, potentially leading to the stiction phenomenon. Indeed, stiction is observed in some devices when they were repeatedly operated at high frequencies exceeding 100 kHz at the cryogenic temperature between 5 K and 10 K. Upon returning to room temperature, some devices resumed normal operation, indicating that the stiction was caused by dielectric charging. Given that high-speed operation is critical for quantum computing applications, further research into materials is required to mitigate dielectric charging and its effects under cryogenic conditions. Alternatively, FEM-based reliability prediction studies may also be necessary22.Fig. 4: Repetitive operation test at cryogenic temperature.a–d Measured dynamic response results at a gate frequency of 10 kHz after 103, 106, 107, and 108 cycles, respectivelyFull size imageFig. 5: SP4T MEMS device operation at cryogenic temperatures.a Applied gate signal (Vgate_1) to the SP4T MEMS device. b Applied input signal (Vinput) to the center input electrode. c Measured output signal (Voutput_1) as function of the applied gate signal (Vgate_1). d Applied gate signal (Vgate_2) to the SP4T MEMS device. e Applied input signal (Vinput) to the center input electrode. f Measured output signal (Voutput_2) as function of the applied gate signal (Vgate_2)Full size imageFig. 6: Logical operations using SP4T MEMS device at cryogenic temperature.a Voltage divider circuit diagram utilizing MEMS switches for NAND operation. b Circuit configuration of the SP4T MEMS device for NAND operation. c Measured output response (Voutput) as a function of input voltages (Vinput_1 and Vinput_2). d Voltage divider circuit diagram utilizing MEMS switches for NOR operation. e Circuit configuration of the SP4T MEMS device for NOR operation. f Measured output response (Voutput) as a function of input voltages (Vinput_1 and Vinput_2)Full size imageDiscussionThis study demonstrates the suitability of commercial MEMS switch as key components for cryogenic multiplexers in large-scale quantum computing systems. The MEMS switches demonstrated stable operation at cryogenic temperature (approximately 5.8 K), exhibiting performance improvements in DC and RF characteristics compared to their room-temperature results. Furthermore, by introducing an engineered waveform, stable dynamic operation was achieved by effectively mitigating the undesired bouncing phenomenon at the cryogenic temperature. Notably, the MEMS switches performed reliably over 100 million cycles at cryogenic temperature. The successful operation of the SP4T MEMS device, along with the demonstration of logical operations such as NAND and NOR gates, underscores the suitability of MEMS switches for complex interconnect architectures in quantum systems. However, challenges such as dielectric charging-induced stiction at high frequencies persist, necessitating further advancements in materials and design optimization for high-speed and long-term reliability. Our findings support the MEMS switches as a promising candidate solution for scalable, reliable quantum systems, paving the way for the realization of large-scale quantum computing with millions of qubits.Materials and methodsFEM simulationFinite Element Method (FEM) simulations were conducted using ANSYS Mechanical to analyze the mechanical stress, beam deflection, and pull-in voltage of the MEMS switch under varying temperature conditions.Electrical measurementsAll electrical measurements were conducted using a cryogenic vacuum probe station (Lake Shore Cryotronics, CRX-VF). To determine the pull-in voltage and on-resistance between the top and bottom electrodes, a sourcemeter (Keithley 2450) was employed, configured with a reading voltage of 1 V and a compliance current limit of 10 mA. A network analyzer (Agilent N5225A) was used to characterize the insertion loss and isolation properties of the MEMS switches. For dynamic response measurements, a gate pulse generated by a function generator (Agilent 33220 A) was amplified using a voltage amplifier (Tegam 2450 Precision Power Amplifier) and subsequently applied to the gate electrode of the MEMS switch. The resulting output signal was monitored and recorded using an oscilloscope (Tektronix 4 Series B MSO Mixed Signal Oscilloscope).