space agencies have only been able to support a limited number of science payloads—far fewer than what’s needed to advance to a truly spacefaring civilization. this scarcity, reinforced by risk-averse cultures, has shaped the historic payload development workflow: a long, bureaucratic process where technologies and research protocols are derisked over years, sometimes decades, before they have a chance to reach orbit. as the science budget continue to dwindle at NASA, earth-based microgravity simulation testing has moved from being a launch prerequisite to a near default for investigating key questions in space biology, simply because it is far cheaper to do so. these simulation tools, however, are fundamentally limited in their ability to reproduce the microgravity environment, limiting the applicability of science findings to human exploration systems.
I want to preface this writing by saying, rather ironically, that I hope the information it contains will become useless soon. as launch costs to low-earth orbit have dropped by roughly an order of magnitude in the past fifteen years—with the potential for at least another order drop once starship (and a single other competitor!) becomes commercially operational—a prime reason emerges to rethink our payload development workflow. instead of optimizing every aspect of a payload for minimal mass and maximal reliability—often at great cost and delay—we can directly prioritize faster iteration, modular design, and greater experimental risk so long as the data collection and downlink methods remain sound. cheaper access to orbit means it may be more effective to launch, learn, and relaunch than to spend years perfecting a single, high-stakes flight with simulations. in short time, we should start treating spaceflight hardware more like software: something to be tested, updated, and evolved rapidly in response to data. but in the meantime, we work under simulated microgravity.
----
several different experimental apparatuses have been developed over the years to study the effects of altered gravity on earth. for larger biological systems like humans, we have facilities like the neutral buoyancy lab at the johnson space center, which can manipulate buoyancy using pools, cranes, and weights to simulate full or partial gravity. simulating microgravity for plants is done through the use of clinostats, which grow plants on platforms that rotate semi-randomly in three dimensions such that the gravity vector averages to zero over time. over the years, we’ve discovered that plant growth and developmental responses are quite similar to what we see in space; however, the expression of certain genes remain slightly different between simulation and spaceflight samples.
in this post though, I want to talk about simulating microgravity for liquid cultures (i.e., microbial communities) on earth for a couple of reasons. first, it’s a somewhat understudied area compared to the potential that could be unlocked with greater understanding at present: if we had better ability to control structure and function in liquid cultures, then we could in theory produce many essential foods, vitamins, and materials in space from widely available feedstock AND upcycle waste in the space environment. we wouldn’t need to rely on pre-packaged space food from resupply or nearly as many expendables as we currently do. darpa’s b-sure program affiliates are doing some great work here, but given the possibility space for biology, there’s always more avenues to explore. the second is a more practical thing: liquid culture simulators are likely better approximators of microgravity for their intended biological model. not only is the gravity vector theoretically averaged to zero over time—mechanical stress is also minimized across the entire organism, imitating the low-shear environment of microgravity caused by the loss of significant convection.
microgravity simulators for liquid cultures maintain the same basic form: a cylindrical vessel mounted horizontally, made to slowly rotate about its central axis. If they are completely filled with liquid culture, turbulent sloshing is eliminated, resulting in laminar flow (Re≪1) in the bulk liquid while exposing contained cells to a roughly zero-averaged gravity vector. but a key engineering design question arises—at what angular speed should these vessels rotate to best approximate microgravity? too slowly, and cells will sediment to the bottom of the vessel, adding pressure and shear not observed in microgravity. too fast, and cells will also experience shear as the centrifugal force (in a rotating reference frame) pushes them toward the vessel walls. just right, and cells will continue to be suspended as they rotate along a circular streamline, minimizing both shear and the net gravity vector to best approximate microgravity.
to properly quantify the angular speed, it turns out that wolf & schwarz modeled this system back in 1991; I’ll summarize the main points here, correct some unfortunate errors in the math, and include some additional information that I feel should have been included for clarity. if we approximate individual cells as roughly spherical (a decent approximation), and the containing medium is a newtonian liquid, we satisfy the conditions for a well-known case in fluid dynamics: steady, incompressible, creeping flow around a sphere. this case greatly simplifies the navier-stokes equations such that an analytical solution can be found while still remaining relatively accurate. when we derive the force balance, we find that net-zero force in a rotating reference frame is achieved when the sedimentation due to gravity equals the effect of the pseudo forces (centrifugal and coriolis), both radially and tangentially.
what matters here is that the number of vessel wall impacts, and therefore shear stress, is minimized. thus, we favor a net radial deviation from the streamline radius of zero. the math works out here such that the deviation induced by gravity has an inverse dependence on angular speed, while the deviation induced by the centrifugal force has a direct dependence on angular speed. the ideal theoretical angular speed, then, is the one in which these two opposing deviations are equal. as it turns out, the optimal rotation rate has a heavy dependence on the distance from the axis of rotation.
while the theory is quite elegant, it isn’t perfect. it relies on many assumptions, with several that may easily be challenged. for example, the no-slip condition at the cell envelope-medium interfaces does apply to most microbial systems at laminar flow; however, it is less confidently asserted for cell types with slippery lipid membranes (e.g., mammalian cells), extracellular matrices, and motility. various corrections have been applied in these special cases, including the use of the partial slip boundary condition, cell deformation modeling, non-newtonian rheological modeling, and other forms of fluid transport modeling. more obviously, cells will form aggregates in microgravity-like conditions, increasing the effective size of the particles and requires higher precision and accuracy in choice of angular speed.