Paper-based microfluidic devices
are popular for their ability
to automate multistep assays for chemical or biological sensing at
a low cost, but the design of paper microfluidic networks has largely
relied on experimental trial and error. A few mathematical models
of flow through paper microfluidic devices have been developed and
have succeeded in explaining experimental flow behavior. However,
the reverse engineering problem of designing complex paper networks
guided by appropriate mathematical models is largely unsolved. In
this article, we demonstrate that a two-dimensional paper network
(2DPN) designed to sequentially deliver three fluids to a test zone
on the device can be computationally designed and experimentally implemented
without experimental trial and error. This was accomplished by three
new developments in modeling flow through paper networks: (i) coupling
of the Richards equation of flow through porous media to the species
transport equation, (ii) modeling flow through assemblies of multiple
paper materials (test membrane and wicking pad), and (iii) incorporating
limited-volume fluid sources. We demonstrate the application of this
model in the optimal design of a paper-based signal-enhanced immunoassay
for a malaria protein, PfHRP2. This work lays the foundation for the
development of a computational design toolbox to aid in the design
of paper microfluidic networks.