DOI: 10.1002/adma.200601528 High-Modulus Spin-On Organosilicate Glasses for Nanoporous Applications** By Hyun Wook Ro, Kookheon Char, Eun-chae Jeon, Hie-Joon Kim, Dongil Kwon, Hae-Jeong Lee, Jin-Kyu Lee, Hee-Woo Rhee, Christopher L. Soles ,* and Do Y. Yoon* Nanoporous glasses, that is, structural glasses with large quantities of nanoscale porosity, are the cornerstone of sever- al emerging technologies. The most prominent example is the semiconductor industry where interlayer dielectric (ILD) in- sulator films are critically needed to both reduce the resis- tance–capacitance time delay of the interconnect circuitry and reduce the cross-talk between adjacent conduction lines. For next-generation semiconductors these materials need an ul- tralow dielectric constant (k < 2.2) while being compatible with patterning processes where the minimum features are smaller than 40 nm. [1–6] These structures, owing to their size and dielectric requirements, can only be realized by introduc- ing large quantities of nanoscale porosity. Other technologies where nanoporous glasses are of significant interest include catalyst supports, [7] separations and membranes, [8] laser mate- rials, [9] optical waveguides, [10] sensors, [11] and applications in biotechnology. [12,13] Adding porosity will generally deteriorate the mechanical properties of a glass, such as the modulus or hardness. [5,6,12] This is especially a problem for the spin-on materials that, even in their nonporous forms, typically have an elastic modu- lus of less than 5 to 10 GPa. [2,5,12,14] Poly(methylsilsesquiox- ane) (PMSQ) is perhaps the most common example of a spin- on glass. [2,5,6,12–14] PMSQ organosilicate glasses (OSGs) are attractive for next generation ILD applications because they can be easily spin-cast into smooth films, have inherently low dielectric constants (k = 2.7 to 2.9) in their nonporous form, can be rendered porous through the addition of a sacrificial pore generating (porogen) material, and exhibit thermal sta- bility up to 500 °C. However, these materials also tend to be mechanically fragile and inducing porosity significantly worsens the problem. An ILD material must possess the mechanical strength to withstand the harsh and abrasive chemical–mechanical planarization (CMP) processes used in semiconductor fabrication. Higher-modulus materials are su- perior for withstanding the loads induced by the CMP process. A high elastic modulus is also a desirable property for coating applications in general. Higher-modulus inorganic glass films or coating can be achieved with chemical–vapor deposition (CVD) techniques. However, it is generally realized that spin- on deposition techniques using OSG materials lead to higher- quality films, especially at high porosity. The extendibility of spin-on OSGs to ultralow k films (very high porosity levels) is promising if the modulus of nonporous starting material can be substantially improved. This is the subject of this manu- script. Another critical property is a low coefficient of thermal expansion (CTE) because of the large temperature changes that are encountered during integration and packaging: a low CTE minimizes the thermal mismatch stresses between the ILD and the other metal or silicon-based materials in the complementary metal oxide semiconductor (CMOS) circuitry that have inherently low CTEs. [1] Most OSGs exhibit CTEs in the range of 100 to 300 × 10 –6 °C –1 , which is about two orders of magnitude greater than the other “hard” materials found in a CMOS device. The resulting thermal mismatch stresses incurred during integration can lead to catastrophic failure. These liabilities in both the mechanical and thermal proper- ties have precluded the integration of nanoporous materials into state of the art semiconductor devices. There is a pressing need to characterize why these types of nanoporous glasses are so fragile and identify opportunities to increase their ther- mal and mechanical properties. [15,16] Our approach is to incorporate organic bridging units into the molecular architecture of PMSQ. PMSQ is typically synthesized through the hydrolytic condensation of methyl- trichlorisilane or methyltrimethoxysilane (MTMS—shown in COMMUNICATION Adv. Mater. 2007, 19, 705–710 © 2007 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim 705 – [*] Dr. C. L. Soles,Dr. H. W. Ro, Dr. H.-J. Lee Polymers Division, National Institute of Standards & Technology Gaithersburg, MD 20899 (USA) E-mail: csoles@nist.gov Prof. D. Y. Yoon,Dr. H. W. Ro, Prof. H.-J. Kim, Prof. J.-K. Lee Department of Chemistry, Seoul National University Seoul 151-747 (Korea) E-mail: dyyoon@snu.ac.kr Prof. K. Char School of Chemical & Biological Engineering Seoul National University Seoul 151-742 (Korea) Dr. E.-c. Jeon, Prof. D. Kwon School of Materials Science & Engineering Seoul National University Seoul 151-747 (Korea) Prof. H.-W. Rhee Department of Chemical Engineering, Sogang University Seoul 121-742 (Korea) [**] This official contribution of the National Institute of Standards and Technology is not subject to copyright in the United States of Ameri- ca. This work was supported in part by the System IC 2010 Project of Korea, the Chemistry and Molecular Engineering Program of the Brain Korea 21 Project, and the NIST Office of Microelectronics Pro- grams. The authors thank Dr. Robert Cook and Jack Douglas for their careful critique of the manuscript. Supporting Information is available online from Wiley InterScience or from the author.