Nanomedicine, Volume I: Basic Capabilities

© 1999 Robert A. Freitas Jr. All Rights Reserved.

Robert A. Freitas Jr., Nanomedicine, Volume I: Basic Capabilities, Landes Bioscience, Georgetown, TX, 1999


10.3.4 Vessel Leakage and Flammability

Drexler10 estimates that valve, gasket or seal mechanisms with opposing faces at an equilibrium separation of 0.3 nm should experience a negligible fluid leakage rate at elevated pressures for all atoms or molecules except helium. In the case of helium at 300 K, the estimated leakage rate of ~10-15 atoms/nm-sec through a seal opposing a ~10-5 atm head pressure (~1020 atoms/m3) rises to ~10-7 atoms/nm-sec at a ~1000 atm head pressure (~1028 atoms/m3)10. Thus at 1000 atm, the mean waiting time for a seal of length 100 nm to pass a single He atom is ~105 sec (~1 day), which is an extremely slow leakage rate given that a 1 micron3 pressure vessel at 1000 atm contains ~1010 helium atoms (Table 10.2). Positive ions such as Li+ and Be++ have smaller radii than helium atoms and thus could in theory pass more easily than helium through a seal, but these ions will rarely be found in large numbers outside a solvating liquid environment.10

Gases might also leak out by diffusion through pressure vessel walls. Hydrogen has the highest coefficient of diffusion of all the gases (Table 3.3). H2 readily diffuses through porous substances such as clay, rubber, and even quartz and silica at elevated temperatures. Gases may diffuse through conventional metals, which have numerous defects, dislocations and grain boundaries, and through silicate glasses, which have open irregular structures that permit substantial diffusion of helium.10 Hydrogen also dissolves in and diffuses through the metals of the nickel, palladium and platinum groups. For example, Pd may dissolve ~900 volumes of H2 at 293 K and 1 atm, and diffusion through a palladium thimble is often used in the laboratory to purify hydrogen gas. H2 diffusion through bulk metals has been widely studied.2056,2057

However, diffusion leakage through defect-free pressure vessel walls comprised of diamond, graphene, or corundum should be negligible. "It is difficult to see how a molecule can pass through these materials without undergoing a reactive transformation."10 The energy required to move atomic hydrogen from free space into a minimum-energy site in diamond is thermally prohibitive (>~800 zJ),2055 and the required energy is even higher for molecular hydrogen. Room temperature gases at ~1300 atm1169 and argon gas at ~600 atm2034 have been stably trapped inside carbon nanotubes, and confirmed by computer simulations,3212-3215 and helium atoms at energies up to ~5 eV (~800 zJ) cannot penetrate a graphene sheet. Ruby (corundum) is commonly used in diamond anvil experiments in which hydrogen is compressed to pressures >106 atm with no apparent leakage.2043 Thus it appears that essentially leakproof pressure vessel walls can be constructed, although no computational studies of diamond/sapphire interfaces had been performed by 1998 [D.W. Brenner, personal communication, 1999]. Double-walled structures with an intervening space containing either getter materials or a pumped vacuum may be used, if necessary, for added safety.10

Microscopic pressure vessels comprised of diamond and filled with compressed hydrogen or oxygen are potentially flammable, even explosive. However, the chemical energy contained within a volume scales as ~L3 while the dissipation of that energy during an explosive event (e.g., pressure, heat, light) occurs across a surface which scales as ~L2, hence energy dissipation per unit area (and thus the relative impact of an explosion on the local environment) scales as ~L-1 which implies declining severity at smaller device sizes (Chapter 17). For example, a liter of nitroglycerine produces a blast intensity of ~108 J/m2, while a spherical 1 micron3 of nitroglycerine gives only ~103 J/m2 upon decomposition.

Heat from operating valves is small and is rapidly conducted away in a diamondoid structure. Corundum employed at all structure/oxidant interfaces virtually eliminates device combustibility. Even for bulk diamond in contact with pure oxygen gas at 1000 atm, the time required to etch a single layer of carbon atoms of thickness ~0.17 nm at 310 K may be crudely estimated from Eqn. 9.55 as ~1011 years; at 400 K, the atomic-layer etch time is still ~400 years. Further discussion of possible nanodevice combustibility is deferred to Chapter 17.


Last updated on 24 February 2003