Dynamic Performance of Hybrid Infrared Modulator Based on Single‐Walled Carbon Nanotubes and Ionic Liquid Enabling Fast Response

Abstract

Herein, a new type of infrared film using single‐walled carbon nanotubes (SWNTs) is developed. The optimal sonication settings to produce SWNTs with the highest possible absorption of infrared light are discovered. A hybrid infrared modulator that combines SWNTs with ionic liquids, with a very fast response time is also developed. By carefully controlling the size of the SWNT bundles and the pore size of the filter membrane used to deposit the SWNTs, a response time of a few hundred microseconds (635 μs) at the half‐power point is achieved, which is the optical 3‐dB point. This technology has the potential to be used in a variety of innovative applications, such as smart windows and military camouflage, due to its excellent infrared optical properties.

Introduction

An infrared (IR) modulator can be understood as a device capable of altering the properties of an IR electromagnetic signal by varying its amplitude, frequency, or phase.^[] By modulating IR signals in this way, an IR modulator allows for detailed information to be encoded and extracted for applications involving detecting and analyzing IR radiation. Examples include security surveillance using IR cameras,^[] medical IR imaging technologies,^[] and nondestructive testing methods employed in industrial inspection.^[] IR modulators have seen increasing use owing to their ability to accurately resolve spatial and temporal variations within IR spectra,^[] thus advancing our fundamental understanding and technological capabilities in domains involving IR optical phenomena.

A hybrid IR modulator can be understood as an advanced device capable of refined manipulation and selective filtering of IR electromagnetic signals through the integrated utilization of multiple constituent materials. In particular, including optically transparent yet electrically conductive components enables the modulated variation of IR optical intensity. Such a configuration affords enhanced control over IR optical signals, rendering hybrid IR modulators well-suited for applications like IR imaging instrumentation, laser optical telecommunication networks, and fiber optic data transmission lines.^[] A key advantage of the hybrid design is the ability to switch signal states with high temporal and amplitude precision. This permits faster modulation compared to conventional electronically-based technologies. Additionally, hybrid IR modulators consume less electrical power, implying greater energy efficiency and more favorable cost profiles relative to traditional modulator designs.^[]

Carbon nanotubes show promise as a material for generating ultrashort optical pulses owing to their ultrafast excited carrier dynamics, allowing them to modulate IR radiation with rapid switching times on the picosecond scale when utilized in electro-optic modulators.^[] Carbon nanotubes also possess a high specific surface area and can be combined with ionic liquids or electrolytes to improve their electro-optic performance in the IR region. However, such combinations result in ion diffusion imposing a limit on switching speeds of carbon nanotube modulators in the millisecond to microsecond range, constraining their integration into various photonic device applications.^[] A key advantage of carbon nanotube-based electro-optic technology is its capability to function at high frequencies, enabling rapid and accurate handling of large data throughputs. Additionally, carbon nanotube modulators can be readily interfaced with existing electronic systems.^[]

Conversely, ionic liquids possess low volatility and high polarities that render them suitable as solvents and electrolytes.^[] For instance, researchers have combined carbon nanotubes with ionic liquids to develop an IR modulator with a fast response time.^[] In another study, a carbon nanotube thin film coated with an ionic liquid was used to demonstrate the modulation of mid-IR radiation, achieving an optical response time of approximately 3.2^[] and 2.5^[] milliseconds which is relatively fast compared to other IR modulators using different materials.

Carbon nanotubes exhibit an optical response time, which refers to the time it takes for the material to change its optical properties in response to an external stimulus such as changes in an electric field or temperature.^[] Carbon nanotubes’ response time varies depending on their diameter, length, and surrounding environment.^[] Typically, carbon nanotubes exhibit fast response times, ranging from picoseconds to nanoseconds.^[] The carbon nanotubes’ chirality, which refers to how the graphene lattice is wrapped to form the tube, can also influence the response time. Further, metallic carbon nanotubes have faster response times than semiconducting ones.^[] It is important to note that the exact response time of an IR modulator based on carbon nanotubes and ionic liquids can be influenced by various factors, including the type and properties of the carbon nanotubes and ionic liquids used, as well as the device's geometry and operating conditions.^[] Furthermore, the application of ionic liquids shows promise for various applications due to their combination of favorable characteristics such as nonflammability over a wide temperature range as liquids, low or nontoxicity, high chemical stability, strong ionic conductivity, powerful solvation of organics and inorganics, and environmental friendliness.^[]

Applying an electric field can alter the energy band-shifting behavior of semiconductor materials in response to an electric field, and this also applies to carbon nanotubes. Interestingly, the direction and strength of the electric field can cause shifts in the absorption spectrum of carbon nanotubes toward higher or lower energies. By tuning these parameters, the absorption spectrum of carbon nanotubes can be adjusted to a desired wavelength range. In this study, ionic gating is achieved by applying transverse electric fields to a network film of carbon nanotubes, which involves an electric field perpendicular to the plane of the carbon nanotube electrodes. This configuration is analogous to supercapacitors, where a dielectric material or spacer separates two parallel oppositely charged electrodes.

In this approach, electrostatic gating induced by an ionic liquid produces electrical double layers at the interface between the ionic liquid and carbon nanotube electrodes. Then, by applying a voltage to the carbon nanotube electrodes immersed in the ionic liquid, ions from the liquid can enter the nanotube network film, inducing shifts in the Fermi level and modifying the electrical and optical properties of the nanotubes. Consequently, these changes affect the absorption and transmission of light by the carbon nanotubes. This ionic gating technique offers the advantage of being nondestructive and reversible, as gating can be removed by changing the voltage polarity. Finally, the study will investigate whether single-walled carbon nanotube (SWNT) network films can serve as electrically configurable optical media that modulate IR transparency in a spatially controlled manner.

In the aforementioned study,^[] it was found that thin films comprised of semiconducting SWNTs (SC-SWNTs) could undergo rapid switching facilitated by an ionic liquid and a counter electrode composed of metallic SWNTs. However, increasing the thickness of the semiconducting layer slowed the switching time, an effect which could be counteracted by also increasing the thickness of the thin film for the metallic counter electrode while maintaining high modulation depth.

The present study reports significant advancements in the fabrication process of an IR modulator based on SWNT films, focusing on the effect of membrane pore size on precise control of SWNT bundle formation during vacuum filtration and investigating their impact on the performance of the IR modulator. We applied low flow rates to gently filter the solution at the low-pressure drop or slow flow rates, i.e., < 30 kPa. The IR-active electrode of the modulator comprises randomly entangled SC-SWNTs, while the counter electrode is composed of metallic SWNTs (MT-SWNTs) and utilizes a specific ionic liquid named 1-Ethyl-3-methylimidazolium bis(trifluoromethylsulfonyl)imide (EMIM-TFSI), which exhibits higher conductivity and lower viscosity than the previously used DEME-TFSI.^[]

Based on our experimental results, the optical modulation and temporal response of the IR modulator have significantly improved. Our findings indicate that the cutoff frequency limitation reported in a previous study^[] of 110 Hz has been surpassed by 2.27 times. Specifically, we have achieved a new record for the cutoff frequency, reaching approximately 250 Hz in our work. These results hold promising implications for the development of SWNT-based IR devices. Moreover, the study suggests that the control of SWNT bundle formation during vacuum filtration, coupled with the utilization of an ionic liquid with unique properties, such as low absorption in the IR region, high conductivity, and lower viscosity, such as EMIM-TFSI, can significantly improve the performance of the IR modulator.

Results and Discussion

To fabricate network films of SWNTs using an aqueous surfactant solution of MT-SWNTs and SC-SWNTs, various techniques were utilized, including sonication, vacuum filtration, transfer, and hot acetone vapor process. First, the SWNTs were sonicated for different time intervals (1 and 30 min) in 10 mL of deionized water at a frequency of 37 kHz with low power (30%) and high power (70%) to obtain well-dispersed and de-aggregated SWNTs. Next, the vacuum filtration technique was employed using a mixed cellulose ester membrane with pore sizes of 0.1 and 0.05 μm, and the control of SWNT film thickness was achieved by carefully regulating several key parameters during the vacuum filtration process. These parameters included the quantity of SWNT material utilized, the working area of the filtration membrane, and the desired SWNT film density, which was set at $1.2\mathrm{g}{\text{cm}}^{-2}$.^[] Afterward, to transfer SWNTs onto a specific substrate, a hot acetone vapor technique is used as previously explained in the literature,^[] where the membrane containing the SWNT films underwent a process in which the hot acetone bath was utilized to selectively dissolve the membrane, thereby facilitating the transfer of the resulting SWNT films onto glass substrates. These glass substrates were prepatterned with Ti (50 nm)/Pt (250 nm) electrodes. The experimental setup assembled the IR modulator by carefully placing approximately 15 μL of ionic liquid in the designated area between the two electrodes. A spacer with a thickness of 250 μm was used during the assembly process to ensure precise alignment and consistent spacing. This procedure was carried out inside a glove box to maintain a controlled environment. The detailed configuration of the IR modulator can be observed in Figure and further illustrated in Figure S1 and S2 (Supporting Information). The diagram provides a visual representation of the arrangement and positioning of the components involved in the modulator assembly. This methodology constructed the IR modulator with the necessary components and spatial configuration to enable its functionality and subsequent characterization. The precise placement of the ionic liquid between the SWNT electrodes, facilitated by the spacer, ensured the proper operation of the modulator and allowed for the succeeding investigation and analysis of its IR modulation capabilities.

[IMAGE OMITTED. SEE PDF]

The study identified the best conditions for obtaining well-dispersed and deaggregated SWNTs through sonication. The size of the pores in the membrane used for vacuum filtration affected the size of the SWNT bundles, with larger pores (0.1 μm) producing smaller bundles and smaller pores (0.05 μm) producing larger ones. The thickness of resulting SWNT films was approximately 61 nm for SC-SWNT and 22 nm for MT-SWNT, with contact resistances of approximately 0.9 and 1.3 kΩ, respectively. Based on the images obtained through scanning electron microscopy (SEM), it can be concluded that the bundles with larger sizes (60–80 nm) and high porosity are more suitable for optical absorption purposes (as seen in Figure S3 and S4, Supporting Information). Conversely, the smaller size bundles (30–40 nm) exhibit lower porosity levels (as observed in Figure S5 and S6, Supporting Information). A sonication time of 1 min with low power was used to achieve a larger bundle diameter. In comparison, a sonication time of 30 min with high power was used for a smaller bundle diameter. However, excessive sonication can lead to new SWNT bundle formation due to attractive van der Waals forces between adjacent tubes.^[]

We utilized the atomic force microscopy (AFM) method to examine the SWNTs suspension before the vacuum filtration process. The AFM images indicated that the SC-SWNT bundle's average size was about 2 nm and a length ranging from 1.0–1.5 μm, while the MT-SWNT bundle's size was approximately 6 nm and a length between 0.5 and 1.5 μm (refer to Figure S7 and S8, Supporting Information). Furthermore, the SEM images taken after the vacuum filtration process can also showcase the SWNT network's large and small bundle size and shape, as shown in Figure S9–S12 (Supporting Information). After analyzing the data from AFM and SEM, it was discovered that incorporating large bundles with high porosity can greatly enhance IR absorption. The validity of this claim will become apparent later when linking these analyses with the dynamic performance of the modulator. In contrast, smaller bundles might be useful for applications such as energy storage.^[]

The purity and quality of SWNTs were assessed by analyzing their absorption spectra in the UV–vis–near-infrared (NIR) region. This was done after the SWNTs were transferred onto glass substrates that were 0.96 mm thick. This analysis involved measuring light absorption across various wavelengths in the electromagnetic spectrum. The electronic transitions in the graphene-like sheets that form the walls of the SWNTs determine their absorption spectra.^[]

Figure illustrates the optical absorption properties of SWNTs. SC-SWNTs exhibit characteristic optical transitions, denoted as S₁₁ and S₂₂, that occur in the NIR and visible wavelength ranges. These intrinsic optical transitions associated with SC-SWNTs can be modulated by applying an electric field or doping to distort the SWNT band structure. In contrast, MT-SWNTs display an M₁₁ optical transition. However, the M₁₁ transition is observed in similar NIR to visible wavelengths as the S₁₁ and S₂₂ transitions of SC-SWNTs.

[IMAGE OMITTED. SEE PDF]

The differences in optical absorption characteristics between SC-SWNTs and MT-SWNTs arise from their distinct electronic band structures. SC-SWNTs exhibit a small bandgap, allowing discrete optical transitions to be modulated by electric fields. MT-SWNTs lack a bandgap, thereby absorbing continuously, but still display characteristic intraband transitions in the NIR to a visible range similar to semiconducting tube transitions. This Figure helps characterize the fundamental optical properties of SWNTs based on their metallic or semiconducting electronic properties.

Based on recent computational investigations utilizing density functional theory (DFT) calculations, valence band photoemission analysis, and DFT outcomes, it has been observed that the electronic structure of ionic liquids incorporating the TFSI anion exhibits a notably intricate nature, which can manifest diverse characteristics contingent upon the specific cation employed.^[] In light of these findings, we experimentally compared the optical transparency of two ionic liquids, EMIM-TFSI and DEME-TFSI, through Fourier transform infrared (FTIR) spectroscopy. FTIR spectroscopy will analyze absorption across the IR wavenumber range from $3000$ to $8000{\text{cm}}^{-1}$ using the same device configuration of the IR modulator. However, the device consists of identical glass substrates separated by a spacer without any SWNT electrodes. This experimental setup intends to assess and compare the intrinsic IR transmission properties of the two ionic liquids via FTIR spectroscopy. Our results showed that EMIM-TFSI had slightly better transmittance in the short-wave infrared (SWIR) region, particularly at a wavenumber of $5482{\text{cm}}^{-}{}^1$ (equivalent to 1824 nm) when compared to DEME-TFSI. The increased transparency of EMIM-TFSI is attributed to its shorter alkyl chain length, which allows for more efficient molecular packaging and reduces light scattering. These findings suggest that EMIM-TFSI could be a suitable choice for applications that require high transmittance in the SWIR region, as illustrated in Figure . The EMIM-TFSI had a wide electrochemical window (from +2.6 to −2.1 V) and low electrode–electrolyte interfacial impedance, facilitating efficient Faradaic and non-Faradaic charge transfer processes at interfaces.^[]

[IMAGE OMITTED. SEE PDF]

Moreover, we extend our investigations by subjecting the ionic liquid (IL; i.e., EMIM-TSFI) to the influence of a direct current (DC) voltage characterized by a constant polarity. This deliberate application of electrical potential facilitates the displacement of positive and negative ions within the ionic liquid, provoking a controlled charge and discharge process. This displacement caused the optical absorption of S₁₁ band to be modulated, resulting in the modulator displaying an “OFF” state when it absorbed light and an “ON” state when it transmitted light. The ability of the modulator to switch between high- and low-absorbing states was determined by the polarity and magnitude of the applied voltage, allowing it to function as a reversible optical modulator.

The maximum absorption state of the modulator corresponded to the SWNT's intrinsic state (OFF) at voltage +1.3 V. In contrast, the minimum absorption state corresponded to the SWNT's p-type state (ON) at −1.1 V, as illustrated in Figure . An ON/OFF state IR modulator that reversibly switches between transmitting and absorbing IR light fully transmits IR light when in the “ON” state and fully absorbs IR light when in the “OFF” state. Switching between these states modifies the modulator's optical properties.

[IMAGE OMITTED. SEE PDF]

The study reveals that the SC-SWNTs (i.e., IR-active electrode) exhibit a substantially higher optical absorption of light in the SWIR range ($4000–10000{\text{cm}}^{-}{}^1$ or 1000–2500 nm) than the MT-SWNTs (i.e., counter electrode), which have very low absorption in this range, enabling SWIR light to penetrate (refer to Figure ). To investigate the effect of ionic gating on the optical absorption band S₁₁, we assembled the IR modulator and placed the ionic liquid in the designated area. We performed UV–vis–NIR measurements on the modulator while applying DC voltages. The zero-voltage state of the modulator corresponded to the ground state, where the ions of the ionic liquid were unpolarized.

In the initial stage of characterizing the IR modulator, a very important step involves quantifying the modulation depth expressed as a percentage. This parameter serves as a fundamental metric in assessing the extent of modulation achieved by the modulator in the IR spectrum, beginning by calculating the difference between the highest and lowest absorption signals of the SC-SWNT. Then, divide this difference by the highest and lowest absorption signals sum. Finally, multiply the result by 100 to represent the modulation depth as a percentage.^[]

Therefore, the modulation depth of the IR modulator was calculated by determining the maximum ($A_{\text{max}}$) and minimum ($A_{\text{min}}$) absorption values according to Equation (), revealing a significant modulation depth of about 94.6% at a wavenumber of $5482{\text{cm}}^{-}{}^1$ (corresponding to 1824 nm), as illustrated in Figure . Nonetheless, having a high modulation depth is essential in optical communication as it enables the transmission of more data within a specific bandwidth. However, to accurately measure the dynamic response of the modulator, it is essential to carefully choose an IR source that closely matches the specific wavenumber of $5482{\text{cm}}^{-1}$ (equivalent to 1824 nm). After careful consideration, we have determined that the nearest commercially available source in the market operates at a complementary wavenumber of $5555{\text{cm}}^{-1}$, corresponding to a wavelength of 1800 nm. This selection allows us to capture and analyze the dynamic behavior of the IR modulator in this specific wavelength.1$\text{Modulation\hspace{0.17em}depth}=\left[\frac{A_{\text{max}}-A_{\min }}{A_{\max }+A_{\min }}\right]\times 100\%$

Figure displays a color plot that outlines the absorbance properties of an IR modulator with the applied DC voltage. It distinctly highlights the influence of DC voltage on the modulator's absorbance characteristics at two specific wavenumbers: $5482{\text{cm}}^{-}{}^1$, corresponding to the S₁₁ bandgap transition, and $10000{\text{cm}}^{-}{}^1$, corresponding to the S₂₂ bandgap transition. The color plot reveals that the absorbance reaches its maximum intensity at a specific DC voltage of +1.3 V, indicative of the modulator's optimal efficiency in absorbing IR radiation during the S₁₁ transition. Conversely, the minimum absorbance, as denoted by the lowest intensity, manifests at a voltage of −1.1 V, signifying a diminished capacity of the modulator to absorb radiation during the S₁₁ transition at this voltage. Such observations have paramount implications for tailoring and fine-tuning the modulator's performance in applications that necessitate either highly efficient or minimal IR radiation absorption at the designated bandgap transitions.

[IMAGE OMITTED. SEE PDF]

The optical hysteresis of an IR modulator is a crucial feature that determines how its optical response is affected by the electric field history it experiences. Specifically, in terms of the DC voltage of the modulator, its optical hysteresis refers to how the modulation or transmission of the IR beam changes as the voltage increases or decreases and how this response is influenced by past voltage history. The voltage and optical response relationship may not be linear in an IR modulator with hysteresis. Instead, the device's response may be delayed as the voltage changes, and the response may also be influenced by the voltage's direction and rate of change. A hysteresis loop typically represents this behavior, a Figure shows how the optical response changes as the voltage is swept up and down at a specific wavenumber of $5482{\text{cm}}^{-}{}^1$, corresponding to the S₁₁ bandgap transition.

The present investigation reports the behavior of an IR modulator under a DC voltage applied with gradually increasing positive polarity until reaching maximum absorption, which was considered complete charging of the modulator. Upon decreasing the voltage, the absorption decreased, indicating the presence of optical hysteresis in the modulator at the S₁₁ bandgap transition. The hysteresis loop was fundamentally attributed to the relaxation times of the IL ions,^[] which respond to changing electric fields. Interestingly, changing the polarity to a negative voltage resulted in the modulator reaching minimum absorption at a full negative charge, as illustrated in Figure . The hysteresis was attributed to different reconfigurations of the IL-SWNT interface during the charging process (forward bias) in Figure versus the discharging process (reverse bias) in Figure .

[IMAGE OMITTED. SEE PDF]

To dynamically characterize the IR modulator, we have successfully developed and implemented a novel external modulation technique, enabling accurate control over the intensity of an IR light-emitting diode (LED) light source within the SWIR region. We can achieve the desired modulation by continuously operating the IR LED light source and applying an AC-modulated signal with a square waveform by integrating the IR modulator as an external modulator (as shown in Figure S13, Supporting Information). In this study, we successfully demonstrated the indirect (i.e., external) modulation of IR LED light in the SWIR region around the 1800 nm wavelength, aligning with the maximum absorption for the S₁₁ band-gap transition of SC-SWNTs.

Our investigation focused on the modulation of optical switches, which plays a significant role in determining the frequency response of the modulator. This involves measuring the frequency at which the output voltage decreases by 0.5 (or −3 dB) of the maximum output voltage. This frequency is called the optical 3-dB point, which marks the cutoff frequency where the spectral density reaches 50% of its maximum value. This represents the highest frequency at which the modulator can operate while maintaining acceptable performance.

Later on, we investigated the impact of membrane pore size on the performance of an IR modulator utilizing the LED modulation technique. The modulator was constructed using small bundles of SWNTs. Two different modulators were examined: one incorporated SWNTs filtered through a membrane with a pore size of 0.1 μm with a cutoff frequency $f_{\mathrm{c}}$ $\simeq$ 39 Hz, while the other utilized SWNTs filtered through a membrane with a pore size of 0.05 μm with a cutoff frequency $f_{\mathrm{c}}$ $\simeq$ 102 Hz, as illustrated in Figure . The frequency response of the modulators was analyzed by plotting the voltage ratio, which means the maximum modulation depth, as a function of modulation frequency to evaluate the modulation performance.

[IMAGE OMITTED. SEE PDF]

Based on the results, it can be concluded that the IR modulator with a smaller pore-size membrane performed better than the one with a larger pore size. This enhancement can be attributed to the significant influence of membrane pore size on the retained bundle size of SWNTs. By utilizing a membrane with smaller pore size, larger SWNT bundles can be retained on the membrane surface, while smaller bundles can freely pass through the pores.^[] Consequently, the filtration of SWNT suspension through a membrane with a small pore size presents a practical and effective technique for generating high-quality SWNT films characterized by low roughness and high conductivity. In contrast, the larger SWNT bundles create a highly porous network that is a conductive scaffold suitable for electrochemical energy storage applications.^[] This porous structure facilitates the unimpeded flow of ions, thereby preventing mass transport limitations in electrochemical devices.^[] The frequency response of the IR modulator in this study corresponds to that of an RC low-pass filter, allowing low-frequency signals to pass through while attenuating or blocking high-frequency signals. We utilized a fitting process to examine the experimental data by employing the frequency-dependent equation of a resistor–capacitor (RC) low-pass filter. Equation () illustrates the magnitude of the IR modulator's impact on the amplitude of every input signal. The following equation^[] accurately describes this:2$\text{Voltage\hspace{0.17em}Ratio}=\frac{1}{\sqrt{{[2\pi f(RC)]}^2+1}}$

Here, the voltage ratio denotes the voltage attenuation of the modulator, f represents the frequency of the input signal in hertz, R represents the resistance in ohms, and C represents the capacitance in farads. Furthermore, the fitting analysis allowed us to extract the time constant $\tau =RC$,^[] which measures the IR modulator's response time. The time constant signifies the time required for the modulator to reach 63.2% of its final output level after applying a step input. By determining the time constant, we calculated the modulator's cutoff frequency $f_{\mathrm{c}}$, representing the frequencies over which the modulator can operate effectively. Moreover, the cutoff frequency $f_{\mathrm{c}}$ is a significant parameter employed to measure the frequency at which the intensity of IR light is reduced by 3-dB (or 50%) compared to its value at DC. The modulator's time constant can be utilized to evaluate its performance, where a faster rise time indicates a more efficient modulator. The relationship between the cutoff frequency and the time constant can be represented by the following equation:^[]3$f_{\mathrm{c}}=\frac{1}{2\pi RC}$

In Equation (), $f_{\mathrm{c}}$ represents the –3-dB cutoff frequency (i.e., half-power frequency) in hertz, R signifies the resistance in ohms, and C denotes the capacitance in farads. As shown in Figure , the modulator represented by blue spheres has a time constant (τ) of approximately 4.10 ms, corresponding to a cutoff frequency ($f_{\mathrm{c}}$) of around 39 Hz. In contrast, the modulator represented by green spheres has a shorter time constant (τ) of approximately 1.55 ms, resulting in a higher cutoff frequency ($f_{\mathrm{c}}$) of approximately 102 Hz.

The other objective of this study was to examine the IR modulation performance of the modulator when the SWNT network films were synthesized using different approaches, resulting in small and large bundles while keeping the filtration membrane with a pore size of 0.05 μm. Several factors, including switching speed, extinction ratio, and porosity, were considered to determine which bundle size yields the best IR modulation performance.

The results demonstrated that the modulator with larger SWNT bundles outperformed the modulator with smaller bundles regarding dynamic performance. The data obtained from the larger bundled modulator exhibited higher consistency and showed good agreement with the fitting formula, as illustrated in Figure , where the modulation performance of two distinct IR modulators, differentiated by the optical 3-dB point. One of the modulators employed small bundles of SWNTs, represented by blue spheres, with a cutoff frequency ($f_{\mathrm{c}}$) of approximately 75 Hz. The response time of this modulator was measured to be 2.12 ms. In contrast, the other modulator utilized large bundles of SWNTs, shown by green spheres, with a higher cutoff frequency ($f_{\mathrm{c}}$) of approximately 123 Hz. This modulator's response time was determined to be 1.3 ms.

[IMAGE OMITTED. SEE PDF]

During the frequency response measurements, careful consideration was given to selecting the modulation amplitude (in volts) to ensure it fell within the electrochemical window of the ionic liquid, and the amplitude modulation was 1.5 V in Figure for both curves so that we can have more reliable data and clearly understand the difference in the performance between the large and small bundles modulators. Additionally, the chosen voltage range was within limits determined by the modulator's absorbance behavior in response to DC voltage (As previously stated, within the context of our discussion in Figure ). The improved switching speed observed in the larger bundled SWNT network films can be attributed to their lower overall resistance.^[] This reduction in resistance is attributed to the increased number of intertube connections and conduction paths within the larger bundles.^[] Consequently, the lower resistance minimized the RC delay and facilitated rapid charge propagation, resulting in faster switching. The larger bundles also exhibited fewer nanotube junctions,^[] enabling more direct conducting pathways and further enhancing switching speed. The higher overall density of electronic states for charge carriers within the larger bundles, due to the increased number of SWNTs, contributed to faster changes in IR transmission during modulation.

Another important parameter that has been measured is the IR modulator's extinction ratio, which measures the efficiency of light absorption and is influenced by the bundle size of SWNTs. Larger SWNT bundles have higher extinction ratios than smaller bundles due to their increased light-absorbing capacity. This is because larger bundles contain more SWNTs, allowing them to absorb more light and achieve a higher ratio of absorbed to transmit light. However, there is a saturation point where further increasing the bundle size does not significantly impact the amount of absorbed light. The higher porosity of the SWNT network films with larger bundles resulted in more consistent IR modulation. This increased porosity provided additional open spaces and direct pathways for propagation in response to an applied voltage. Consequently, the larger bundles demonstrated enhanced consistency in IR modulation across the entire film. This analysis highlighted the higher IR modulation performance of SWNT network films with larger bundles. The larger bundles exhibited advantages such as faster switching speeds, higher extinction ratios, lower resistance, and increased porosity, collectively leading to more efficient and consistent IR modulation.

The extinction ratio, an essential measure for optical modulation, is expressed in decibels (dB) and represents the ratio between the maximum optical power transmitted when the modulator is in the “ON” state and the minimum optical power transmitted when the modulator is in the “OFF” state. A high extinction ratio indicates the modulator's ability to switch effectively between these states, which is essential for minimizing noise and interference. In Figure , it can be observed that the modulator with larger bundles exhibits a higher extinction ratio, approximately 2.68 dB, indicates a better signal-to-noise ratio, compared to the modulator with smaller bundles, which has an extinction ratio of about 1.85 dB. In certain situations, long-distance optical communications may require a higher extinction ratio. However, the extinction ratio measures the contrast between the “ON” and “OFF” states of the modulator. Mathematically, the extinction ratio is defined in the following equation:^[]4$\text{Extinction\hspace{0.17em}ratio}=10\text{log}\left(\frac{P_{\text{ON}}}{P_{\text{OFF}}}\right)$where $P_{\text{ON}}$ represents the optical power when the modulator is in the “ON” state, and $P_{\text{OFF}}$ represents the power when the modulator is in the “OFF” state. It is important to note that altering the applied voltage can impact the optical power. When referring to an applied voltage with a frequency of 0.1 Hz, it must be understood that the voltage oscillates from positive to negative polarity (or vice versa) 0.1 times per second.

Figure illustrates that the modulator with larger bundles demonstrates a significantly higher modulation depth, approximately 94.6%, compared to the modulator with smaller bundles, which exhibits approximately 68.7%. In optical modulation, the modulation depth refers to the ratio of the maximum absorption to the minimum absorption of the modulator. It quantifies the extent to which the optical power emitted by a device can be varied. A larger modulation depth implies that a greater change in optical power can be achieved by adjusting the applied DC voltage. This study shows that modulators with larger SWNT bundles and those with higher porosity tend to possess higher optical modulation depths. This can be attributed to their enhanced ability to absorb and emit IR light, resulting from increased SWNTs within the bundles. Consequently, these modulators can produce more pronounced variations in optical power, making them valuable for efficient optical communication applications.

In measuring frequency response, the modulation amplitude (peak-to-peak voltage) was systematically increased while ensuring compliance with the electrochemical window limits of the ionic liquid. However, the absorbance behavior of the modulator, also controlled by DC voltages, provided valuable insights into the modulation depth and the range of these DC voltages, which corresponded to the difference between maximum and minimum absorption levels. This information was instrumental in determining the appropriate maximum modulation amplitude for the frequency response measurements. The ionic liquid's electrochemical window largely determined the choice of modulation amplitude. Exceeding this boundary could lead to irreversible breakdown of the ionic liquid, with negative outcomes like electrolyte damage, gas release, electrode wear, reduced ionic movement, and weakened efficiency and stability. Modulators with larger bundles showed maximum absorption at +1.3 V and minimum at −1.1 V. The voltage gap between these readings was about 2.4 V, which was critical for setting the ideal modulation amplitude for improved optical modulation and response time. Additionally, the selected ionic liquid, EMIM-TFSI, had a broad electrochemical range from +2.6 to −2.1 V, a difference of over 4 V. This deep insight into the system's electrochemical properties offered a clear edge in preserving the IR modulator's durability, functionality, and reversibility without degradation. The formation of SWNT bundles in IR modulators plays a key role in determining the modulation amplitude voltage of these optoelectronic devices. Various factors related to SWNT bundles can significantly influence the modulation amplitude voltage. First, optical absorption is affected by the size and porosity of SWNTs. Larger bundles with higher porosities exhibit more efficient light absorption, allowing for a greater change in optical power or absorption with a given change in applied voltage. Second, conductivity is influenced by the size of SWNT bundles. Larger bundles tend to have higher conductivity, enabling faster response times.

Figure shows the results obtained by applying an amplitude modulation voltage of 2.5 V, which was anticipated to yield the maximum cutoff frequencies for both modulators under investigation. Regarding the large bundle modulator, the measured cutoff frequency ($f_{\mathrm{c}}$) was approximately 250 Hz, indicating its meaningful response to signals up to around 250 Hz. The response time, quantified as 635 μs, signifies the speed at which the modulator reacts to changes in the input signal. A shorter response time indicates a more favorable performance. In contrast, the cutoff frequency was lower for the small bundle modulator, around 190 Hz. The response time was slightly longer, as measured at 837 μs. Consequently, when subjected to the same applied modulation voltage, the large bundle IR modulator exhibited the following: A higher cutoff frequency enables it to respond more effectively to higher-frequency signals, and a shorter response time indicates a faster transient response. These findings align with intuitive expectations, considering the larger size of the large bundle modulator. Hence, the large bundle modulator exhibits better high-frequency performance than the small one. The crucial findings emphasize that bundle size is key in determining the cutoff frequency and response time.

[IMAGE OMITTED. SEE PDF]

The study also investigated the impact of varying the amplitude modulation voltage on the cutoff frequency of an IR modulator. The term “amplitude modulation” refers to altering the maximum voltage of the signal applied to the IR modulator. Experimental results demonstrated a linear relationship between the amplitude modulation voltage and the cutoff frequency of the IR modulator, specifically within the range of hundreds of hertz. This relationship is graphically represented in Figure . It is important to note that exceeding a particular voltage limit presents a substantial hazard of seriously damaging the IR modulator. Surpassing this threshold voltage could damage the IR modulator. Therefore, a maximum modulation voltage must be upheld to guarantee the modulator functions safely. Exceeding the maximum voltage could potentially compromise the modulator's integrity and operation. Maintaining a voltage below the threshold is necessary to avoid adverse effects on the performance and lifespan of the IR modulator device.

[IMAGE OMITTED. SEE PDF]

In contrast, the cutoff frequency is the highest frequency to which the IR modulator can effectively respond. Hence, an increased cutoff frequency implies that the modulator can accurately process higher-frequency signals. The observed linear correlation between the modulation voltage and cutoff frequency implies a direct cause-and-effect association between these two parameters, at least for the tested lower frequencies and voltage amplitudes. However, it is important to acknowledge that the linear trend may not persist at higher frequencies and voltages due to the potential for irreversible damage to the IR modulator beyond a certain point. The experimental findings validate the effectiveness of utilizing amplitude modulation of the input voltage to control the IR modulator's cutoff frequency response. This relationship holds for modest variations in modulation voltage and frequency.

Our study also aimed to explore and understand the correlation between the cutoff frequency and time constant of the IR modulator, utilizing the LED modulation technique, to confirm the frequency response curve shown in Figure . We have visually represented the relationship by plotting the cutoff frequency as a function of the time constant, as shown in Figure . Equation () describes the relationship between the cutoff frequency and the time constant. This equation suggests that knowledge of one parameter allows for predicting the other. We hypothesized an inverse relationship between the cutoff frequency and the time constant. Specifically, a shorter time constant indicates a faster response time and is likely associated with a higher cutoff frequency. Conversely, a longer time constant, indicating a slower response, corresponds to a lower cutoff frequency. Figure shows the characterization of an IR modulator's dynamic response and time constant. A voltage modulation signal of 2.5 V was applied to test the IR modulator's responsiveness. The modulator switched ON and OFF states at different frequencies (75, 85, 90, 150, and 160 Hz). The aim was to evaluate how quickly and dynamically the IR modulator could operate by determining its ability to turn ON and OFF at those frequencies. The time constant of the IR modulator was extracted from the results. The time constant was determined by measuring when the output voltage reached zero (when the IR modulator turned OFF). Then, measure when the output voltage reaches 63.2% of its final value (when the IR modulator is turned ON). The difference between these two times gives the time constant of the IR modulator. A shorter time constant indicates a faster response time since the output voltage reaches 63.2% of its final value more quickly. From our data analysis, it is evident that there's a significant relationship between the cutoff frequency and time constant of this specific IR modulator and its consistent performance over multiple cycles.

[IMAGE OMITTED. SEE PDF]

Conclusion

In this study, we successfully fabricated network films of SC- and MT-SWNTs dispersed in an aqueous surfactant solution for IR applications. Extensive testing allowed the determination of optimal fabrication conditions tailored for IR technology. We demonstrated modulation of the S₁₁ optical absorption band by applying a direct current voltage to an ionic liquid. The IR modulator exhibited optical hysteresis attributed to relaxation dynamics at the ionic liquid-SWNT interface during charging and discharging. Our thorough investigations conclusively determined EMIM-TFSI to be a higher-level ionic liquid for controlling SWIR light compared to DEME-TFSI. Larger SWNT bundles exhibited enhanced performance due to lower resistance, increased connections, and fewer junctions. These bundles have higher electronic state densities, enabling faster-infrared transmission changes and higher extinction ratios during modulation. The bundle size and porosity of the SWNT network film critically impacted modulator performance. Increased bundle size and porosity achieved optimal modulation depths responding to 250 Hz signals with an impressive 635 microsecond response time. Further enhancements may include lowering ionic liquid IR viscosity, increasing ionic conductivity, and reducing spacer thickness between electrodes. Modulation voltage significantly affects cutoff frequency, with shorter time constants enabling faster responses. Optimization of these parameters is critical for developing efficient optoelectronic devices. Reversible voltage modulation prevents ionic liquid degradation and modulator failure, preserving viability over repeated cycles. Our work establishes bundle size significantly impacts performance and develops methods to control bundle size reliably. This study demonstrates the potential for SWNTs in advanced IR applications.

Experimental Section

This study utilized a hybrid IR modulator comprising two layers of SWNTs semiconducting and metallic with exceptional conducting properties. The filtration membrane we used for the vacuum filtration was from MF-Millipore Membrane Filter, 0.1 and 0.05 μm pore size. The SWNT layers were securely affixed to glass substrates via transfer technique and then using a 250 μm thick adhesive frame (Frame-Seal Incubation Chambers, Bio-Rad Laboratories) as a spacer between the electrodes. The inner space of the modulator was filled with a specific ionic liquid known as EMIM-TFSI, which was procured from IoLiTec-Ionic Liquids Technologies GmbH. A SC-SWNTs film was used to construct the modulator's IR-active electrode obtained from Nanointegris Inc. – IsoNanotubes-S 99%. In contrast, the counter-electrode was fashioned from a thin film of highly isolated MT-SWNTs from Nanointegris Inc.– IsoNanotubes-M 99%. The fabrication process involved vacuum filtration of the SC-SWNT and MT-SWNT films, which were transferred onto glass substrates with predeposited Ti (50 nm)/Pt (250 nm) electrodes. The resulting configuration of Ti/Pt lines on opposite sides of the IR modulator created a small window for IR transmission. To study the surface features of SWNTs, we used the Veeco MultiMode model No. MMAFM-2 AFM in conjunction with the NanoScope IIIa controller. This AFM method uses a sharp tip to scan the sample and obtain information on its topography. For measuring the size of bundles of SWNTs, we employed the ZEISS-SUPRA 40VP SEM with a GEMINI Column. For analyzing the transmission characteristics of ionic liquids, we used the Thermo Fisher Scientific Nicolet 6700 FTIR Spectrometer, which measures IR light transmission. The absorption spectra of the SWNT thin film-based device were measured under varying applied voltages to evaluate the modulator's performance and behavior using a Cary 5000 UV–vis–NIR spectrophotometer from Agilent Technologies. A Keithley model 617 source meter supplied the required DC gate voltage. The dynamic properties of the modulator's response were examined using an IR LED (Model 1800P, Thorlabs) as a NIR radiation source with a wavelength close to 1800 nm, which aligned with the maximum absorption band (S₁₁) of the SC-SWNTs. A series of experiments were conducted to study the IR modulator. A square-wave voltage modulation generated by a function generator (Model DS345, Stanford Research Systems) was used to drive the modulator. This voltage modulation served as the modulator's input signal, allowing for the modulation of transmitted IR light. An InGaAs photodiode (Model FD10D, Thorlabs) was chosen to detect the modulated light for its suitability in capturing IR signals. The output signal from the photodiode was then connected to a lock-in amplifier (SRS 830, Stanford Research Systems). This lock-in amplifier allowed for precise measurements of the amplitude and frequency dependence of the IR response. This allowed for a detailed characterization of the modulator's performance. To capture the temporal traces of the frequency response, a Tektronix TDS 1001C-EDU oscilloscope was used. The oscilloscope was connected to the AC output of the lock-in amplifier. This enabled the visualization and recording of the modulated IR signal over time. This setup provided crucial data for analyzing the dynamic behavior and temporal characteristics of the IR modulator's response to different frequencies.

Acknowledgements

The authors thank the Ministry of Education in Saudi Arabia's Institutional Funding Program for Research and Development (grant no. IFPIP: 1372-130-1443) and the Deanship of Scientific Research (DSR) at King Abdulaziz University for their technical and financial support. The valuable contributions of those who helped make this project a success are also greatly appreciated.

Conflict of Interest

The authors declare no conflict of interest.

Data Availability Statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.

References

J. Olenewa, Guide to Wireless Communications, Cengage Learning, Boston, MA 2013.

A. Kumar, A. Goswami, K. Singh, R. McGee, T. Thundat, D. Kaur, ACS Appl. Electron. Mater. 2019, 1, 2226.

B. F. Jones, IEEE Trans. Med. Imaging 1998, 17, 1019.

X. Maldague, S. Marinetti, J. Appl. Phys. 1996, 79, 2694.

Z. Wang, W. Ren, Light: Sci. Appl. 2023, 12, 7.

M. Montesinos-Ballester, L. Deniel, N. Koompai, T. H. N. Nguyen, J. Frigerio, A. Ballabio, V. Falcone, X. Le Roux, C. Alonso‐Ramos, L. Vivien, A. Bousseksou, G. Isella, D. Marris‐Morini, ACS Photonics 2022, 9, 249.

J. Federici, L. Moeller, J. Appl. Phys. 2010, 107, 6.

M. A. Richardson, J. A. Coath, Opt. Laser Technol. 1998, 30, 137.

H. Gong, W. Li, G. Fu, Q. Zhang, J. Liu, Y. Jin, H. Wang, J. Mater. Chem. A 2022, 10, 6269.

Y. Duan, B. Liu, Adv. Mater. 2018, 30, 1802394.

J. Niu, Y. Wang, X. Zou, Y. Tan, C. Jia, X. Weng, L. Deng, Appl. Mater. Today 2021, 24, 101073.

J. Hu, Y. Hu, Y. Ye, R. Shen, Chem. Eng. J. 2023, 452, 139147.

B. Zeng, Z. Huang, A. Singh, Y. Yao, A. K. Azad, A. D. Mohite, A. J. Taylor, D. R. Smith, H. ‐T. Chen, Light: Sci. Appl. 2018, 7, 51.

F. Valdez, V. Mere, X. Wang, N. Boynton, T. A. Friedmann, S. Arterburn, C. Dallo, A. T. Pomerene, A. L. Starbuck, D. C. Trotter, A. L. Lentine, S. Mookherjea, Sci. Rep. 2022, 12, 18611.

P. Avouris, M. Freitag, V. Perebeinos, Nat. Photonics 2008, 2, 341.

M. Chernysheva, A. Rozhin, Y. Fedotov, C. Mou, R. Arif, S. M. Kobtsev, E. M. Dianov, S. K. Turitsyn, Nanophotonics 2017, 6, 1.

Y. Gladush, A. A. Mkrtchyan, D. S. Kopylova, A. Ivanenko, B. Nyushkov, S. Kobtsev, A. Kokhanovskiy, A. Khegai, M. Melkumov, M. Burdanova, M. Staniforth, J. Lloyd‐Hughes, A. G. Nasibulin, Nano Lett. 2019, 19, 5836.

E. Durán-Valdeiglesias, W. Zhang, C. Alonso-Ramos, S. Serna, X. Le Roux, D. Maris-Morini, N. Caselli, F. Biccari, M. Gurioli, A. Filoramo, E. Cassan, L. Vivien, Sci. Rep. 2018, 8, 11252.

S. Kruss, A. J. Hilmer, J. Zhang, N. F. Reuel, B. Mu, M. S. Strano, Adv. Drug Delivery Rev. 2013, 65, 1933.

E. Jónsson, Energy Storage Mater. 2020, 25, 827.

P. J. Lynch, M. Tripathi, A. Amorim Graf, S. P. Ogilvie, M. J. Large, J. Salvage, A. B. Dalton, ACS Appl. Mater. Interfaces 2023, 15, 11225.

T. Fukushima, T. Aida, Chem.–A Eur. J. 2007, 13, 5048.

Y. Sun, H. Chang, J. Hu, Y. Wang, Y. Weng, C. Zhang, S. Niu, L. Cao, Z. Chen, N. Guo, J. Liu, J. Chi, G. Li, L. Xiao, Adv. Opt. Mater. 2021, 9, 2001216.

M. L. Moser, G. Li, M. Chen, E. Bekyarova, M. E. Itkis, R. C. Haddon, Nano Lett. 2016, 16, 5386.

D. Stekovic, B. Arkook, G. Li, W. Li, E. Bekyarova, M. E. Itkis, Adv. Mater. Interfaces 2018, 5, 1800861.

A. H. Brozena, M. Kim, L. R. Powell, Y. Wang, Nat. Rev. Chem. 2019, 3, 375.

N. Anzar, R. Hasan, M. Tyagi, N. Yadav, J. Narang, Sens. Int. 2020, 1, 100003.

G. Speranza, Nanomaterials 2021, 11, 967.

L. C. Qin, X. Zhao, K. Hirahara, Y. Miyamoto, Y. Ando, S. Iijima, Nature 2000, 408, 50.

A. Takakura, K. Beppu, T. Nishihara, A. Fukui, T. Kozeki, T. Namazu, Y. Miyauchi, K. Itami, Nat. Commun. 2019, 10, 3040.

M. F. De Volder, S. H. Tawfick, R. H. Baughman, A. J. Hart, Science 2013, 339, 535.

V. I. Artyukhov, E. S. Penev, B. I. Yakobson, Nat. Commun. 2014, 5, 4892.

J. Charoenpakdee, O. Suntijitrungruang, S. Boonchui, Sci. Rep. 2020, 10, 18949.

N. Manikandan, V. P. Suresh Kumar, S. Siva Murugan, G. Rathis, K. Vishnu Saran, T. K. Shabariganesh, Mater. Today: Proc. 2021, 47, 4682.

P. W. Barone, S. Baik, D. A. Heller, M. S. Strano, Nat. Mater. 2005, 4, 86.

A. Janissek, J. Lenz, F. D. Giudice, M. Gaulke, F. Pyatkov, S. Dehm, F. Hennrich, L. Wei, Y. Chen, A. Fediai, M. Kappes, W. Wenzel, R. Krupke, R. Thomas Weitz, Appl. Phys. Lett. 2021, 118, 063101.

K. Chatterjee, A. D. Pathak, A. Lakma, C. S. Sharma, K. K. Sahu, A. K. Singh, Sci. Rep. 2020, 10, 9606.

S. Mallakpour, Z. Rafiee, J. Polym. Environ. 2011, 19, 447.

H. Shimada, K. Ono, Y. Ootuka, J. Appl. Phys. 2003, 93, 8259.

B. Zhao, I. Mönch, H. Vinzelberg, T. Mühl, C. M. Schneider, Appl. Phys. Lett. 2002, 80, 3144.

Z. Wu, Z. Chen, X. Du, J. M. Logan, J. Sippel, M. Nikolou, K. Kamaras, J. R. Reynolds, D. B. Tanner, A. F. Hebard, A. G. Rinzler, Science 2004, 305, 1273.

Q. Cheng, S. Debnath, E. Gregan, H. J. Byrne, J. Phys. Chem. C 2010, 114, 8821.

C. L. Pint, N. W. Nicholas, S. Xu, Z. Sun, J. M. Tour, H. K. Schmidt, R. G. Gordon, R. H. Hauge, Carbon 2011, 49, 4890.

J. C. Blancon, M. Paillet, H. N. Tran, X. T. Than, S. A. Guebrou, A. Ayari, A. S. Miguel, N.‐M. Phan, A.‐A. Zahab, J. L. Sauvajol, N. Del Fatti, F. Vallée, Nat. Commun. 2013, 4, 2542.

S. A. Hussain, Sci. Rep. 2019, 9, 17282.

I. Kuusik, M. Kook, R. Pärna, A. Kivimäki, T. Käämbre, L. Reisberg, A. Kikas, V. Kisand, J. Mol. Liq. 2019, 294, 111580.

K. C. Lethesh, A. Bahaa, M. Abdullah, M. O. Bamgbopa, R. A. Susantyoko, Front. Chem. 2022, 10, 556.

S. Winder, J. Carr, Newnes Radio and RF Engineering Pocket Book, Newnes, Woburn, MA 2002.

M. Belotti, X. Lyu, L. Xu, P. Halat, N. Darwish, D. S. Silvester, C. Goh, E. I. Izgorodina, M. L. Coote, S. Ciampi, J. Am. Chem. Soc. 2021, 143, 17431.

Y. Zhu, W. Xie, S. Gao, F. Zhang, W. Zhang, Z. Liu, J. Jin, Small 2016, 12, 5034.

E. Senokos, D. B. Anthony, N. Rubio, M. C. Ribadeneyra, E. S. Greenhalgh, M. S. Shaffer, Adv. Funct. Mater. 2023, 33, 2212697.

V. Egorov, C. O’Dwyer, Curr. Opin. Electrochem. 2020, 21, 201.

A. R. Hambley, N. Kumar, A. R. Kulkarni, Electrical Engineering: Principles and Applications, Prentice Hall, New Jersey 2013.

G. Long, W. Jin, F. Xia, Y. Wang, T. Bai, X. Chen, X. Liang, L.‐M. Peng, Y. Hu, Nat. Commun. 2022, 13, 6734.

E. X. Ding, Q. Zhang, N. Wei, A. T. Khan, E. I. Kauppinen, R. Soc. Open Sci. 2018, 5, 180392.

S. Jiang, P. X. Hou, C. Liu, H. M. Cheng, J. Mater. Sci. Technol. 2019, 35, 2447.

E. X. Ding, A. Hussain, S. Ahmad, Q. Zhang, Y. Liao, H. Jiang, E. I. Kauppinen, Nano Res. 2020, 13, 112.

N. Yamanaka, High-Performance Backbone Network Technology, CRC Press, Boca Raton, FL 2004.