Elettra | BEAR Chemistry for microelectronics: how photoresists for extreme ultraviolet lithography work at the molecular level Electronics are an increasingly large part of our daily life, from the smartphone in our pockets to the huge servers distributing dozens of exabytes (1 EB = 1 billion TB) of data every day all around the world. From the low speed GSM network developed more than 30 years ago delivering kilobits per second to modern 5G networks capable of hundreds of megabits per second, everything from phones to networking infrastructure and servers had to be significantly upgraded to accommodate increased performance. Of course, as performance increases, the power envelope needs to remain fit for their purpose; mass adoption would be difficult if the better performance would require larger batteries. To that end, new technology requires to be both more performant and power efficient to continue the perpetuation of Moore's law that has been governing the field of computing electronics for more than half a century. The main parameter used to increase performance and reduce power consumption is the miniaturization of the electronic circuits, and more specifically the ability to print them using the smallest features possible. For that, a process called photolithography is used, in which light transfers information into a physical pattern on a layer of the substrate. First, a substrate is coated with a very thin layer of photoresist. This material has the ability to undergo chemical changes upon exposure to light changing its ability to be dissolved by using a solvent called developer. If the photoresist exposed to light becomes insoluble to the developer, it will form a so-called negative pattern, conversely if the exposed photoresist becomes soluble it will leave a positive pattern. Once the pattern is formed on the substrate, subsequent steps are performed to etch, add material or dope the desired parts of the substrate to modify their properties and ultimately obtain the desired functional structures. To obtain patterns with smaller and smaller features and reach nanometer sizes, the wavelength of the light itself has to be reduced. For this reason, the wavelength used for lithography has been steadily reduced from mercury lamps at 436 nm in the 1960s to deep UV argon fluoride lasers at 193 nm which still represent a large portion of the manufacturing today. However, to print even finer features in a single costefficient process step, current cutting edge chips are made using extreme ultraviolet (EUV) which has a wavelength of 13.5 nm. This EUV radiation is much more energetic than conventional deep UV light, and this leads to new photochemical processes that require the exploration of new, more suitable photoresists. Our research focuses on understanding these photochemical processes and give mechanistic insights on the chemical reactivity of new EUV-suitable photoresists. Our work is performed on one of the most promising negative tone candidates for EUV lithography: tin-oxo-hydroxo cage molecules (Fig. 1 a). This cluster has the particularity of being positively charged and thus requiring negatively charged anions. These anions can be exchanged, allowing versatile tuning of the physical and chemical properties of the photoresist. In the present work we exchanged the anions with large pentafluorophenyl borate anions which contain a large number of fluorine atoms. One reason for this choice is that fluorine has a much higher EUV absorption crosssection than the typical organic elements like oxygen, carbon and hydrogen. This is important because a thin film of the photoresist must absorb as large a fraction EUV photons as possible to allow the photochemical reactions to occur. Upon EUV exposure, we observed a surprising result: this photoresist behaved as a positive tone resist rather than the expected negative tone. To probe the changes occurring in such photoresists upon EUV irradiation and gain mechanistic insight, synchrotron experiments were used. Two beamlines at the Paul Scherrer Institute were used (1) to expose the photoresist thin films (on which infrared spectroscopy was performed ex situ) and (2) to measure variations in the x-ray absorption spectra at the carbon and fluorine K-edges, enabling to monitor the chemical changes in these elements. At Elettra, on the BEAR beamline, we monitored the chemical changes of the photoresists thin films by exposing them to EUV and using in situ x-ray photoelectron spectroscopy (Fig. 1 b). Analyzing the samples immediately after EUV exposure without air contact or time delay is a greatly advantage, as reactive species such as radicals can be formed upon EUV irradiation. By complementing these results with other techniques such as mass spectrometry and quantum chemical calculations we were able to obtain mechanistic insight, clarifying the difference in reactivity between the tin oxo-hydroxo cluster with hydroxide anions and with pentafluorophenylborate anions. In tin clusters with small counterions, the usual mechanism for the solubility switch involves cleavage of the Sn–C bond, which presumably leads to cross-linking of the clusters and reduces the photoresist’s solubility. In the photoresist with pentafluorophenylborate anions, however, this process competes with degradation of the anion itself, which grafts onto the cluster and thereby increases its solubility. This uncommon behaviour shows the significant influence of the anion on the tin-based photoresist reactivity and introduces an alternative approach to significantly tune tin clusters reactivity by changing the anion instead of the tin cluster itself. Further mechanistic research along these lines has been performed at the BEAR beamline and will be published soon. Elettra Highlights 2024–2025 18
RkJQdWJsaXNoZXIy MjQ4NzI=