water · energy · family: making one is easy. making a million is the problem
one atom thick,impossible to manufacture
Nanoporous graphene membranes cannot scale from lab to industrial desalination
Problem statement
Single-atom-thick graphene membranes with sub-nanometer pores can, in simulations, filter salt from seawater at a fraction of the pressure conventional reverse osmosis needs — the Lockheed patent puts the reduction at a factor of five or more — but no one has yet manufactured them at the scale needed for real desalination plants. Lockheed Martin patented its "Perforene" graphene membrane technology in 2013, but a 2018 review of the field found single-layer graphene membranes still confined to centimeter-scale laboratory devices, Lockheed's own product page today still puts individual devices at several square centimeters, and the patents have since lapsed. The core problem is that creating uniform nanopores (about 0.5–1 nm across) in large-area graphene sheets free of substantial defects is extremely difficult, and tears and defects in the single-atom-thick membrane let salt leak through.
Why this matters
State-of-the-art seawater reverse osmosis consumes about 2.2 kWh per cubic meter of freshwater produced, close to the theoretical minimum of about 1.8 kWh per cubic meter at 50% recovery, and energy is about 40% of the total cost of seawater desalination. A more permeable membrane operating at lower pressure could cut that energy demand by perhaps 10–20%, which the 2018 review translates into a cost saving of only about 5–10%; the review argues that effort should go instead to pretreatment, fouling and plant design. Google Patents lists both patents named here as expired for non-payment of maintenance fees, even though Lockheed Martin's product page still describes Perforene as expanding out of the lab and partnering with outside organizations on real-world applications.
What’s been tried and why it hasn’t worked
Lockheed Martin's patents describe perforating a graphene sheet with apertures of about 0.6 nanometers (to block sodium ions) or 0.9 nanometers (to block chloride ions), spaced about 15 nanometers apart, by etching in dilute oxygen at elevated temperature, laser drilling, or a self-assembling block-copolymer mask with reactive ion etching. Academic groups have grown single-layer graphene by chemical vapor deposition (CVD) on copper foil and opened pores with oxygen plasma or gallium-ion bombardment followed by chemical etching. At centimeter scale, transferring a single layer needs careful processing to avoid tears and extrinsic defects (100–200 nm) through which leaking occurs, and intrinsic defects (1–15 nm) from growth limit sub-nanometer pore control; the patent itself notes that as pore density rises the sheet weakens locally and can rupture. Researchers have also explored graphene oxide (GO) laminate membranes, which are cheaper and easier to produce, but fine control of the interlayer spacing within the sub-nanometer range remains challenging, and for desalination the spacing must be decreased and stabilized. The 2018 review found that most of the papers it examined did not discuss or demonstrate significant improvements in the critical real-world factors: scaling, fouling, and chemical or thermal stability. Lockheed's claim of a membrane that would desalinate water at a fraction of the cost of standard reverse osmosis was criticized by the editor of the Water Desalination Report.
What would unlock progress
A breakthrough in large-area graphene synthesis with atomic-level pore control would change everything — this likely requires moving beyond CVD to a scalable, defect-tolerant manufacturing process. Alternatively, hybrid approaches combining graphene with conventional polymer membranes could capture some energy savings without requiring perfect graphene. Advances in self-assembling nanoporous materials or block copolymer templating might offer paths to uniform sub-nanometer pores at scale.
Entry points for student teams
A student team could design and test a hybrid membrane combining graphene oxide flakes with a polymer support, measuring salt rejection and water flux as a function of GO layer thickness and cross-linking chemistry. This is feasible with standard lab membrane casting equipment and bench-scale filtration cells. Skills in materials science, membrane characterization (SEM, permeability testing), and water chemistry would be most relevant. A team without a wet lab could start at the desk: take the pore sizes, spacing and five-fold pressure reduction claimed in the Lockheed patents named above and the permeability and salt-rejection comparison of commercial and nanostructured membranes in the 2018 npj Clean Water review, and work out what the review's warning implies — that a full-scale plant may not be designed with membranes more permeable than today's, so that new plant designs, not just new membranes, would be needed.
Genes
Structural cousins: stuck the same way, in other fields
Sources
US9193587B2, "System and method for water purification and desalination," Lockheed Martin Corporation, Google Patents, accessed 2026-02-10. Related: US8361321B2, "Perforated graphene deionization or desalination." Both Expired – Fee Related. go to source ↗
Working notes
Notes made while writing and checking this brief, shown as written. They use working shorthand.
Related patents: US8361321B2, "Perforated graphene deionization or desalination," inventors John B. Stetson, Jonathan Mercurio, Alan Rosenwinkel and Peter V. Bedworth, granted 2013-01-29 (also Expired – Fee Related). US9193587B2 itself, inventor Rex Bennett, was granted 2015-11-24 and describes a hand-pressed canister purifier whose membrane includes perforated graphene with 1–50 nm openings. Neither patent uses the name Perforene; the 2018 review (Boretti et al., "Outlook for graphene-based desalination membranes," npj Clean Water 1:5, 2018, https://www.nature.com/articles/s41545-018-0004-z) is the source that says Lockheed Martin patented "Perforene" in 2013; US8361321B2 is the Lockheed graphene-desalination patent granted that year. That review concluded that the state of the field "places graphene-based membranes at a quite far in time potential commercial viability." The Perforene product page on Lockheed Martin's website still exists but has not been updated with commercial availability. It was read on 2026-09-30 at https://www.lockheedmartin.com/en-us/products/perforene-graphene-membrane.html; it says individual devices can be as large as several square centimeters, that Perforene is partnering with outside organizations on real-world applications, and gives no commercial availability; the datasheet address the review cites returns 404. This problem connects to the broader challenge of translating nanomaterials research into industrial products.
Reconciliation 2026-09-30: Fact repair after reading the whole write-up against every source it names, by a first reader and then a second reader on the same day. Fetched on 2026-09-30 and read in full: the Google Patents pages for US9193587B2 (https://patents.google.com/patent/US9193587B2/en) and US8361321B2 (https://patents.google.com/patent/US8361321B2/en); the 2018 npj Clean Water review named in the notes above (Boretti et al., https://www.nature.com/articles/s41545-018-0004-z, open access); and Lockheed Martin's Perforene product page (https://www.lockheedmartin.com/en-us/products/perforene-graphene-membrane.html, HTTP 200, which the first reader had looked for at two other addresses and reported missing). The Water Desalination Report item the review cites (desalination.com/issues/2420, 2014 issue 9) is paywalled; only its headlines could be read. (1) The text said graphene membranes could filter seawater "using 100x less energy" than reverse osmosis. No source says this. US8361321B2 says simulated results show the perforated graphene sheet reduces the required pressure "by at least a factor of five"; the sentence now says that. (2) "Predicted commercialization by 2014–2015" is in no source and was removed. "Over a decade later the technology remains confined to centimeter-scale lab demonstrations" now says that the 2018 review found single-layer membranes confined to centimeter-scale laboratory devices and that Lockheed's product page today puts individual devices at several square centimeters (the page's words: "individual devices can be as large as several cm^2"); the page also says Perforene is "expanding out of the lab," so "lab demonstrations" is not claimed for today. (3) The pore size "1 nm diameter" became "about 0.5–1 nm": the review gives 0.5÷1 nm and the patent 0.6 and 0.9 nanometers. "Any defect ... renders it useless" is not in the sources; the review says tears and defects let leaking occur, and the sentence now says that. (4) Removed because no source contains them: over 2 billion people without safely managed drinking water; desalination as the only option for arid coastal regions; 3–6 kWh per cubic meter; energy as the dominant cost; a cut of "20% or more"; viability for communities that cannot afford it; and "$50B" for Lockheed Martin. The paragraph now uses the review's figures: about 2.2 kWh per cubic meter for state-of-the-art seawater reverse osmosis, a minimum of about 1.8 kWh per cubic meter at 50% recovery, energy about 40% of total cost, an energy gain of 10–20% and a cost gain of 5–10% from a more permeable membrane, and the review's call to put effort into pretreatment, fouling and plant design instead. (5) "A signal that even a $50B defense contractor could not find a path to commercial viability" is an inference no source makes, and Lockheed's product page still describes Perforene as partnering with outside organizations on real-world applications; the sentence now says only that Google Patents lists both patents as expired for non-payment of fees while the product page still describes Perforene that way. "Both its core Perforene patents" became "both patents named here," because neither patent uses the word Perforene. (6) The text said Lockheed's approach used CVD on copper, then ion bombardment or plasma etching, that CVD is costly and gives mainly multilayer graphene, and that bombardment and plasma etching make irregular pores. The patents and the product page mention none of these methods; US8361321B2 describes oxygen etching at elevated temperature, laser drilling, and a block-copolymer mask with reactive ion etching, with 0.6 and 0.9 nanometer apertures spaced 15 nanometers apart, and warns that higher aperture density can locally weaken the sheet and cause rupture. CVD on copper foil, oxygen plasma and gallium-ion bombardment with chemical etching are the academic methods the review describes, and the review's centimeter-scale figures (extrinsic defects 100–200 nm, intrinsic defects 1–15 nm) replace the unsupported ones. (7) "GO membranes swell in water, changing their pore geometry unpredictably" is in no source; the review says GO is cheaper and easier to produce, that fine control of the interlayer spacing in the sub-nanometer range is challenging, and that for desalination the spacing must be decreased and its stability increased. The sentence now says that. (8) "Most peer-reviewed studies have not demonstrated improvements" now says most of the papers the review examined did not discuss or demonstrate significant improvements, which is the review's wording. (9) Quotation not found: the words "ridiculous and very premature" attributed to the Water Desalination Report editor appear in no source that could be read; the review says only that Lockheed's claims "were criticized by the editor of the Water Desalination Report," the Report item itself is paywalled, and the sentence now paraphrases the review. (10) Invented quotation: the review does not say "at a quite distant potential commercial viability"; its words are "places graphene-based membranes at a quite far in time potential commercial viability," and the notes now quote that. (11) The claim that Lockheed pivoted Perforene to oil and gas wastewater treatment needing 50–100 nm holes is in no source and was removed (the product page says only that it typically perforates from single-digit nanometers up to 100 nanometers, with no application named). (12) The notes now give the two patents' inventors and grant dates and the review's citation, copied from the records, and record that US9193587B2 describes a hand-pressed canister purifier with graphene perforated at 1–50 nm. The first reader had written one inventor for US8361321B2; the record lists four, and the notes now name all four. (13) A desk-based entry point was added, built only from the patents and the review already named, so that the section holds two doors. Checked by the second reader and put back as originally written: "The Perforene product page on Lockheed Martin's website still exists but has not been updated with commercial availability" (the first reader had replaced it with a report that the page could not be found; it is live, and its address is now given). Confirmed in the sources and left as written: the Source line (both patent numbers, titles and assignee; both shown as Expired – Fee Related); Perforene patented in 2013 (the review's statement, and US8361321B2's 2013-01-29 grant); the review's point that the mass production of large defect-free single-layer sheets is extremely difficult at present; block-copolymer templating as a route to uniform pores (named in US8361321B2). Not checked, because the item is paywalled: whether the Water Desalination Report editor used the words "ridiculous and very premature." No title changed, though both titles say more than the sources do ("Cannot Scale" and "Impossible to Manufacture," where the review says extremely difficult at present and not expected in the short term); this is left for Beth. Genome tags untouched.