Microneedle Array Geometries for CGMs

A microneedle array's geometry decides whether a glucose sensor reaches the right layer of skin and how it feels going in. The variables are needle length, tip sharpness, spacing, needle count and material. Reviews describe microneedles as usually shorter than 1,500 micrometers (1.5 millimeters), set in arrays of up to hundreds [1]. The microneedle glucose sensor we found with FDA authorization, the Biolinq Shine, uses just seven microsensors, which sit in the papillary dermis, the top layer of the dermis [2]. The FDA granted it as a prescription device on September 23, 2025, and cleared it for over-the-counter use on September 18, 2026 [3][4]. Our piece on microneedle array sensor innovation covers Shine's clinical study and the wider field.

Length sets depth, and depth drives pain

The target is the dermis, where interstitial fluid, the fluid that surrounds cells, carries glucose the sensor can read [2]. Needles that stop in the upper skin usually don't reach the nerves and blood vessels deeper in the dermis, which is the main argument for keeping them short [1].

The clearest human data on length comes from a 2008 Georgia Tech study. Researchers inserted single microneedles 480 to 1,450 micrometers long, plus arrays of 5 or 50 needles, into the inner forearms of 10 healthy volunteers [5]. Every design hurt less than a 26-gauge hypodermic needle, scoring 5% to 40% of its pain score. Length mattered most. Tripling it raised the pain score sevenfold [5].

Here's what surprised us: in the same study, tip angle (tested from 20 to 90 degrees), needle width and needle thickness had no significant effect on pain [5]. Those features still matter, but for insertion and strength rather than comfort.

Biolinq's published patent application for a microneedle array sensor gives example needles about 500 micrometers tall and about 170 micrometers across. It describes an insertion depth of about 0.25 millimeters, so the electrodes sit in the upper dermis [6]. The filing argues that the upper dermal layers lack pain receptors and that shallow placement shortens the delay before glucose reaches the sensor [6]. A patent application records what a company filed. It doesn't show that Shine uses those dimensions, and the FDA decision summary for Shine lists no needle dimensions at all [2].

Tip sharpness controls insertion

A sharper tip goes in with less force and to a more predictable depth. In a 2004 study, the force needed to push microneedles into living skin rose in a straight line with the area of the needle tip. Measured forces ran from about 0.1 to 3 newtons, low enough to insert by hand [7]. The same team found that thicker walls made needles harder to break, and that the safety margin (fracture force divided by insertion force) grew with thicker walls and smaller tip radius [7].

A 2014 study in donated human skin showed why sharpness matters for depth. Needles with 5-micrometer tips slid in smoothly. Blunter tips, up to 37 micrometers, went in only superficially at first and then suddenly plunged deeper [8]. Insertion force rose with tip diameter, from 20 to 167 millinewtons, and the authors concluded that sharp tips are essential for reaching a chosen depth in a controlled way [8].

For a sensor, the tip has a second job: it carries the electrode. Biolinq's application describes solid microneedles with a tapered tip whose very point is insulated, with the electrode on the sloped surface set back from that point [6]. It shows columnar, pyramidal and hybrid columnar-pyramidal needle shapes [6].

Spacing, needle count and the applicator

Pitch is the center-to-center distance between neighboring needles. Biolinq's application says needles should be spaced to spread the force applied to the skin and avoid a "bed of nails" effect [6]. In its account, wider pitch tends to lower insertion force and deepen penetration, but pitch only starts to change insertion force below about 150 micrometers [6]. One layout it describes puts seven microneedles in a hexagon, about 750 micrometers apart [6]. That count matches the seven microsensors the FDA describes for Shine [2], but we found no primary source that gives Shine's actual layout.

Research arrays have gone much denser. One 2004 silicon design used a 20-by-20 grid of needles 200 to 350 micrometers tall, with 120-micrometer bases on 300-micrometer centers. A later three-electrode glucose patch packed about 9,500 silicon microneedles into each square centimeter [1]. More needles mean more possible sensing sites, but comfort pushes the other way: in the Georgia Tech study, ten times as many needles raised pain just over twofold [5].

Biolinq has also filed a patent application for an applicator that accelerates a microneedle array sensor toward the skin with a specified impact force and velocity [9]. Shine goes on with a single-use applicator, though the FDA summary doesn't describe how it works [2].

Materials and needle types

Microneedles have been made from silicon, glass, ceramic, metal, polymers and carbohydrates. They come in solid, hollow, dissolving, swellable and coated forms [1]. The review literature sorts the trade-offs roughly like this:

• Silicon is strong and can be etched into many lengths and shapes, but parts of silicon needles can break off under the skin [1].
• Metals such as stainless steel and titanium offer good mechanical strength [1].
• Polymers are biocompatible and easy to manufacture, and some are made to dissolve or swell in the skin [1]. An Imperial College London team cast a sensor array in an epoxy photoresist called SU-8, coated it with platinum and silver to form electrodes, and trapped the enzyme glucose oxidase in a polymer film. They tested it in one healthy volunteer for 3 to 6 hours [10].
• Hollow needles are harder to build. Tall, thin ones lack the internal support of solid needles, so they can bend or insert unevenly [1]. A team at KTH Royal Institute of Technology in Stockholm went the other way and put a tiny sensing probe inside a single hollow silicon microneedle. The part that entered the skin was 50 times smaller than in commercial sensors, and readings trailed capillary (fingerstick) blood by about 10 minutes [11].

Biolinq also holds a granted US patent on combining silicon-fabricated solid microneedle sensors with CMOS circuitry, the chip technology behind most electronics [12]. That shows what Biolinq claimed, not what Shine contains.

What geometry doesn't settle

Even well-shaped needles disturb the skin. Short-term studies of microneedle glucose sensors reported minor swelling and redness that cleared without treatment [13]. Shine's authorization is also narrow: up to five days of wear, color-coded glucose ranges instead of numbers, and adults 22 and older who don't take insulin [2][4]. Whether a sensor like that fits someone's diabetes care is a decision for their care team.

The biggest gap we found is public data: every pain, force and breakage figure above comes from lab studies or patent filings. As of October 2026, Shine's FDA decision summary names its seven microsensors and the skin layer they sit in, but not their length, tip shape or spacing [2]. So the geometry of the one authorized microneedle glucose sensor we found still isn't published.