Self-pressurized rapid freezing (SPRF) has been proposed as a simple alternative to traditional high pressure freezing (HPF) protocols for vitrification of biological samples in electron microscopy and cryopreservation applications. Both methods exploit the circumstance that the melting point of ice reaches a minimum when subjected to pressure of around 210 [MPa], however, in SPRF its precise quantity depends on sample properties and hence, is generally unknown. In particular, cryoprotective agents (CPAs) are expected to be a factor; though eschewed by many SPRF experiments, vitrification of larger samples notably cannot be envisaged without them. Thus, in this study, we address the question of how CPA concentration affects pressure inside sealed capillaries, and how to design SPRF experiments accordingly. By embedding a fiber-optic probe in samples and performing Raman spectroscopy after freezing, we first present a direct assessment of pressure buildup during SPRF, enabled by the large pressure sensitivity of the Raman shift of hexagonal ice. Choosing dimethyl sulfoxide (DMSO) as a model CPA, this approach allows us to demonstrate that average pressure drops to zero when DMSO concentrations of 15 wt % are exceeded. Since a trade-off between pressure and DMSO concentration represents an impasse with regards to vitrification of larger samples, we introduce a sample architecture with two chambers, separated by a partition that allows for equilibration of pressure but not DMSO concentrations. We show that pressure and concentration in the fiber-facing chamber can be tuned independently, and present differential scanning calorimetry (DSC) data supporting the improved vitrification performance of two-chamber designs. Lay version of abstract for "Self-pressurized rapid freezing at arbitrary cryoprotectant concentrations" Anyone is familiar with pipes bursting in winter because the volume of ice is greater than that of liquid water. Less well known is the fact that inside a thick-walled container, sealed and devoid of air bubbles, this pressure buildup will allow a fraction of water to remain unfrozen if the sample is also cooled sufficiently rapidly far below the freezing point. This phenomenon has already been harnessed for specimen preparation in microscopy, where low temperatures are useful to immobilize the sample, but harmful if ice-formation occurs. However, specimen preparation cannot always rely on this pressure-based effect alone, but sometimes requires addition of chemicals to inhibit ice-formation. Not enough is known directly about how these chemicals affect pressure buildup: Indeed, rapid cooling below the freezing point is only possible for small sample volumes, typically placed inside sealed capillaries, so that space is generally insufficient to accommodate a pressure sensor. By means of a compact sensor, based on an optical fiber, laser and spectrometer, we present the first direct assessment of pressure inside sealed capillaries. We show that addition of chemicals reduces pressure build-up and present a two-chambered capillary to circumvent the resulting trade-off. Also, we present evidence showing that the two-chambered capillary design can avoid ice-formation more readily than a single-chambered one. This article is protected by copyright. All rights reserved.