Modulating Surface Chemistry of Al Powders for Elastomeric Composites with Applications in Electronic Cooling

Elastomeric composites are an important class of thermal interface materials as they are shape‐adaptive, which can fit uneven interfaces and achieve ideal interfacial heat transfer in electronic cooling. There is often a trade‐off between mechanical and thermal performances in highly filled composites. Here, the surface chemistry of Al powders via introducing dodecyltrimethoxysilane (DTS) upon different grafting densities is precisely controlled. The results show that the DTS grafting density of 0.24 molecule nm−2 endows the composites with optimized mechanical properties. The tensile stress, elongation at break, and Young's modulus are enhanced by 180%, 56%, and 94.4% after DTS grafting, respectively. Meanwhile, the composite possesses a high fracture energy of 356.2 J m−2 with an enhanced strain energy density of 37.6 kPa. The DTS grafting on Al also contributes to higher thermal conductivity (4.4 W mK−1) and lower contact thermal resistance (0.52 °C cm2 W−1) of the composites and significantly improves the stability of the composites upon high‐temperature storage. Guidance is provided here on methods and data references for optimizing the performance of elastomeric composites, which helps to expand their practical application in electronic cooling.


Introduction
Electronic components and devices will produce varying levels of heat during their operation, thus electronic cooling has played a significant role. [1] Where significant amounts of heat www.advmatinterfaces.de on the performance of the elastomeric composites still needs to be further clarified.
In this study, via the precise introduction of a quantifiably assessable DTS layer of 0.24 molecule nm −2 on Al powders to enhance the transfer of interfacial loads and heat carriers, we have achieved simultaneous improvements on mechanical strength (≈180%), elongation at break (≈56%), Young's modulus (≈94.4%), fracture energy (≈215%), strain energy density (≈355%), and thermal conductivity (≈7.9%), along with reduced contact thermal resistance (≈20.1%). The result composites exhibit satisfactory mechanical and thermal stability, high thermal conductivity (4.4 W mK −1 ) and low contact thermal resistance (0.52 °C·cm 2 W −1 ), good thermal shock resistance and fast thermal response in cyclic T3Ster tests, demonstrating great potential for application in the heat dissipation of electronic devices under scenarios containing variable external coupling fields. Our study has provided data references for exploring how the surface chemistry of filler affects the overall performance of elastomeric composites.

Variation of the Surface Chemistry of Al Powders
The grafting mechanism is shown in Figure 1a. There are two distinct segments on the thermogravimetric analysis (TGA) curves in Figure 1b. DTS shows less thermally stable and has all decomposed at close to 250 °C ( Figure S1, Supporting Information). When the temperature reaches 200 °C, the weight loss of raw Al powders is ≈1.51 wt%, which is due to the physical moisture adsorbed on Al surface. The weight loss of all the modified Al is smaller due to the depletion of certain hydroxyl groups after the DTS grafting. As the temperature further www.advmatinterfaces.de increases and reaches 650 °C, the overall weight loss of raw Al powders is ≈2.43 wt% due to further detachment of hydroxyl groups. On the other hand, the weight loss of modified Al is 2.74 wt% (y = 0.13), 2.89 wt% (y = 0.19), 3.02 wt% (y = 0.24), and 3.15 wt% (y = 0.29), respectively, which is mainly due to the thermal decomposition of grafted DTS molecules. The magnified curve ( Figure S2, Supporting Information) indicates a clear difference in the residual quality of the different samples after heat treatment. The Fourier-transform infrared spectroscopy (FT-IR) spectra of Al before and after DTS grafting are shown in Figure 1c. Compared to raw Al, the additional peaks at 2950 and 2880 cm −1 in modified Al correspond to the CH stretching vibration. Meanwhile, the additional peak at 1460 and 1130 cm −1 corresponds to the CH bending vibration and the SiOH vibration, [11] respectively, demonstrating the successful grafting of DTS on Al surface. [12] With the increase of the DTS grafting density from 0.13 to 0.29 molecule nm −2 , the intensity of the above peaks shows a distinct enhancement which lays a verification for precise regulation of the surface chemistry. The splitting peaks of X-ray photoelectron spectroscopy (XPS) spectra of raw Al and DTS 0.24 are shown in Figure 1d. On the basis of AlOH peak (74.76 eV) and AlOAl peak (74.25 eV) in raw Al, [13] a newly emerged peak with a binding energy of 75.48 eV is found in DTS 0.24, which can be attributed to the structure of AlOSi, indicating the chemical bonding between DTS and Al. [14] Elemental scanning of Si is further proceeded to confirm the existence of DTS grafting. Traces of Si compounds were inevitably introduced in raw Al powders due to the production process, thus Si peaks can be seen in the XPS spectra of raw Al. [15] As shown in Figure 1e, the intensity of the Si2p peak is significantly enhanced upon higher DTS grafting density which coincides with the varying pattern of Si content in the inset table. The variation of surface chemistry is also demonstrated by testing the contact angle with water ( Figure 1f) as the grafting of organic molecules inevitably changes the hydrophobic nature of Al surface. The contact angle becomes distinctively larger upon the increasing DTS grafting density. The above characterization indicates that the spectral profiles of Al powders are distinctly different upon different DTS grafting densities, which is beneficial for us to monitor the surface chemistry of Al. The morphology of Al powders unchanged due to the nano thickness of grafted DTS (Figure 1g). Scanning electron microscope (SEM)-Element mapping spectrum and XPS mapping is also conducted in Figure S3 in the Supporting Information. The silicon element is evenly distributed on the surface of the aluminum spheres, indicating the successful grafting of DTS.
Interfacial bonding is essential to determine the composites' overall performance. As interfacial adhesion is different from being measured directly, the fracture morphology can assist in evaluating the interaction between Al and silicone matrix. In Figure 2a, the interfacial separation could be observed along with the holes and gaps in TIM@Aluminum, indicating relatively weak interfacial bonding. Meanwhile, the agglomeration of Al powders also demonstrates the mismatch in surface energy between Al and silicone matrix. The holes between Al and silicone matrix almost disappeared in TIM@DTS 0.13 (Figure 2b), which illustrates that DTS grafting has significantly improved the interfacial compatibility. Meanwhile, the trend of Al aggregation has also slowed down. Upon higher DTS grafting densities, Al powders are becoming more evenly distributed as clearly shown in Figure 2c,d. The above phenomena confirm that DTS grafting leads to stronger interfacial bonding, facilitating the transfer of loads and lattice vibrations at the interface.

The Effect of DTS Grafting on Mechanical Performance
The effect of DTS grafting of Al on the tensile properties of the composites is investigated as shown in Figure 3. The stressstrain curves ( Figure 3a) demonstrate that the introduction of raw Al powders into silicone matrix results in significant decay of elongation at break but improved tensile stress. This trend in data is similar to that of the previously reported studies. [6,16] Higher content of Al makes the composite less tensile via damaging the entanglement of polymer chains. Meanwhile, Al possesses higher strength than silicon matrix and acts as a crosslinking center to transfer the load effectively, improving the tensile strength. The mechanical performance of TIM@Aluminum and TIM@DTSy filled with 90 wt% Al is further characterized ( Figure 3b). DTS grafting contributes to significantly improved tensile properties with a maximum tensile strength over 0.26 MPa corresponding to a 180% enhancement and a maximum elongation at the break over 150% corresponding to an increase of 98%. Except TIM@DTS0.13, Young's modulus of other composites containing DTS-grafted Al is all higher than that of TIM@Aluminum ( Figure 3c) where TIM@DTS0.24 exhibits the highest Young's modulus of ≈3.5 kPa. Strong interfacial bonding allows for better stress transfer between Al and silicone matrix. In this case, there is an optimum value for the thickness of the DTS layer which ensures an effective stress transfer but is not so thick as to cause interface peeling. By evaluating stress, elongation at break, and Young's modulus together, TIM@DTS 0.24 is considered to have the best combined mechanical properties. The stress-strain test of TIM@DTS 0.24 at different tensile rates and strains is further conducted as shown in Figure 3d,e, respectively. Upon the tensile www.advmatinterfaces.de rate increasing from 5 to 100 mm min −1 , the tensile strength of TIM@DTS 0.24 exhibits a visible enhancement while the elongation at break shows a sharp decline (Figure 3d). The viscoelasticity of the elastomeric composites dominates the behavior in response to an external load, thus the strain rate greatly correlates with the relaxation time of the segment of polymer chains. When the external load is fixed and the tensile rate is relatively high, the composite does not share enough time to make viscoelastic deformation happen. Therefore, the same deformation would require a more enormous external load. Figure 3e shows the stress-strain property of TIM@DTS 0.24 upon different strains from 30% to 120% where TIM@DTS 0.24 exhibits desired elastic recovery. We further obtained the strain-stress curves of TIM@DTS 0.24 at 90% strain for 1000 cycles (Figure 3f). The composite exhibits satisfactory elastic recovery, with the stress after 1000 cycles still maintaining 54% of the initial value. Meanwhile, no cracks appear in the composite, indicating that TIM@DTS 0.24 shares good resistance to cyclic tensile loading.
Fracture energy is another important factor for evaluating the stability of elastomeric composites to resist external tearing. Experiments guided by Rivlin-Thomas's fracture energy theory were performed to analyze the effect of the subtle variations of surface chemistry in Al. [17] The stress-extension ratio curves of the composite samples with and without the crack upon different DTS grafting densities are shown in Figure 4 with summarized fracture energy (Γ) and strain energy density (U). The increasing DTS grafting density contributes to the significant and simultaneous enhancement of Γ and U. Compared to the TIM@Aluminum, Γ has increased from 113.0 to 472.6 J m −2 as the DTS grafting density reaching 0.29 molecule nm −2 , which corresponds to an enhancement of 318%. Meanwhile, www.advmatinterfaces.de U increases from 11.4 to 51.9 kPa, representing an increase of 355%.
Inorganic fillers and organic polymers possess different chemical properties. The silane coupling agent acts as a bridge between the two and has both organic and alkoxy groups in its molecular structure. [18] The DTS forms covalent bonds with the aluminum spheres after hydrolysis, while the resultant silanol group of DTS-aluminum and the vinyl silicone oil undergo hydrophobichydrophobic interactions. In this way, DTS improves the interfacial adhesion between the aluminum sphere and silicone matrix. The improved interfacial adhesion significantly optimized the mechanical properties of the material, [19] including tensile stress, elongation, modulus, and fracture energy. On the one hand, DTS reduces the polarity difference between the aluminum and silicone matrix, contributing to a more uniform distribution of aluminum spheres and a continuous silicone network. Each aluminum sphere can be regarded as a macroscopic crosslinking point and a homogeneous network of aluminum spheres can transfer external loads quickly to each crosslinking point. On the other hand, the enhanced interfacial adhesion strengthens the interface between the aluminum spheres and the silicone matrix, making the two less likely to peel off and increasing the mass transfer efficiency between them. Covalent bonds between DTS and aluminum spheres have greater bonding energy than van der Waals forces; thus, they can consume more internal stresses caused by external loads and maintain stability. Therefore, when external loads are applied to composites containing DTS-modified aluminum spheres, they are more likely to be propagated and consumed efficiently. Intuitively the composite will exhibit greater mechanical performances.
The present work was also conducted with the interesting conclusion that not all mechanical parameters improve with increasing DTS grafting amounts. Different mechanical parameters involve different energy transfer and release processes. An increase in DTS layer thickness changes the ratio between the aluminum sphere-DTS and DTS-silicone matrix interfacial layer, thus affecting the efficiency of interfacial mass transfer. This would require further in-depth study.

The Effect of DTS Grafting on Thermal Performance
The effect of DTS grafting on thermal performance was also investigated. The intrinsic thermal conductivity of silicone matrix is 0.23 W mK −1 . Figure 5a shows a general variation pattern of thermal conductivity in the composites. The thermal conductivity increases with Al content as more Al powders are involved in constructing a thermally conductive network. The thermal conductivity of the elastomeric composite reaches about 4.0 W mK −1 with the addition of 90 wt% Al. It should be noticed that DTS grafting leads to a relatively small margin of enhancement in thermal conductivity compared to the improvements in tensile strength and fracture energy (Figure 5b). Although the DTS grafting improves the interfacial adhesion and promotes heat transfer, the introduction of DTS also generates new Al/DTS and DTS/silicone matrix interfaces, leading to more intensified interfacial phonon scattering. [20] Thus, the positive and negative effects cancel each other out to some extent. 0.13 and 0.24 molecule nm −2 of DTS grafting both contribute to nearly 8% higher thermal conductivity than that of TIM@Aluminum. Besides the thermal conductivity, the contact thermal resistance is also an essential factor for assessing the thermal transferring performance. Figure 5c highlights the advantage of TIM@DTS 0.24 for its lowest contact resistance (0.52 °C cm 2 W −1 ). For comparison, we have also summarized the contact resistance of some available TIMs and materials in literature reports. Table S1 in the Supporting Information illustrates that the performance of our samples is among the better ones. It is www.advmatinterfaces.de important to note, however, that contact thermal resistance is related to interface roughness, temperature, and pressure. For example, the higher the pressure, the lower the contact thermal resistance normally is. Previous comparisons under uniform conditions are lacking in the literature. Therefore, Table S1 in the Supporting Information only provides a rough comparison for reference purposes. The above data illustrates the balance of mechanical and thermal performance achieved upon a DTS grafting density of 0.24 molecule nm −2 . Elastomeric composites used as thermal interface materials normally suffer from high temperatures in real application scenarios. Thus, the thermal stability of the composites under high-temperature treatment is essential which is evaluated via the variation of the overall thermal resistance. Figure 5d shows the test device where we forged a 0.5 mm deep, 1 in. square groove at the bottom of the mold, consistent with the size of the heating module and our composite samples for the characterization of thermal resistance. TIM@Aluminum and TIM@DTS 0.24 were then stored at high temperatures for various times with the overall thermal resistance recorded in Figure 5e. Silicone matrix contains specific volatile components which will run out as small molecules upon high-temperature treatment, making the composites harder and the interfacial adhesion degraded. Therefore, the overall thermal resistance of TIM@Aluminum and TIM@DTS 0.24 both increase upon longer storage time. The thermal resistance of TIM@Aluminum before and after high-temperature storage of 1000 h are 83.8 and 92.4 °C cm 2 W −1 , respectively, illustrating a performance degradation of 10.3%. On the other hand, the thermal resistance of TIM@DTS 0.24 before and after high-temperature storage are 83 and 88.9 °C cm 2 W −1 , respectively. The rise in overall thermal resistance is less than that of TIM@Aluminum. Meanwhile, TIM@DTS 0.24 consistently exhibits lower thermal resistance than TIM@Aluminum. This is mainly due to the stronger interfacial bonding between DTS-grafted Al and silicone matrix, suppressing the precipitation of small molecules and delaying the aging behavior of the composites. The volume resistivity of our sample upon different voltages is measured as shown in Figure S4 in the Supporting Information. The composites present electrical insulation even at high filler loadings as the value is much larger than the insulator's critical value (10 9 Ω cm).
To illustrate the practical application in heat dissipation of the chip, we have applied the T3Ster method to evaluate the thermal shock stability of TIM@DTS 0.24 during cyclic heating and cooling. TIM@DTS 0.24 is placed between the chip and the heat sink (Figure 6a), where the input source power of the chip can be precisely controlled to produce accurate thermal metrics. A photograph of the physical installation is shown in Figure 6b. By applying cyclically varying current into the chip, a rapid junction temperature transition (2 min per cycle for 1000 cycles) can be achieved with the chip temperature-time curve recorded in Figure 6c. The heat generated from the chip will be transferred through TIM@DTS 0.24 to heat sink and further to the outside environment via the fan. If there is a significant variation of the intrinsic thermal conductivity or interfacial adhesion of TIM@DTS 0.24 during the cyclic test, this will result in a sudden change in the chip temperature, indicating that the material itself is less resistant to thermal shock. The fragment curves drawn from Figure 6c show that the maximum and minimum values of the chip temperature do not change significantly (Figure 6d), indicating that TIM@DTS 0.24 has good reliability for use as thermal interface material upon longterm operation. Moreover, our composites can also be made into various patterns to suit different application scenarios by the dispensing process before thermal curing ( Figure S5 and Movie S1, Supporting Information).

Conclusion
In this study, we have thoroughly investigated the effects of different DTS grafting densities of Al powders on the mechanical and thermal performances of the elastomeric composites. The results show that the DTS grafting endows the composites with satisfactory mechanical strength and tensile deformation and significantly enhanced tearing resistance. The above improvements are mainly due to the stronger interfacial bonding between Al and silicone matrix after DTS grafting, leading to a more homogeneous distribution of Al, as well as the DTS layer itself both of which help to transfer the load uniformly. Meanwhile, DTS grafting also leads to improvement in thermal conductivity and contact resistance, partially contributing to the satisfactory reliability upon high-temperature storage. T3Ster results further indicate that the composites possess www.advmatinterfaces.de good thermal shock resistance and response speed when used as thermal interface material between chip and heat sink upon 1000 cyclic tests. The present work has been carried out to provide a data reference for the practical application of elastomeric composites in electronic cooling.

Experimental Section
Materials: Raw Al powders with an average size of 1-2 µm were purchased from Hunan Ningxiang Jiweixin Metal Powder Co., Ltd. DTS, 1-ethynyl-1-cyclohexanol, platinum catalyst, and ethanol of A. R. grade were offered by Shanghai Aladdin Biochemical Technology Co., Ltd. Vinyl and hydrogen-containing silicone oils were provided by Zhejiang Runhe Chemical Co., Ltd. All the chemicals are used without further purification.
Surface Modification of Al Powders via DTS: First, a certain amount of DTS and ethanol aqueous solution were mixed followed by stirring and the addition of Al powders. Upon treatment of sonication for 5 min, the mixture was then transferred into water bath under 60 °C for several hours with stirring. The as-modified Al powders were collected by filtration and washed by ethanol for several times. Finally, the modified Al powders were obtained after vacuum-assistant drying.
Calculation of DTS Grafting Density: The DTS grafting density is mainly determined by the weight ratio of DTS/Al and reaction time. Increasing the weight ratio of DTS or the reaction time can increase the DTS grafting density. The number of hydroxyl groups on Al surface would limit the maximum grafting density. Table S2 in the Supporting Information shows the reaction conditions for different DTS grafting densities. The grafting density is calculated on the basis of thermogravimetric data. Within the certain temperature range, the weight loss of raw Al powders is 2.43%. The following equation can be obtained Fabrication of the Elastomeric Composites: Vinyl and hydrogen silicone oils were first mixed according to Table S3 in the Supporting Information. A certain amount of the resultant silicone oil and Al powders were mixed in a planetary mixer for 10 min. The mixture was then heated to 90 °C under vacuum and stirring for removal of air and better blending. Platinum catalyst was added to the mixture as the temperature reduced to room temperature, followed by being stirred for further 30 min. Samples of elastomeric composites were prepared by placing the mixture into three-roller calendar followed by thermal curing. Shapes can be cut to suit the needs.
Calculation of the Fracture Energy: A rectangular sample without a notch was first stretched by one side until break. Meanwhile, the curve of stress-extension ration was recorded. For comparison, sample with the same composition but with a crack of 2 mm long was also tested with the elongation remarked when the crack notch started to move. For a normal rectangular sample with a crack in the direction perpendicular to the side, the fracture energy (Γ) can be expressed as [21,22]

Γ = 2kUc
(3) where c is the length of the crack and k is a constant calculated by Characterization: The weight loss of raw and modified Al powders was obtained via SDT-Q600 Simultaneous TGA equipment (TA Instruments, USA). The grafting condition of DTS was confirmed by VERTEX 70v FT-IR spectrometer (Bruker, USA). The analysis of surface chemistry of modified Al powders was conducted by ESCALAB 220I-XL X-ray photoelectron spectroscopy (Thermo Scientific, USA). OCA 20 Contact Angle Meter was applied to characterize the wetting angle of Al (DataPhysics Instruments, Germany). Scanning electron microscopy (Field Electron and Ion Company, FEI, Nova NanoSEM450, USA) was used to characterize the morphology of the composites. Mechanical performance was recorded via Autograph AGS-X Series Precision Universal Tester (Shimadzu Corp., Japan). LW-9389 TIM Thermal Resistance and Conductivity Measurement Apparatus (LongWin, Taiwan) was applied to measure thermal resistance and calculate apparent thermal conductivity based on ASTM D 5470-06 Standard. Stability testing was proceeded on GPS-3 Temperature Cycling Chambers (Espec Corp., Japan). The thermal shock resistance of the composites was recorded via a T3Ster Thermal Transient Tester (Siemens, Germany).

Supporting Information
Supporting Information is available from the Wiley Online Library or from the author.