Wearing the Water Harvester: Hierarchical Textile Fibers That Keep Their Speed at Scale
Original Authors: Chuxin Lei, Yongzheng Zhang, Lu He, Yuyang Wang, Juan Wu, Weixin Guan, Yaxuan Zhao, Qiang Fu, Keith P. Johnston, Kai Wu, Guihua Yu
Original paper is accessible at: https://doi.org/10.1126/sciadv.aed9949
The scale-up problem in sorption-based AWH
Sorption-based atmospheric water harvesting (SAWH) has produced impressive materials. MOFs, salt composites, and hydrogels can capture large amounts of water quickly when tested as milligram-scale samples in a laboratory instrument.
Real devices are different. A practical harvester needs hundreds of grams to kilograms of sorbent. Once a sorbent is packed into a thick bed or large panel, vapor has to travel further, mass transfer resistance rises, and the fast kinetics measured in the laboratory largely disappear. This gap between small-sample performance and device performance is one of the main obstacles to practical SAWH.
This paper asks:
Can a sorbent be designed so that its fast kinetics survive scale-up, and even be worn?
The authors' answer is to build the sorbent as a fiber and then weave it into a textile.
Why fibers, and why ordinary fibers are not enough
Fibers have a useful property: structure engineered at the microscale is repeated continuously along their length. When fibers are woven, the textile has interconnected air channels between them, so vapor can reach every fiber even in a large or multilayer fabric.
Within a single fiber, though, there is a problem. Conventional wet-spun fibers usually have a skin–core structure: a dense, nonporous outer surface and a uniform nanoporous interior. Vapor must first condense on the surface and then diffuse inward. A dense skin limits surface condensation and slows the inward transport. That limitation at the level of one fiber carries over to the whole textile.
The fiber therefore needs two features at once:
an open, porous surface where vapor can condense quickly
a hierarchy of internal pores that pulls liquid water inward fast
Building the hierarchy with a molecular switch
The authors make these fibers, called hierarchical open porous fibers (HOP-Fibers), by mixing cellulose with hydroxypropyl cellulose (HPC), an amphiphilic polymer.
In plain water, HPC chains stay extended because they form hydrogen bonds with water. In the strongly alkaline LiOH/urea solution used to dissolve cellulose for spinning, hydroxide ions weaken those hydrogen bonds. HPC chains then fold onto themselves through hydrophobic interactions, in what is called a coil-to-globule transition.
These HPC globules stay stable in the spinning solution, so the dope can still be spun. During wet spinning they form phase-separated domains. When the solvent is removed, these domains leave behind large pores both on the fiber surface and inside it, alongside the smaller pores of the cellulose network.
Two controls tune the structure:
LiOH concentration: below about 1 wt%, HPC barely changes and the fiber has uniform nanopores. Higher concentrations drive stronger phase separation and larger pores.
HPC-to-cellulose ratio: more HPC produces a more pronounced hierarchy. The best version, HOP-Fiber-4, uses an HPC/cellulose ratio of 3:5.
The fibers are then loaded with LiCl for hygroscopicity and freeze-dried.
How water moves through the fiber
The authors describe sorption in three stages, supported by Gibbs free energy analysis, optical microscopy, and lattice Boltzmann flow simulations:
Surface liquefaction. Open surface pores lower the free energy barrier for vapor to condense. Under the microscope, HOP-Fibers formed many microdroplets within minutes at 30% RH.
Intake through large pores. Liquid water enters the large pores first because they offer lower flow resistance.
Redistribution into small pores. Capillary (Young–Laplace) pressure then draws water from large pores into small pores, where it is stored.
In short, the large pores act as fast entry channels and the small pores act as storage.
Single-fiber performance
Tested in a dynamic vapor sorption (DVS) instrument at 25 °C, HOP-Fiber-4 achieved:
1.16 g/g at 15% RH
1.50 g/g at 30% RH
2.68 g/g at 60% RH
It also absorbed water quickly. It reached 80% of its saturated uptake in 64 min at 15% RH, 19 min at 30% RH, and 20 min at 60% RH, which the authors report as faster than most state-of-the-art sorbents.
Water was released with mild heating at 60 °C. At 30% RH, the fiber completed 36 sorption–desorption cycles per day, with about 20 minutes per cycle, and performance remained stable.
The key result: performance that survives scale-up
This is where the paper differs from many sorbent studies. The authors tested woven HOP-Fiber-4 textiles of increasing size and thickness, each for one hour at 30% RH:
A 400 cm² textile captured 1.04 g/g, about 75% of the value for a 1 cm² sample.
A 24-layer textile still reached 0.96 g/g.
Compared with conventional sorbents at similar mass and humidity, the woven textiles showed a 3- to 10-fold improvement in water uptake rate.
Manufacturing also scaled. Using a 20-nozzle spinneret on industrial-type equipment, the team produced HOP-Fiber-4 continuously at 509 m per hour, replacing LiOH with lower-cost NaOH. The scaled-up fibers performed comparably to laboratory fibers. They were then woven into square-meter-scale textiles that tolerated rolling, folding, and twisting.
A harvesting jacket in the field
To demonstrate practical use, the authors built a jacket with four detachable moisture-capture units, two of 0.04 m² and two of 0.01 m², each covered with porous PTFE. The wearer can run, climb, or hike while the units capture moisture.
Water is collected in a separate foldable, inflatable collection unit. The textile units are placed on a flexible DC-powered heater at 60 °C inside the sealed unit, and the released vapor condenses on a hydrophobic inner surface and drains into a collection bag. The whole system packs into a portable bag.
Field tests used continuous cycles of 1.5 hours of sorption and 1 hour of desorption:
Xichang, China (arid): 3.76 L per kg of sorbent per day (4.10 L/m²/day), or 410 mL per day
Additional tests in Austin, USA (semi-arid) and Chengdu, China (humid)
Across 20–80% RH: 3.76 to 7.45 L/kg/day and 410 to 894 mL per day
The collected water had minimal lithium residue and met WHO drinking water guidelines for that parameter. The authors also used global average RH data to estimate the jacket's water production potential around the world.
Open questions
Energy for desorption. Capture is passive while the jacket is worn, but release uses a DC-powered heater at 60 °C. The paper does not report the energy consumed per liter. For an off-grid wearable, the power source and its weight are part of the real system. The authors point to solar-thermal and low-grade heat regeneration and on-body energy harvesting as future directions.
Practical scale. The reported numbers imply roughly 0.1 m² of textile and about 110 g of sorbent per jacket. That is a useful supplementary supply, but it is well below the daily drinking water requirement of an active person in a hot, dry place.
Operating routine. The field protocol cycles every 2.5 hours, with the units moved to the collector for each desorption. How that fits into real use while traveling or working remains to be shown.
Water quality. Lithium is reported. A broader panel, including microbiology and possible leaching from the fibers and the PVC condenser, would strengthen potability claims.
Long-term durability of LiCl retention, especially with motion, sweat, abrasion, and washing, still needs testing.
Why this paper matters
The main contribution of this work is not a single uptake number. It addresses directly the problem that limits most SAWH materials: fast kinetics in the laboratory but slow kinetics at device scale.
By controlling pore structure at the molecular level during spinning, and then using weaving to keep vapor pathways open across large and thick textiles, the authors keep most of their small-sample performance at the scale of square meters. They also show that the fibers can be manufactured continuously with industrial methods.
The wearable jacket is an attention-grabbing demonstration, but the broader lesson applies to any AWH device: design the sorbent architecture for scale from the start. Fibers and textiles may become a practical format not only for clothing but also for tents, shelters, and panels where large, breathable sorbent surfaces are needed.





Comments