A two-stage insulation method for suppressing thermal crosstalk in microarray sensitive units

Thermal crosstalk between array structures is a key factor in limiting the sensitivity of micro-nano array sensors. We propose a two-stage thermally isolated structure with thermal holes and heat dissipation layer and pulsed voltage heating to reduce thermal crosstalk. Through finite element thermal simulation analysis as well as thermal interference test, the results show that the thermal crosstalk of the two-stage structure is reduced by 12.89% and 39.67%, respectively, in the steady state compared to the structure with no thermal isolation, and pulsed voltage heating leads to the thermal crosstalk of the two-stage structure to be <10%.

Thermal crosstalk between array structures is a key factor in limiting the sensitivity of micro-nano array sensors.We propose a two-stage thermally isolated structure with thermal holes and heat dissipation layer and pulsed voltage heating to reduce thermal crosstalk.Through finite element thermal simulation analysis as well as thermal interference test, the results show that the thermal crosstalk of the two-stage structure is reduced by 12.89% and 39.67%, respectively, in the steady state compared to the structure with no thermal isolation, and pulsed voltage heating leads to the thermal crosstalk of the two-stage structure to be <10%.© 2024 The Author(s).][11] However, when each unit is in the independent heating operating state they will generate a thermal field around it, due to chip materials such as Si or SiO 2 having a certain degree of thermal conductivity. 12,13)][20] The use of polyimide (PI) films as a substrate instead of traditional silicon chips enables effective thermal isolation from the heating elements to the substrate.However, for adjacent sensors on the same level, just reducing the thermal conductivity of the substrate material is still quite a big gap in achieving the desired thermal isolation effect.][22] Existing studies have focused on sensor array unit spacings within the range of hundreds of micrometers.Limited research addresses thermal crosstalk issues in sensor arrays with densely packed spacings, in the range of tens of micrometers or even smaller.Further investigation is crucial to assess their practical effectiveness.[29][30][31] This paper proposes a two-tier thermal isolation structure design with a PI substrate and Pt electrodes, along with a method of pulsed voltage heating.This structure is modeled and optimized using COMSOL software.Simulation studies analyze the dimensions of the thermal isolation holes in the first structure, the material of the heat dissipation layer in the second structure, and the thickness of the metal film.The impact of these factors on sensor thermal crosstalk is evaluated to determine optimal parameter designs.Finite element thermal simulations are performed before and after thermal isolation, along with thermal interference testing of the thermal isolation units under steady-state and transient conditions.The thermal crosstalk impact of the two-stage structure is investigated under transient conditions with pulsed voltage heating, comparing experimental results with those obtained from an analytical model.
The proposed anti-thermal crosstalk structure designed (Fig. 1) features an array with a PI substrate (thickness of 20-30 μm) and Pt electrodes.The primary structure [Fig.1(a)] involves etching rectangular thermal isolation trenches around each heating unit, simulating air filling within the trenches.Due to the extremely low thermal conductivity of the air (0.026 W mK −1 ), these trenches effectively provide thermal shielding between adjacent units.The secondary structure [Fig.1(b)], built upon the first structure, introduces a patterned heat dissipation layer beneath the thermal isolation holes.This results in square voids below the heating region, creating a high thermal conductivity layer between adjacent heating units and forming a heat dissipation band.When a unit operates and generates heat, the heat diffuses radially on the substrate, leading to a higher temperature in inactive units.The patterned heat dissipation layer guides the heat to the dissipation layer, optimizing the conduction path of the electrode layer.This design effectively directs heat in a specific direction, both reducing the thermal crosstalk effect between units in the horizontal direction, improving thermal management efficiency, and ensuring the heating efficiency of the heating units.
Pulsed voltage heating [Fig.1(c)] simultaneously generates heating effects by applying transient voltage pulses.Here, the voltage undergoes a rapid rate of change, resulting in brief yet intense voltage flowing through the heating object.The resistive heat generated by these voltages causes the heating object's temperature to increase.Pulsed heating enables controlled adjustment of the thermal load based on requirements.By finely tuning the pulse amplitude and width, the release of thermal energy for each pulse can be precisely controlled.This ensures effective heating while minimizing the impact of thermal load on the surrounding environment, thereby reducing thermal crosstalk.
Using the COMSOL Multiphysics finite element simulation software, a dual-stage structure for a microheater array was established, and its resistance to thermal crosstalk was analyzed.Structural optimization was performed to achieve optimal performance.Under ideal environmental conditions (ambient temperature of 25 °C), thermal convection between the heating plate and the air was modeled as natural convection with a convective coefficient of 10 W/(m 2 •K).Minor thermal radiation effects were neglected.The thermal conductivity of the PI substrate and Pt electrodes were 0.35 W/(m•K) and 71.6 W/(m•K), respectively.
The anti-thermal crosstalk performance of a thermal isolation structure is closely related to the geometric dimensions of the isolation holes, the chosen heat dissipation materials, and the thickness of the heat dissipation layer.Therefore, these factors must be comprehensively considered. 32)The selection criterion is based on the thermal crosstalk rate between heating units, defined as the temperature rise in an adjacent unit when an adjacent unit is electrically heated: Here, nm AE is the thermal crosstalk rate to the nth unit when the mth unit operates, T n ∆ is the temperature rise of the nth heated unit due to the operation of the mth unit, and T n is the temperature of the nth unit when it operates.
Considering adjacent units and applying a stable input voltage of 1 V, a parametric scan of the thermal isolation hole width from 0 to 20 μm is conducted in the first-level structure.The thermal crosstalk rates for different-sized thermal isolation holes are illustrated in Fig. 2(a).As the hole width increases, the thermal crosstalk rate between adjacent units gradually decreases, demonstrating improved thermal isolation performance.However, this is accompanied by a decrease in the structure's mechanical stability.Considering device miniaturization, integration, and mechanical stability, the optimal hole width is determined to be 10 μm, achieving the best thermal crosstalk rate of 36.085%.
Simulation comparison was made to determine the best metal material for optimal heat dissipation in the second-level structure: Au, Ag, and Cu were tested.The highest temperature reached on the heated unit during electrical activation was compared.Materials exhibiting high temperatures When Cu was used as the heat dissipation layer material, the temperature on the heating unit was the lowest (175.4 °C), indicating the best heat dissipation performance.Although silver (Ag) has good thermal conductivity, it is expensive, making it too costly for large-scale applications.In addition to being expensive, gold (Au) is susceptible to oxidation and corrosion at high temperatures or in humid environments, which can affect the performance and lifetime of the heat sink.Thus, Cu emerged as the optimal choice for the heat dissipation layer.
The thickness of the heat dissipation layer also significantly influences thermal crosstalk.A parametric scan of this layer was conducted with thicknesses ranging from 0 to 20 μm [Fig.2(c)].As the thickness increased 1 to 10 μm, the thermal crosstalk rate significantly decreased, followed by gradual stabilization.Simultaneously, stress also increased with increasing thickness.Therefore, 10 μm was determined as the optimal thickness for the heat dissipation layer, resulting in a thermal crosstalk rate of 33.264%.
The micro-heater array is configured as a 4 × 4 array with a 30 μm spacing between units (Fig. 3).The electrode width and  The steady-state interference simulation results are presented in Fig. 3(c).Under stable input voltages of 0.1-0.5 V, the proposed structure exhibits progressively improved antithermal crosstalk effects compared to the structure without thermal isolation.Specifically, at the same input voltage, the second-level structure reduces the thermal crosstalk by approximately 50% relative to the structure without thermal isolation.
Traditional steady-state heating methods in array structures lead to highly concentrated and accumulated heat energy, exacerbating the adverse effects of thermal crosstalk.This study uses pulsed voltage heating, precisely manipulating pulse voltage amplitudes and widths to manage temperature and time during the heating process.This allows the device to quickly reach the desired operating temperature.Moreover, given this method's transient nature, heat energy is only transferred to neighboring units within a limited time, reducing the heat transfer window within the system and consequently lowering the impact of thermal crosstalk caused by heat conduction.
Transient temperature field analysis was conducted using COMSOL Multiphysics, with a voltage amplitude of 1 V, pulse width of 100 ms, and a sampling point every 5 ms.The thermal crosstalk rates between units for the traditional structure and the proposed second-level structure are shown in Fig. 3(d).Pulsed heating effectively reduces thermal crosstalk between device units, with the second-level structure exhibiting the best anti-thermal crosstalk performance (crosstalk rate of 6.69%).For a 3 V pulse voltage of 3 V and 100 μs

037001-4
© 2024 The Author(s).Published on behalf of The Japan Society of Applied Physics by IOP Publishing Ltd pulse width (sampling every 10 μs), increasing the pulse voltage amplitude and decreasing the width significantly reduces the thermal crosstalk rates for the two-stage structure with thermal isolation [Fig.3(e)].Even for the structure without thermal isolation, changing the heating method results in a marked reduction in thermal crosstalk rates.
Figure 4 shows the processed structural components.Figure 4  The miniature working region of the arrayed micro-heater makes it difficult to use conventional contact-based or IR measurement methods for characterizing its working area temperature.Therefore, the temperature is assessed utilizing the linear resistance variation property of the Pt heating electrode. 33,34)If the temperature coefficient α is temperature-independent, the relationship between the material's resistance and temperature (R-T) can be expressed as α represents the temperature coefficient of resistance (TCR), and R(T 0 ) is the resistance at the initial temperature T 0 .The resistance of the Pt electrode increases linearly with temperature, demonstrating a well-defined linear relationship between resistance and temperature (Fig. 5).The TCR for the Pt electrode is determined to be 0.0071 (1 K −1 ) using a least squares error fit to the regression line obtained via fitting.
The microheater array was powered using the Keysight B2900A Precision Source/Measure Unit (SMU) with the heating units.The B2900A series features compact and costeffective desktop SMUs designed for high-resolution and high-precision current and voltage (I-V ) measurements.Here, I-V curves were used to characterize the thermal performance of the microheater array.As the microheater operates as a pure resistive device, resistance values can be obtained from the I-V curves.The temperature of the heating unit is then quantitatively determined by combining the I-V curves, TCR values, and the measured resistance changes, using Eqs.(1) and ( 2).This allows for a quantitative assessment of thermal crosstalk between heating units in the array.The steady-state experimental range covered thermal crosstalk rates between microheater array units at input voltages of 0.1-0.5 V in increments of 0.1 V [Fig.5(b)].At 0.5 V, the thermal crosstalk rate for the secondary structure was 25.74%, a reduction of approximately 30.6% compared to the primary structure and 40% compared to the no-isolation structure.Under transient condition [Figs.5(c) and (d)], in the transient state, power is supplied using a pulse voltage with a period of 200 ms and a duty cycle of 0.5, when the voltage amplitude is 1 V and the pulse width is 100 ms, the crosstalk rates were 25.96% and 6.93% for the primary and secondary structures, demonstrating the effective reduction of thermal crosstalk between array units using pulsed voltage heating.As the pulse voltage amplitude increased, the crosstalk rate also increased.However, even at 2 V, the thermal crosstalk rate for the secondary structure remained below 10%, proving the feasibility of pulsed voltage heating.Future research can focus on further reducing the crosstalk impact by increasing the voltage amplitude or reducing the pulse width.
Notably, the experimentally obtained anti-thermal crosstalk efficiency is consistently lower than the simulated results.In the simulations, the PI substrate under the Cu heat dissipation layer was not considered for simplicity.Additionally, the actual temperature in the laboratory differed from the simulated preset temperature, leading to a suboptimal anti-thermal crosstalk rate.Furthermore, in the physical fabrication process, the precision and accuracy of the microstructure play a crucial role in influencing experimental outcomes.
In this paper, we design, simulate, and test a two-stage thermal isolation structure for a 16-unit sensor array with PI as the substrate and platinum (Pt) as the heating electrode material, where the primary structure consists of heatinsulating holes, and the secondary structure consists of heat-insulating holes and a patterned heat dissipation layer.Simulation studies were used to optimize design parameters including hole width, metal film material, and metal film thickness.Steady-state and transient thermal interference simulations were performed on the two-stage structure.Under a steady-state voltage of 0.5 V, the thermal crosstalk rates for the two stages were 36.06% and 24.99%.Compared to the no-isolation structure, the thermal crosstalk for the twostage structure decreased by 12.89% and 39.67%.Under transient conditions with a 1 V pulsed voltage, the thermal crosstalk rate of the two-stage structures is 23.09% and 6.69%, respectively.The second-stage structure, in combination with the pulse voltage heating mode of operation, can be realized to achieve good thermal isolation, so that the thermal crosstalk can be successfully mitigated.Experimental testing using the temperature-resistance characteristics of Pt electrodes and I-V curves produced results consistent with the simulations, validating the rationality of the two-stage design and highlighting the benefits of using pulsed voltage heating to achieve precise, reliable, and low-crosstalk temperature output, holding significant implications for enhancing the operational stability and reliability of array sensors.

Fig. 3 . 3 ©
Fig. 3. (a) The simulation model of the primary structure.(b) The simulation model of the secondary structure.(c) The thermal crosstalk rate at 0.1-0.5 V under steady-state conditions for uninsulated and two-stage structures.(d) The thermal crosstalk rate for uninsulated and two-stage structures under transient conditions with 1 V voltage amplitude and 100 ms pulse width.(e) The thermal crosstalk rate for uninsulated and two-stage structures under transient conditions with a 3 V voltage amplitude and 100 μs pulse width.

Fig. 4 .
Fig. 4. The two-stage structure samples.(a) The single-story primary structure.(b) The two-tier secondary structure.
(a) represents the first-level structure, featuring insulation holes around heating units.The top and bottom holes have the same size of 60 μm × 10 μm, and the right and left holes have the same size of 80 μm × 10 μm.

Figure 4 (
b) illustrates the second-level structure, consisting of an upper insulation hole layer, identical to the first-level structure.The lower orange section represents the Cu thermal film, and the black region beneath the heating unit corresponds to the 80 μm × 80 μm square cutout.

Fig. 5 .
Fig. 5. Thermal crosstalk performance testing of two-stage structures.(a) The electrical resistance ratio of Pt heating electrode versus applied temperature.The calculated TCR is 0.0071 K−1.(b) No thermal crosstalk from insulated and two-stage structures at a constant voltage of 0.1-0.5 V. (c) The crosstalk rate of primary structure at a voltage amplitude of 1-2 V and pulse width of 100 ms.(d) The crosstalk rate of secondary structure at a voltage amplitude of 1-2 V and pulse width of 100 ms.