To appreciate why this matters, it helps to understand what a nanopore sensor actually does. The principle is deceptively simple: a membrane thin enough and porous enough to present a single nanoscale opening separates two fluid reservoirs. When a voltage is applied, ions flow through the pore, producing a steady baseline current. When a molecule—say, a strand of DNA or an individual protein—passes through the opening, it briefly blocks or modulates that current. Because each molecule disturbs the ionic flow in a way that reflects its size, shape, and charge, the sensor can identify and characterize molecules one at a time, without the chemical labeling or amplification steps that many conventional assays require. The smaller and more precisely controlled the opening, the finer the discrimination the device can achieve.

The work, led by doctoral student Xinxin Liu and professor Frank Niklaus of KTH’s Division of Micro and Nanosystems, addresses a well-recognized bottleneck in the fabrication of nano-scale openings in thin membranes. A typical sensor device uses a membrane with a single nanopore, which functions as a detector for individual molecules. Because these openings must be carved with extraordinary precision, they are currently produced one at a time using techniques such as electron-beam drilling and sculpting with a Transmission Electron Microscope (TEM). Both approaches are slow, expensive, and difficult to scale to industrial volumes—limitations that have kept nanopore-based sensing from reaching its full potential outside of well-funded laboratories.

The economics of these legacy methods are worth spelling out. An electron microscope capable of drilling a hole a few nanometers wide represents a capital investment in the millions of dollars, and each pore requires an operator-guided session of beam exposure, imaging, and verification. A skilled technician might spend hours producing a handful of usable devices, and every additional pore multiplies the time and cost. That arithmetic is tolerable when the goal is a research prototype; it becomes prohibitive when the goal is thousands or millions of sensors for clinics, laboratories, or field-deployable diagnostic kits. The mismatch between how nanopores are made and how many of them the growing field demands has been one of the quiet constraints on the technology’s expansion.

Niklaus says the study demonstrates a fundamentally different approach: using built-in mechanical stress to create the nanoholes. Rather than removing material with a beam of electrons or ions, the KTH team lets physics do the work. The technique exploits fracture mechanics—the same rules that govern how materials crack and tear—to generate openings in the sensor membrane, and it does so across many sites on a wafer simultaneously, producing large arrays of nanopores in parallel.

The process begins with a stack of layers deposited on a silicon wafer. The top layer is deliberately placed under tension, stretched like a drawn rubber band. Beneath it sits a second layer of material that is etched away until only a small, pre-defined shape remains. This remnant acts as an anchor and is meant to define the shape of the nanopore that will ultimately form in the third layer underneath—the sensor membrane itself. When the tension in the top layer is released, the stressed layer behaves like a rope attached to that anchor.

“The beam pulls on until it tears a nanoscale fragment from the membrane, creating the hole,” Liu says. She offers a vivid analogy for the mechanism: it is much like removing a tree stump with a rope. The stressed beam attaches to the nanoscale anchor, pulling it away from the membrane, which serves as the ground. As the anchor is wrenched free, it tears a fragment out of the membrane, leaving behind a hole whose dimensions are dictated by the lithographically defined anchor shape. The image of uprooting stumps is more than a flourish—it captures the essential physics of a tensile failure event harnessed as a manufacturing step.

The elegance of the scheme lies in how it converts an engineering liability into an asset. Mechanical stress in thin films is ordinarily something chipmakers work hard to manage and minimize, because uncontrolled stress can warp wafers, delaminate layers, and crack devices. Instead of suppressing it, the KTH team engineers it deliberately, tuning the tension so that fracture occurs exactly where and when it is wanted. Every structural element involved—the stressed beam, the sacrificial anchor, the membrane—is defined by standard lithographic patterning, which means the geometry of each future pore is set down in advance by the same design files that govern the rest of the chip. Fracture, a phenomenon usually treated as a failure mode, becomes a deterministic manufacturing step.

What makes the result notable is not just the elegance of the analogy but the breadth of its applicability. The researchers tested the technique successfully in multiple materials, including dielectric, metallic, and semiconductor sensor membranes. That versatility matters because different sensing applications favor different membrane materials, and a method confined to a single material system would have limited reach. Some sensing chemistries work best with insulating dielectrics; others benefit from the conductive properties of metals or the semiconductor behavior of silicon-based films. The study also shows that the resulting nanopores were effective at DNA analysis and molecule detection, demonstrating that pores made this way are not merely structural curiosities but functional components of working sensors.

The significance of parallel fabrication is hard to overstate for the field. Nanopores are valuable tools because their extremely small openings in ultra-thin materials can be used to detect and analyze individual molecules such as DNA and proteins. They underpin a range of sensing and analysis technologies used in medical research and, increasingly, in specialized clinics where sequencing systems are deployed for outbreak surveillance, cancer genomics, and infectious disease identification, among other applications. If the pores at the core of these instruments can be mass-produced on standard semiconductor lines rather than sculpted one by one under an electron microscope, the cost and availability of such systems could shift substantially.

The route toward commercialization is already under way. Liu says the researchers have patented the technique and formed a startup to progress it toward commercial implementation. “We have already shipped samples to a collaborator for testing, so it has moved beyond the lab stage,” she says. That step—from demonstration to external evaluation—is often where promising fabrication methods stall, and the team’s early move to put devices in a collaborator’s hands suggests confidence that the process can survive contact with real-world requirements.

The publication itself carries the hallmarks of a carefully documented study. Appearing in a peer-reviewed journal with a formal article title and digital object identifier, the work positions the stress-driven fracture method alongside established nanofabrication techniques as a credible alternative rather than a laboratory demonstration destined to remain on the bench. The researchers’ framing of the method as compatible with conventional semiconductor manufacturing is central to its promise: wafer-scale processes, layer deposition, and lithographic patterning are the stock-in-trade of the chip industry, and a nanopore method built from those same ingredients inherits decades of industrial maturity. Semiconductor fabs routinely pattern features far smaller than the anchors and beams involved here, deposit thin films with precisely controlled stress states, and inspect finished wafers with automated optical tools. Borrowing that infrastructure means the marginal cost of each additional pore on a wafer falls toward the cost of the lithography itself—which is to say, toward very little.

For the clinics and research laboratories that depend on nanopore sensing, the implications could be far-reaching. Sequencing systems used for outbreak surveillance need to be deployable quickly and at scale; cancer genomics and infectious disease identification benefit from lower per-instrument costs; and medical research labs without access to specialized nanofabrication facilities could gain access to sensors that were previously out of reach. A method that turns nanopore fabrication from an artisanal craft into a batch process aligns with exactly those needs. It also opens the possibility of designing devices with many pores operating in parallel on a single chip, multiplying throughput in the same way that arrays of sensors have transformed other areas of diagnostics.

There are, of course, caveats worth keeping in view. The study reports pores in the range of roughly 6 to 10 nanometers, which is well suited to many single-molecule applications, but the full distribution of pore sizes achievable across a wafer, the long-term stability of stress-engineered membranes, and the yield of usable devices in high-volume production remain questions that commercialization will need to answer. Single-molecule sensing is unforgiving of variability: even small differences in pore dimensions can shift signal characteristics from device to device, and any manufacturing method intended for scale must demonstrate tight, repeatable control. Fracture, for all its determinism in principle, is a physical event with inherent statistical character, and characterizing how tightly it performs across thousands of sites on a wafer will be an essential part of the industrialization story. The researchers’ own decision to ship samples for external testing acknowledges that laboratory performance must be validated by independent users before the method can be considered proven at scale.

Still, the core achievement stands: a way to make nanoscale holes by pulling, rather than drilling—harnessing built-in stress and fracture mechanics to tear open pores defined by lithography, in parallel, across multiple material systems. Like the tree stump that yields to a well-anchored rope, a stubborn fabrication bottleneck has been dislodged by applying force in the right place. If the technique continues to advance through patenting, startup development, and collaborator testing, the sensors that detect individual DNA strands and proteins may soon be made the way computer chips are: by the wafer, in the thousands, far from the cleanrooms where they were once coaxed into existence one pore at a time.