A Harvard-led team turned an ordinary semiconductor into a tiny, parallel "DNA printer" that builds 64 genetic strands at once — without the harsh chemicals that dominate today's synthesis.
Making custom DNA is a routine but messy business. Drug developers, diagnostic labs and data-storage researchers all need short, precisely ordered strands of genetic code, and today that usually means flowing toxic solvents through plastic columns. A new chip suggests a quieter way: let electricity and water do the writing.
The device looks unremarkable — a silicon wafer etched with dozens of concentric ring electrodes. The trick is what happens in the thin film of water sitting on top of it. DNA is built letter by letter, and each addition depends on the local chemistry being just right. The team controlled that chemistry by running a tiny electric current through the rings, which nudged the water's pH up or down in precise, localized zones.
In the right pH pocket, a water-based enzyme dutifully added the next DNA base. Because each ring could be tuned independently, the chip grew 64 different sequences side by side in a single pass — strands up to about 39 letters long. The work, published in Nature Electronics, is the first demonstration of parallel enzymatic DNA synthesis driven entirely by a standard semiconductor chip.
Conventional synthesis relies on chemical reagents that are effective but generate hazardous waste and need careful purification. Enzymatic methods are cleaner, but scaling them has been hard. By using the chip itself to localize the reaction electrochemically, the researchers avoided shipping reagents across a big surface — each sequence gets its own miniature, addressable reaction chamber defined by voltage rather than by plastic walls.
The most immediate payoff is cleaner, cheaper, on-demand DNA for research and diagnostics — imagine a benchtop chip that prints the primers a lab needs in minutes instead of waiting on a shipment. Looking further out, DNA is being explored as an ultra-dense information medium, and a programmable chip-writer fits naturally into that vision.
The current limit of roughly 39 bases per strand is modest; many applications need longer, more complex sequences. But the architecture — standard silicon, water, enzymes, electric control — is the kind of unglamorous, scalable foundation that biotech likes. If the length barrier falls, the "DNA printer" could become as ordinary as the inkjet beside it.
The broader lesson is that biology's tools are increasingly being built from the same materials as our computers. When the same wafer can both compute and synthesize life's code, the line between chip and laboratory keeps getting thinner.