A low-damage cleaning method for carbon nanotube wafers based on substrate material properties

Abstract

The application discloses a low-damage cleaning method for carbon nanotube wafers based on substrate material properties. The method is used for carbon nanotube wafers with residual polymers after polymer-assisted separation, dispersion or assembly. The substrate material properties are obtained, and a rapid thermal processing annealing atmosphere window is selected according to the properties. The window is different for different substrates, and at least includes oxygen-containing atmosphere treatment parameters and inert atmosphere replacement parameters. Oxygen-containing atmosphere treatment and inert atmosphere replacement are sequentially performed in a rapid thermal processing annealing furnace. The residual O2 is lower than the equipment safety threshold at the exhaust end, and H2 / N2 atmosphere treatment can be optionally performed. The high-resistance silicon, SiO2 / Si and HfO2 / Si substrates respectively adopt a window with different adjustment directions, such as medium-high temperature and low oxygen, low temperature and high oxygen for a short time, or medium-high temperature and low oxygen, and a moderate controlled window. The application can reduce the polymer residue to about 1% to 2%, convert the carbon nanotube into a measurable conduction state, and inhibit ablation, bundling and dielectric layer leakage degradation.

Description

Technical Field

[0001] This invention relates to the field of semiconductor manufacturing technology, and specifically to a low-damage cleaning method for carbon nanotube wafers based on the properties of the substrate material. Background Technology

[0002] Carbon nanotube materials produced by polymer-assisted separation, purification, and wafer-level assembly often contain residual conjugated polymers, non-conjugated polymers, surfactants, or organic solvents. These residues can affect metal contacts, dielectric layer deposition, interface state density, threshold stability, device leakage current, and long-term reliability. Therefore, a cleaning process needs to be established in the device fabrication or wafer post-processing procedures.

[0003] For the removal of polymers or dispersants from the surface of carbon nanotubes, various routes have been reported in this field, including solvent immersion and weak acid protonation, hydrogen plasma reduction, inert atmosphere or vacuum furnace annealing, and rapid heating and cooling annealing. These routes can remove residues to varying degrees under their respective conditions, but their treatment conditions are independent of the substrate. Existing annealing cleaning processes often use a uniform temperature, time, or atmosphere to treat carbon nanotube samples on different substrates, ignoring the differences in substrate material properties. At the same time, the ability of inert or vacuum atmospheres to oxidize and decompose the conjugated polymer backbone tightly bound to carbon nanotubes is limited, while introducing an oxygen-containing atmosphere without distinguishing the substrate makes it difficult to avoid oxidative damage to the heat-sensitive dielectric layer and the electrical properties of the device. In reality, substrates such as high-resistivity silicon, SiO2 / Si, and HfO2 / Si exhibit significant differences in thermal conductivity, surface chemistry, oxygen adsorption behavior, reduction sensitivity, and carbon nanotube adhesion mechanisms: silicon has a thermal conductivity of approximately 149 W / (m·K), SiO2 approximately 1.4 W / (m·K), and HfO2 approximately 1.1 W / (m·K). For composite wafer substrates, carbon nanotubes actually contact the surface oxide or dielectric layer. Local thermal diffusion, interfacial oxygen activity, and surface adsorption states cannot be solely determined by the silicon wafer itself. Low thermal conductivity oxide layers or high dielectric constant dielectric layers may cause localized heat accumulation and different surface reaction pathways. Therefore, the same annealing conditions may effectively remove residues on one substrate but may cause carbon nanotube ablation, breakage, bundling, stacking, or interfacial electrical degradation on another substrate. Furthermore, the damage modes of carbon nanotube arrays and carbon nanotube network films differ: carbon nanotube arrays are more concerned with alignment maintenance, bundle formation, stacking, and localized ablation, while carbon nanotube network films are more concerned with network connectivity, node contacts, resistance drift, and localized fracture. Providing only a single annealing condition is insufficient to cover the requirements of wafer-level carbon nanotube processes. Nitrogen-containing reactive atmospheres such as ammonia may also react with defect sites or edge sites of carbon nanotubes, introducing carbon-nitrogen bonds, nitrogen-related defects, or doping effects, thereby altering the intrinsic electrical properties of carbon nanotubes.

[0004] Therefore, existing technologies still have the drawback of failing to simultaneously achieve polymer residue removal and low-damage protection of carbon nanotubes. It is necessary to propose a semiconductor-compatible cleaning method that can design rapid thermal annealing windows based on the properties of the substrate material. Summary of the Invention

[0005] The present invention aims to provide a low-damage cleaning method for carbon nanotube wafers based on the properties of the substrate material, so as to overcome the shortcomings of the prior art. The technical problem to be solved by the present invention is achieved through the following technical solution.

[0006] This invention provides a low-damage cleaning method for carbon nanotube wafers based on the properties of the substrate material. This method removes polymer residues remaining on the surface of carbon nanotube wafers after polymer-assisted separation, dispersion, or assembly. The specific steps are as follows: Obtain the material properties of the substrate supporting the carbon nanotubes; select a rapid heat treatment annealing atmosphere window according to the material properties of the substrate. Different rapid heat treatment annealing atmosphere windows correspond to different substrates with different material properties. The rapid heat treatment annealing atmosphere window includes at least oxygen-containing atmosphere treatment parameters and inert atmosphere replacement parameters. In a rapid heat treatment annealing furnace, the carbon nanotube wafer is subjected to oxygen-containing atmosphere treatment and inert atmosphere replacement sequentially according to the selected rapid heat treatment annealing atmosphere window. The inert atmosphere replacement is carried out until the residual O2 concentration at the exhaust end is lower than the equipment safety threshold, so as to reduce carbon nanotube damage while removing polymer residues on the surface of carbon nanotubes.

[0007] Furthermore, the material properties of the substrate include one or more of the following: thermal conductivity, thermal diffusivity, surface oxide composition, surface hydroxyl density, oxygen adsorption and release capacity, reduction sensitivity, dielectric layer stability, and adhesion strength between the carbon nanotubes and the substrate.

[0008] Furthermore, for substrates that can withstand high annealing temperatures and do not have heat-sensitive dielectric layers on their surface, a rapid heat treatment annealing atmosphere window with a high annealing temperature and a low O2 volume fraction is selected. For substrates with a thermally oxidized SiO2 layer on the surface, select a rapid heat treatment annealing atmosphere window with a lower annealing temperature and a higher O2 volume fraction, or a rapid heat treatment annealing atmosphere window with a higher annealing temperature and a lower O2 volume fraction. For substrates with a high dielectric constant dielectric layer on the surface, a rapid heat treatment annealing atmosphere window with a lower annealing temperature and controlled O2 volume fraction and holding time is selected.

[0009] Furthermore, the substrate is a high-resistivity silicon substrate, a SiO2 / Si substrate, a HfO2 / Si substrate, an Al2O3 / Si substrate, a ZrO2 / Si substrate, a Si3N4 / Si substrate, a quartz substrate, or a sapphire substrate. The oxygen-containing atmosphere used in the oxygen-containing atmosphere treatment is an O2 / N2 mixture, an O2 / Ar mixture, dry air, diluted O2, or pure O2, and the volume fraction of O2 in the oxygen-containing atmosphere is 0.5% to 100%. Further, the heating rate of the annealing treatment is 5~100 ℃ / s; the heating rate is 10~50 ℃ / s; or, when the substrate is an HfO2 / Si substrate, the heating rate is 5~30 ℃ / s.

[0010] Furthermore, the inert atmosphere replacement is carried out by introducing N2 or Ar, and the replacement time is determined based on the effective cavity volume of the rapid heat treatment annealing furnace, the dead volume of the pipeline, the set total flow rate, and the O2 concentration at the exhaust end; the replacement time is 30~180 s.

[0011] Furthermore, the rapid heat treatment annealing atmosphere window also includes H2 / N2 atmosphere treatment parameters; after the inert atmosphere replacement, the carbon nanotube wafer is subjected to H2 / N2 atmosphere treatment, wherein the volume fraction of H2 in the H2 / N2 atmosphere is 0.5%~5%.

[0012] Furthermore, the volume fraction of H2 in the H2 / N2 atmosphere is 3%~4%, and the H2 / N2 atmosphere treatment temperature is 350~520 ℃, and the time is 0.5~3 min.

[0013] Furthermore, when the substrate is a high-resistivity silicon substrate, the oxygen-containing atmosphere treatment adopts a medium-high temperature low oxygen window with an O2 volume fraction of 1%~5%, a temperature of 580~620 ℃, and a time of 2.5~5 min, or a mild pre-screening window with a temperature of 430~530 ℃ and a time of 1~4 min. When the substrate is a SiO2 / Si substrate, the oxygen atmosphere treatment adopts a low temperature and high oxygen short time window with an O2 volume fraction of 20%~100%, a temperature of 300~400 ℃, and a time of 0.5~3 min, or adopts a medium temperature and high temperature and low oxygen window with an O2 volume fraction of 0.5%~5%, a temperature of 500~650 ℃, and a time of 2~7 min. When the substrate is an HfO2 / Si substrate, the oxygen-containing atmosphere treatment adopts a window of O2 volume fraction of 5%~40%, temperature of 350~450 ℃, and time of 1~5 min.

[0014] Further, the carbon nanotubes are carbon nanotube arrays, carbon nanotube network films, or a combination of both, wherein the carbon nanotube arrays include wafer-level carbon nanotube arrays; when the carbon nanotubes are carbon nanotube arrays, compared to carbon nanotube network films, at least one of the following is used: reducing the volume fraction of O2 in the oxygen-containing atmosphere, shortening the holding time, and reducing the heating rate; when the carbon nanotubes are carbon nanotube network films, the oxidation heat budget is increased while maintaining network connectivity; and one or more of the following are used as criteria for low-damage cleaning effect: polymer residue ratio, G / D change, polarized Raman intensity ratio, alignment angle, sheet resistance, two-point resistance, contact resistance, off-state current, threshold drift, dielectric layer leakage current, and microstructure.

[0015] The embodiments of the present invention have the following advantages: (1) Introducing the properties of the substrate material into the window design of the carbon nanotube cleaning process is different from the traditional annealing method with uniform temperature, uniform time or uniform atmosphere. The same method framework can achieve a balance between residue removal and low damage on different substrates such as high-resistivity silicon, SiO2 / Si, and HfO2 / Si. Moreover, the adjustment directions of the window for the three representative substrates in terms of annealing temperature and O2 volume fraction are different. The window selection rule cannot be obtained by monotonically extrapolating from uniform conditions. Thus, the scattered empirical annealing conditions are transformed into a semiconductor-compatible process system that can be replicated, transferred and optimized by feedback.

[0016] (2) The rapid heat treatment annealing furnace achieves high controllable heating and short-term heat budget, which is conducive to completing the removal of polymer residues in a short time and reducing the damage caused by long-term heat exposure. The polymer residue can be reduced from 100% to about 1%~2%, and the carbon nanotube array and carbon nanotube network film are transformed from a stable resistance that cannot be measured by conventional resistance meters to a measurable conductive state. The two-point resistance is reduced to about 0.5 MΩ and about 0.1 MΩ, respectively.

[0017] (3) By combining and controlling the volume fraction of O2, temperature, time and heating rate, polymer residue removal can be achieved with a lower total heat budget; by replacing the oxidation step with N2 or Ar, the risk of oxidation overshoot and atmosphere switching can be reduced.

[0018] (4) By using short-time H2 / N2 post-treatment, the effects of adsorbed oxygen or weakly bound oxygen residues can be reduced on some substrates; for sensitive substrates such as HfO2 / Si, this step can be reduced by lowering the temperature, shortening the time or omitting it, thereby reducing the risk of interface states, resistance drift and device leakage on the dielectric substrate.

[0019] (5) By classifying and evaluating carbon nanotube arrays and carbon nanotube network films, bundle formation, stacking, arrangement disruption, network breakage, node contact degradation and electrical drift can be controlled respectively. Bundle formation, stacking and arrangement disruption can be suppressed in carbon nanotube arrays, and network connectivity and node contact stability can be maintained in carbon nanotube network films.

[0020] (6) This method is suitable for expansion to different wafer sizes such as 2-inch, 4-inch, 8-inch and 12-inch, and can be converted using equivalent thermal budget based on the actual wafer temperature curve and atmosphere switching response of the rapid thermal processing equipment. Attached Figure Description

[0021] Figure 1 A flowchart for selecting the annealing atmosphere window for carbon nanotubes based on the properties of the substrate material, provided for embodiments of the present invention; Figure 2 A comparative schematic diagram of rapid heat treatment annealing windows for different representative substrates provided in embodiments of the present invention; Figure 3 This is a schematic diagram of the time-temperature and atmosphere switching sequence during rapid heat treatment annealing provided in an embodiment of the present invention. Detailed Implementation

[0022] It should be noted that, unless otherwise specified, the embodiments and features described in this application can be combined with each other. The present invention will now be described in detail with reference to the accompanying drawings and embodiments.

[0023] like Figure 1As shown, the low-damage cleaning method for carbon nanotube wafers provided in this embodiment of the invention is executed according to the following process: sample input S1, material property determination S2, window selection S3, annealing S4, and result evaluation and feedback S5. Sample input S1 includes substrate type, carbon nanotube morphology, and polymer residue status; material property determination S2 is used to obtain or determine the material properties of the substrate supporting the carbon nanotubes. These material properties include, but are not limited to, thermal conductivity, thermal diffusivity, surface oxide composition, surface hydroxyl density, oxygen adsorption and release capacity, reduction sensitivity, dielectric layer stability, adhesion strength between carbon nanotubes and the substrate, as well as carbon nanotube density, carbon nanotube morphology, and polymer residue level. The material properties of the substrate can be determined based on the substrate type and dielectric layer structure information of the wafer, or a pre-screening annealing test can be performed on the wafer. After being categorized into similar substrate types, the window selection (S3) determines the rapid heat treatment annealing atmosphere window based on the aforementioned material properties. The window includes at least the rapid heat treatment temperature, holding time, heating rate, O2 volume fraction, inert atmosphere replacement time, H2 volume fraction in the H2 / N2 atmosphere, H2 / N2 treatment temperature, and treatment time. Annealing (S4) is performed in the order of oxidation step, intermediate replacement, and optional H2 / N2 post-treatment. Result evaluation and feedback (S5) optimizes the window based on indicators such as polymer residual ratio, G / D, resistance, morphology, leakage current, and threshold drift.

[0024] In this embodiment, the processing atmosphere consists of three modules: an oxidizing residue removal module, an inert isolation module, and an H2 / N2 post-treatment module. The oxidizing residue removal module uses one of the following: an O2 / N2 mixture, an O2 / Ar mixture, dry air, diluted O2, or pure O2, with an O2 volume fraction of 0.5% to 100%; the inert isolation module uses N2, Ar, or other inert gases; the H2 / N2 post-treatment module uses an H2 / N2 mixture with an H2 volume fraction of 0.5% to 5%, and in some embodiments, the H2 volume fraction is 3% to 4%. In all embodiments of the present invention, the oxygen-containing atmosphere is a mixture of O2 and N2, a mixture of O2 and Ar, diluted O2, or pure O2. Ammonia or nitric oxide is not used as a cleaning atmosphere to avoid the reaction between nitrogen-containing active species and the defect sites or edge sites of carbon nanotubes, which could introduce carbon-nitrogen bonds, nitrogen-related defects, or doping effects, thereby altering the intrinsic electrical properties of carbon nanotubes. Even though ammonia is used as a hydrogen source in some existing processing routes, the present invention still excludes ammonia and nitric oxide from the rapid thermal annealing atmosphere window, and the H2 / N2 post-treatment uses only a mixture of H2 and N2 as the gas source.

[0025] The material properties of different substrates dictate different window design principles. Silicon substrates have a thermal conductivity of approximately 149 W / (m·K) and exhibit rapid thermal diffusion; SiO2 has a thermal conductivity of approximately 1.4 W / (m·K), and HfO2 has a thermal conductivity of approximately 1.1 W / (m·K). For composite substrates such as SiO2 / Si and HfO2 / Si, the carbon nanotubes actually contact a low-thermal-conductivity surface oxide or dielectric layer, resulting in significantly different local thermal diffusion, interfacial oxygen activity, and surface adsorption states compared to high-resistivity silicon. Figure 2 As shown in the table below, the design principles for the three types of representative bases are as follows.

[0026]

[0027] like Figure 3 As shown, the typical process sequence of this embodiment is as follows: Step 1: Place the wafer with a carbon nanotube array or carbon nanotube network film into the reaction chamber of a rapid thermal annealing furnace. Unlike tube furnaces, rapid thermal annealing furnaces feature rapid heating, quick temperature tracking of the set value, smaller effective chamber volume, and faster atmosphere switching response. Step 2: Introduce N2 or Ar for pre-replacement or pre-flow. The pre-replacement time is determined based on the effective chamber volume of the rapid thermal annealing furnace, the dead volume of the pipeline, and the O2 concentration at the exhaust end, typically 30-120 s. Step 3: Heat at a rate of 5-100 °C / s, or 10-50 °C / s in some embodiments; for sensitive substrates such as HfO2 / Si, a milder heating rate of 5-30 °C / s can be used. Step 4: Introduce an O2 atmosphere for the first annealing step, i.e., oxidation. The O2 volume fraction, temperature, and time are selected according to the properties of the substrate material. Step 5: Switch to N2 or Ar for intermediate purging (i.e., inert atmosphere purging), with a purging time of 30–180 s, or continue purging until the O2 concentration at the exhaust end is below the equipment safety threshold. Step 6: If necessary, introduce H2 / N2 for a second-step annealing or post-treatment. Step 7: After cooling to a safe temperature in an N2 or Ar atmosphere, remove the sample and evaluate the balance between cleanliness and damage using Raman spectroscopy, atomic force microscopy, scanning electron microscopy, electrical testing, or device parameter analysis.

[0028] It should be noted that, since this method uses a rapid heat treatment annealing furnace, the atmosphere switching time is not directly extrapolated from the volume of the tubular furnace, but should be determined based on the effective cavity volume of the rapid heat treatment annealing furnace, the inlet and outlet structure, the dead volume of the pipeline, the set total flow rate, the O2 monitoring results at the exhaust end, and the equipment safety interlock requirements. For the H2 / N2 step, the residual O2 should be reduced to below the equipment safety threshold by N2 or Ar before introducing H2 to comply with the equipment safety specifications.

[0029] The main process parameter ranges of this invention are shown in the table below.

[0030]

[0031] The testing and evaluation methods common to all embodiments are described below. The polymer residue ratio can be estimated by the polymer characteristic peak intensity, the organic residue signal after background subtraction, or other equivalent spectral indicators: with the polymer residue signal of the sample before annealing as 100%, and the background of the thoroughly cleaned substrate or a low-residue sample as the baseline, the residue ratio is calculated as: (polymer characteristic signal after annealing - baseline signal) / (polymer characteristic signal before annealing - baseline signal) × 100%. This invention uses the intensity ratio of the G peak to the D peak in Raman spectroscopy, i.e., G / D, as a structural indicator. Unannealed carbon nanotube samples contain significant polymer coatings and organic residues. These residues affect both the G and D peaks in Raman testing, with a stronger apparent contribution to the G peak. Therefore, unannealed samples may exhibit a high apparent G / D, typically around 40. After annealing, as the polymer residues are removed, the G / D decreases significantly. This decrease does not necessarily indicate structural damage to the carbon nanotubes, but rather a gradual shift in the apparent Raman signal from being dominated by polymer contributions to being dominated by the intrinsic signal of the carbon nanotubes. Whether damage has occurred requires a comprehensive assessment considering polymer residues, microstructure, and resistance changes. Before annealing, due to polymer coatings and contact obstruction, the two-point resistance of carbon nanotube samples typically cannot be measured using conventional resistance meters, often exhibiting an over-range or open-circuit state. After annealing, the two-point resistance of carbon nanotube arrays typically drops to the order of approximately 0.5 MΩ, and for carbon nanotube network films, due to the presence of multiple conductive pathways, the two-point resistance typically drops to the order of approximately 0.1 MΩ. The two-point resistance method is used for comparative analysis before and after a process under the same electrode spacing, test structure, and test conditions, and is not limited to the intrinsic resistivity of carbon nanotube materials.

[0032] Example 1: Cleaning of carbon nanotube wafers on a high-resistivity silicon substrate.

[0033] The polymer-dispersed or assembled carbon nanotube samples are placed on a high-resistivity silicon wafer. High-resistivity silicon has high thermal conductivity and rapid heat diffusion, making it suitable as a representative substrate with relatively high thermal diffusivity. It can be used in rapid heat treatment windows with medium to high temperatures, low oxygen concentrations, and short holding times. Specifically, N2 or Ar is first introduced for pre-replacement at a flow rate set by the rapid heat treatment equipment for 30-120 s, or until the O2 at the exhaust end reaches the initial level set by the equipment. Then, the temperature is increased at a rate of 10-50 ℃ / s, and in this embodiment, 20-50 ℃ / s is used. The oxidation step uses an O2 / N2 mixture with an O2 volume fraction of 1%-5%, a temperature of 580-620 ℃, and a time of 2.5-5 min. Mild pre-screening can be carried out at 430-530 ℃ for 1-4 min. After the oxidation step, N2 or Ar is introduced for intermediate replacement for 30-180 s, or until the residual O2 is lower than the equipment safety threshold. Optional post-treatment uses an H2 / N2 mixture with an H2 volume fraction of 3%-4%, a temperature of 430-520 ℃, and a time of 0.5-3 min. Evaluation indicators include reduction of polymer characteristic peaks, G / D falling back to the target range and not below the damage threshold, and microstructure and two-point resistivity changes being within acceptable ranges. Table 1 lists a comparison of key indicators on high-resistivity silicon substrates before annealing and after different annealing conditions, where the heating rate for each annealing condition is approximately 30 °C / s.

[0034] Table 1 Data from high-resistivity silicon substrate examples

[0035] As shown in Table 1, the apparent G / D ratio of the samples on the high-resistivity silicon substrate before annealing was high, mainly due to polymer residue. After annealing, the polymer residue could be reduced from 100% to about 1%~2%, and the two-point resistivity changed from unmeasurable to about 0.5MΩ, indicating that rapid heat treatment with controlled O2 concentration can effectively remove insulating polymer residue and improve the conductivity path. When the holding time was extended to 7 min, the residue could be further reduced, but the resistance increased and the signs of local thermal stress increased. Therefore, a holding time of 2.5~5 min at 600 °C and 1% O2 volume fraction can achieve a balance between residue removal and low damage on the high-resistivity silicon substrate.

[0036] Example 2: Cleaning of carbon nanotube wafers on SiO2 / Si substrate.

[0037] Carbon nanotube samples are placed on silicon wafers with thermal oxide layers or deposited SiO2 layers. The surface of this type of substrate is a low thermal conductivity oxide layer with more surface hydroxyl groups and polar sites. The polymer residues and carbon nanotube adhesion states are different from those of high-resistivity silicon, and two equivalent windows can be used: low temperature and high oxygen for short time or medium temperature and low oxygen for longer time. The low-temperature, high-oxygen, short-time window uses an O2 volume fraction of 20%–100%, a temperature of 300–400 °C, a time of 0.5–3 min, and a heating rate of 10–50 °C / s for rapid oxidation of polymer residues. The initial process window is 300 °C for 1–2 min. The medium-high-temperature, low-oxygen window uses an O2 / N2 mixture with an O2 volume fraction of 0.5%–5%, a temperature of 500–650 °C, a time of 2–7 min, and a heating rate of 10–50 °C / s to reduce oxidative activity and improve the degree of thermally driven residue removal. Optional post-treatment uses an H2 / N2 mixture with an H2 volume fraction of 3%–4%, a temperature of 400–500 °C, and a time of 0.5–3 min to adjust the effect of adsorbed oxygen or weakly bound oxygen residues, and this should be confirmed by electrical indicators. Table 2 lists a comparison of key performance indicators for carbon nanotube arrays on SiO2 / Si substrates, where the heating rate for each annealing condition is approximately 30 °C / s.

[0038] Table 2 Data from SiO2 / Si substrate examples

[0039] As shown in Table 2, significant residue removal can be achieved on SiO2 / Si substrates under both low-temperature, high-oxygen, short-time and medium-high-temperature, low-oxygen conditions. Compared to high-resistivity silicon, the oxide layer on the SiO2 / Si surface diffuses thermally more slowly, and prolonged heat treatment is more likely to cause an increase in resistance and changes in the interface state. Therefore, for SiO2 / Si substrates, it is advisable to select 400 °C, pure O2, 2 min, or 600 °C, 1% O2 volume fraction, 2.5 min as the equilibrium window, rather than using prolonged medium-high temperature treatment.

[0040] Example 3: Cleaning of carbon nanotube wafers on HfO2 / Si substrate.

[0041] Carbon nanotube samples are placed on an HfO2 / Si substrate. The HfO2 dielectric layer can be a high-dielectric-constant dielectric layer formed by atomic layer deposition or other methods, with a thickness of 2–50 nm, and in some embodiments 5–20 nm. HfO2 / Si is sensitive to oxygen vacancies, interface states, and reducing atmospheres; therefore, lower temperatures, shorter processing times, and controlled oxygen activity windows are employed. Specifically, N2 or Ar is first introduced for pre-purification at the flow rate set by the rapid heat treatment equipment for 30-120 s; then the temperature is increased at a rate of 5-30 ℃ / s, which is 10-30 ℃ / s in this embodiment; the oxidation step uses an O2 / N2 mixture with an O2 volume fraction of 5%-40%, a temperature of 350-450 ℃, and a time of 1-5 min, where an O2 volume fraction of 5%-20%, a temperature of 380-420 ℃, and a time of 1-3 min are recommended starting points; after the oxidation step, N2 or Ar is introduced for intermediate purification, and the purification time is determined based on the effective cavity volume, total flow rate, and O2 concentration at the exhaust end of the rapid heat treatment equipment, usually 30-180 s; optional post-treatment uses an H2 / N2 mixture with an H2 volume fraction of 3%-4%, a temperature of 350-430 ℃, and a time of 0.5-2 min. This step can be omitted if leakage, threshold drift, or D peak increase occurs. Evaluation metrics include dielectric leakage current, threshold drift, D-peak variation, polymer characteristic peak retention, and morphological continuity. Table 3 lists the data of the examples on the HfO2 / Si substrate, and uses the same high-resistivity silicon window processing conditions as Comparative Example 1 to illustrate the different protection windows of HfO2 / Si compared to high-resistivity silicon and SiO2 / Si.

[0042] Table 3. Data and comparative examples of HfO2 / Si substrates

[0043] As shown in Table 3, the HfO2 / Si substrate can reduce polymer residue to approximately 1%–2% under relatively mild conditions. Without specifying a particular mechanism, oxygen vacancies, oxygen adsorption sites, or oxygen exchange sites on the HfO2 surface may promote the oxidative cracking or desorption of organic residues, thus achieving lower residues under relatively mild oxygen-containing conditions. This effect is not limited to a specific mechanism. On the other hand, HfO2 / Si is more sensitive to interface states and dielectric leakage. Comparative Example 1 shows that when the 600 °C, 1% O2 volume fraction, and 2.5 min window, suitable for high-resistivity silicon, are directly applied to HfO2 / Si substrates, although the residue can be reduced to about 1%, the leakage current increases to more than 3 times and the threshold drift reaches about 0.20 V, which is significantly worse than the results of about 1.1 times leakage current and about 0.03 V threshold drift under the mild window. This indicates that uniform annealing conditions cannot meet the low-damage requirements of different substrates. That is, cleaning HfO2 / Si substrates according to the idea of ​​uniform treatment will only result in damaging results. This comparison constitutes a reverse example of the uniform annealing route. In addition, the apparent G / D of about 40 before annealing is mainly contributed by the Raman signal of the polymer. The change in apparent G / D after annealing mainly reflects the removal of polymer residues and does not necessarily represent the deterioration of carbon nanotube structure. It should be judged in combination with electrical indicators such as dielectric leakage current and threshold drift. Based on comprehensive balance, the HfO2 / Si substrate adopts a mild oxygen-containing window of 380~420 ℃, O2 volume fraction of 5%~20%, and 1~3 min, and the H2 / N2 post-processing is omitted based on leakage current and threshold drift.

[0044] Example 4: Differential treatment of carbon nanotube arrays and carbon nanotube network films.

[0045] For carbon nanotube arrays, reducing the oxidation step time and oxygen activity, and avoiding the bundling and stacking of high-density adjacent carbon nanotubes under thermal, capillary, or electrostatic effects, allows for a suitable reduction in the heating rate within the limits of rapid heat treatment equipment, exemplarily 5–30 °C / s. For carbon nanotube network films, due to their inherent random overlap and node structure, the oxidation heat budget can be appropriately increased while maintaining network connectivity. The main risks of carbon nanotube arrays are bundling, stacking, widening of the alignment angle, and localized ablation. Evaluation indicators include the intensity ratio of the G peak to the D peak, alignment angle, microstructure, contact resistance, and off-state current. The main risks of carbon nanotube network films are network breakage, node contact degradation, and sheet resistance drift. Evaluation indicators include sheet resistance, two-terminal resistance, network connectivity, D peak variation, and area uniformity, with particular emphasis on controlling localized overburning. Table 4 lists the differential data for the two morphologies under the same representative annealing conditions.

[0046] Table 4. Differentiation data between carbon nanotube arrays and network thin films

[0047] As shown in Table 4, the same residual removal window produces different electrical results for carbon nanotube arrays and carbon nanotube network films. The low-damage criterion for array samples should not be based solely on the reduction in resistance, but should also be judged in combination with alignment preservation, beamforming, and microstructure; while for network films, sheet resistance or two-point resistance stability is more suitable as the main screening index.

[0048] Based on the above embodiments, the windowing conclusions obtained from the embodiment data are shown in Table 5.

[0049] Table 5 Windowing Conclusions Based on Example Data

[0050] The above data shows that the technical effect of this invention does not come from a single annealing temperature, but from a rapid heat treatment atmosphere window determined by the properties of the substrate material and the sample morphology. A high G / D ratio before annealing cannot be directly used as evidence of low defects; after annealing, the G / D ratio decreases, polymer residue is reduced to about 1%~2%, the resistance changes from unmeasurable to measurable conductive state, and the microstructure remains continuous, which together constitute the quantitative basis for the low-damage cleaning effect of this invention.

[0051] This invention is not limited to the three types of substrates mentioned above. For other dielectric / semiconductor composite substrates such as Al2O3 / Si, ZrO2 / Si, Si3N4 / Si, quartz, and sapphire, they can be classified into similar categories based on thermal conductivity, surface oxide stability, reduction sensitivity, and carbon nanotube adhesion strength. Adjustable parameters include O2 volume fraction, rapid heat treatment temperature, holding time, heating rate, inert atmosphere replacement, and optional H2 / N2 post-treatment. The initial screening window can be set according to the table below and optimized based on actual residue, G / D ratio, resistance, and device parameters.

[0052]

[0053] This invention is not limited to specific polymers. For different conjugated polymers, non-conjugated polymers, surfactants, or organic residues, the oxygen-containing steps and H2 / N2 steps can be adjusted according to their thermal decomposition temperature, oxidation sensitivity, and binding strength with carbon nanotubes. This invention is also not limited to single-wafer processing or batch processing, and is suitable for scaling up to different wafer sizes such as 2-inch, 4-inch, 8-inch, and 12-inch wafers. Equivalent thermal budgets can be used based on the actual wafer temperature profile and atmosphere switching response of the rapid thermal processing equipment. Any rapid thermal processing atmosphere window selection logic driven by substrate material properties that achieves a balance between polymer residue removal and low-damage protection falls within the scope of this invention.

[0054] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A low-damage cleaning method for carbon nanotube wafers based on substrate material properties, used to remove polymer residues on the surface of carbon nanotube wafers after polymer-assisted separation, dispersion, or assembly, characterized in that: Obtain the material properties of the substrate supporting the carbon nanotubes; select a rapid heat treatment annealing atmosphere window according to the material properties of the substrate. Different rapid heat treatment annealing atmosphere windows correspond to different substrates with different material properties. The rapid heat treatment annealing atmosphere window includes at least oxygen-containing atmosphere treatment parameters and inert atmosphere replacement parameters. In a rapid heat treatment annealing furnace, the carbon nanotube wafer is subjected to oxygen-containing atmosphere treatment and inert atmosphere replacement sequentially according to the selected rapid heat treatment annealing atmosphere window. The inert atmosphere replacement is carried out until the residual O2 concentration at the exhaust end is lower than the equipment safety threshold, so as to reduce carbon nanotube damage while removing polymer residues on the surface of carbon nanotubes.

2. The low-damage cleaning method for carbon nanotube wafers based on substrate material properties according to claim 1, characterized in that: The material properties of the substrate include one or more of the following: thermal conductivity, thermal diffusivity, surface oxide composition, surface hydroxyl density, oxygen adsorption and release capacity, reduction sensitivity, dielectric layer stability, and adhesion strength between carbon nanotubes and the substrate.

3. The low-damage cleaning method for carbon nanotube wafers based on substrate material properties according to claim 1, characterized in that: For substrates that can withstand high annealing temperatures and do not have heat-sensitive dielectric layers on their surface, a rapid heat treatment annealing atmosphere window with a high annealing temperature and a low O2 volume fraction should be selected. For substrates with a thermally oxidized SiO2 layer on the surface, select a rapid heat treatment annealing atmosphere window with a lower annealing temperature and a higher O2 volume fraction, or a rapid heat treatment annealing atmosphere window with a higher annealing temperature and a lower O2 volume fraction. For substrates with a high dielectric constant dielectric layer on the surface, a rapid heat treatment annealing atmosphere window with a lower annealing temperature and controlled O2 volume fraction and holding time is selected.

4. The low-damage cleaning method for carbon nanotube wafers based on substrate material properties according to claim 1, characterized in that: The substrate is a high-resistivity silicon substrate, a SiO2 / Si substrate, a HfO2 / Si substrate, an Al2O3 / Si substrate, a ZrO2 / Si substrate, a Si3N4 / Si substrate, a quartz substrate, or a sapphire substrate. The oxygen-containing atmosphere used in the oxygen-containing atmosphere treatment is an O2 / N2 mixture, an O2 / Ar mixture, dry air, diluted O2, or pure O2, and the volume fraction of O2 in the oxygen-containing atmosphere is 0.5% to 100%.

5. The low-damage cleaning method for carbon nanotube wafers based on substrate material properties according to claim 1, characterized in that: The heating rate of the annealing treatment is 5~100 ℃ / s; the heating rate is 10~50 ℃ / s; or, when the substrate is an HfO2 / Si substrate, the heating rate is 5~30 ℃ / s.

6. The low-damage cleaning method for carbon nanotube wafers based on substrate material properties according to claim 1, characterized in that: The inert atmosphere replacement is carried out by introducing N2 or Ar. The replacement time is determined based on the effective cavity volume of the rapid heat treatment annealing furnace, the dead volume of the pipeline, the set total flow rate, and the O2 concentration at the exhaust end. The replacement time is 30~180 s.

7. The low-damage cleaning method for carbon nanotube wafers based on substrate material properties according to claim 1, characterized in that: The rapid heat treatment annealing atmosphere window also includes H2 / N2 atmosphere treatment parameters; after the inert atmosphere replacement, the carbon nanotube wafer is subjected to H2 / N2 atmosphere treatment, wherein the volume fraction of H2 in the H2 / N2 atmosphere is 0.5%~5%.

8. The low-damage cleaning method for carbon nanotube wafers based on substrate material properties according to claim 7, characterized in that: The volume fraction of H2 in the H2 / N2 atmosphere is 3%~4%, and the H2 / N2 atmosphere treatment temperature is 350~520 ℃, and the time is 0.5~3 min.

9. The low-damage cleaning method for carbon nanotube wafers based on substrate material properties according to claim 1, characterized in that: When the substrate is a high-resistivity silicon substrate, the oxygen-containing atmosphere treatment adopts a medium-high temperature low oxygen window with an O2 volume fraction of 1%~5%, a temperature of 580~620 ℃, and a time of 2.5~5 min, or a mild pre-screening window with a temperature of 430~530 ℃ and a time of 1~4 min. When the substrate is a SiO2 / Si substrate, the oxygen atmosphere treatment adopts a low temperature and high oxygen short time window with an O2 volume fraction of 20%~100%, a temperature of 300~400 ℃, and a time of 0.5~3 min, or adopts a medium temperature and high temperature and low oxygen window with an O2 volume fraction of 0.5%~5%, a temperature of 500~650 ℃, and a time of 2~7 min. When the substrate is an HfO2 / Si substrate, the oxygen-containing atmosphere treatment adopts a window of O2 volume fraction of 5%~40%, temperature of 350~450 ℃, and time of 1~5 min.

10. The low-damage cleaning method for carbon nanotube wafers based on substrate material properties according to claim 1, characterized in that: The carbon nanotubes are carbon nanotube arrays, carbon nanotube network films, or a combination of both, wherein the carbon nanotube arrays include wafer-level carbon nanotube arrays; when the carbon nanotubes are carbon nanotube arrays, compared to carbon nanotube network films, at least one of the following is used: reducing the volume fraction of O2 in the oxygen-containing atmosphere, shortening the holding time, and reducing the heating rate; when the carbon nanotubes are carbon nanotube network films, the oxidation heat budget is increased while maintaining network connectivity; and one or more of the following are used as criteria for low-damage cleaning effect: polymer residue ratio, G / D change, polarized Raman intensity ratio, alignment angle, sheet resistance, two-point resistance, contact resistance, off-state current, threshold drift, dielectric layer leakage current, and microstructure.

Images