{"id":3609,"date":"2026-10-23T09:37:26","date_gmt":"2026-10-23T01:37:26","guid":{"rendered":"http:\/\/www.noithatdonghai.com\/blog\/?p=3609"},"modified":"2026-10-23T09:37:26","modified_gmt":"2026-10-23T01:37:26","slug":"how-is-an-echelle-grating-manufactured-4488-686fcc","status":"publish","type":"post","link":"http:\/\/www.noithatdonghai.com\/blog\/2026\/10\/23\/how-is-an-echelle-grating-manufactured-4488-686fcc\/","title":{"rendered":"How is an Echelle Grating manufactured?"},"content":{"rendered":"

Hey everyone, if you\u2019ve ever wondered how those super powerful Echelle gratings end up in every spectroscopy lab, astronomical observatory, and even some fancy semiconductor manufacturing lines\u2014you\u2019re in the right spot. I\u2019m [skip the company name, just stick to being your go-to grating supplier], and I\u2019ve spent the last 15 years standing right next to our production line, troubleshooting weird hiccups during coating, and cheering when we nail a perfect groove density that makes a customer\u2019s experiment work for the first time. So today, I\u2019m breaking down exactly how we make these things\u2014no stuffy textbook jargon, just the real, messy, sometimes weird process that turns blank glass into the backbone of so much important work. Echelle Grating<\/a><\/p>\n

<\/p>\n

First off, let\u2019s get one thing straight: Echelle gratings aren\u2019t your run-of-the-mill holographic gratings you buy for a high school physics project. Those are for seeing basic spectra of light bulbs. Echelles? They\u2019re the heavy hitters. They have way coarser grooves (like 30 to 300 grooves per mm, compared to the tiny 1,200+ you see in regular gratings) and a super steep, blazed angle\u2014usually around 75 degrees. That\u2019s what lets them spread light across multiple orders so efficiently, which is why astronomers use them to split starlight into tiny, precise bits to figure out what stars are made of, and why chemical analysts use them to detect parts per billion of pollutants in water samples. If you mess up even one tiny step in making an Echelle, the whole thing\u2019s useless. I\u2019ve seen a whole batch get scrapped because we nicked a groove with a diamond tool that was worn out\u2014ruined 3 weeks of work, all because I forgot to check the tool\u2019s edge that morning. Oops.<\/p>\n

Let\u2019s start with the very beginning, because you can\u2019t make a top-tier Echelle out of a random piece of soda-lime glass. We use what\u2019s called \u201clow-expansion borosilicate crown glass\u201d (wait, but I\u2019ll just say \u201cULE glass\u201d for short, since that\u2019s what everyone in the game uses\u2014stands for Ultra-Low Expansion). Why? Because Echelles have to stay perfectly stable, no matter the temperature shifts, like when an observatory\u2019s telescope cools down overnight or a lab\u2019s laser heats up the setup. If the glass expands or shrinks even a tiny bit, the grooves get misaligned, and your spectrum is garbage. We order it in big 12-inch square blanks first, and the first step for us is cutting it down to the exact size the customer needs\u2014like 2 inches round, 4 inches square, whatever. Then we do a whole lot of polishing, way more than regular glass. We\u2019re talking mirror-level flatness\u2014like, less than a quarter of a wavelength of light across the entire surface. I\u2019ve been on the polishing line at 2 a.m. once when we had a last-minute order for a satellite, and the spec was so tight we had to stop and re-polish three times because a single speck of dust landed on the surface. Yeah, that\u2019s the kind of detail we live with.<\/p>\n

Next up is the part that people usually think is the magic: ruling the grooves. But wait\u2014don\u2019t confuse this with old-school mechanical ruling, where you drag a diamond tool across glass over and over. We still use ruling engines, but they\u2019re not your grandpa\u2019s contraption. Ours is a 20-ton machine in a temperature-controlled room that never varies by more than 0.01 degrees Celsius. Any tiny temperature shift would throw the ruling alignment off. The ruling head holds a diamond tool that\u2019s sharpened to a point so fine, you could barely see it under a microscope. And here\u2019s the thing: to make an Echelle\u2019s steep groove angle, you don\u2019t just move the diamond straight back and forth. The ruling engine tilts the blank at that 75-degree blaze angle while it\u2019s cutting, so each groove has that sharp, precise slope that makes the Echelle work. But here\u2019s the hard part: we\u2019re ruling thousands of grooves in a straight line, over and over, across a surface that could be 6 inches wide. If the machine stutters even once, or the diamond gets a tiny chip, the whole thing has to get scrapped. Last year, we had a diamond tool chip mid-ruling, and we didn\u2019t even notice until the customer called a week later saying their spectrum had weird lines. We had to apologize, remake the whole order, and now we check each tool under a scanning electron microscope before we even load it into the ruling engine. No exceptions. That\u2019s the kind of mistake you don\u2019t get to make twice.<\/p>\n

Once the grooves are ruled, we have a huge problem: the glass itself doesn\u2019t reflect enough light, especially at the wavelengths Echelles are used for (like UV, visible, and near-IR). So we need to coat the grooves with a super thin layer of metal, usually aluminum, because it\u2019s cheap and reflective, but sometimes gold or silver depending on what wavelength the customer needs. Wait, but it\u2019s not just a thin layer\u2014we do what\u2019s called \u201covercoating\u201d sometimes, with a tiny layer of magnesium fluoride, to protect the metal from scratches or oxidation. How do we do this coating? In a vacuum chamber, of course. We pump out all the air, because air molecules would mess up the metal vapor as it\u2019s being deposited. Then we heat the metal (or whatever material we\u2019re using) until it turns to vapor, and it condenses onto the grating surface. But here\u2019s the cool part: we don\u2019t just coat the flat surface\u2014we coat the sides of the grooves too, which is super important for the blaze. The vacuum coating process has to be uniform across the entire surface, so every groove gets the same thin layer. If one side of a groove has a thicker coating, the light reflects unevenly, and you lose efficiency. We test every coated grating with a laser to check reflectivity, and if it\u2019s off by even 1%, we send it back for a redo. I\u2019ve seen a batch of 10 gratings get sent back because the coating machine had a slight leak during one run\u2014we had to re-coat all of them, and it set us back 10 days. Not fun, but it\u2019s worth it to keep customers happy.<\/p>\n

Wait a second\u2014what about holographic Echelles? Some suppliers use that method now, right? Instead of ruling grooves with a diamond, they use two lasers to create an interference pattern on a photoresist-coated glass, then etch the pattern into the surface. I won\u2019t lie, that\u2019s faster for high volume, but the groove angles aren\u2019t as precise as mechanically ruled Echelles, and they have higher stray light\u2014meaning extra light that gets scattered and muddles the spectrum. So we still stick to mechanically ruled Echelles for most of our customers, especially the ones doing high-resolution work like exoplanet hunting or pharmaceutical testing. We do some holographic ones for lower-end applications, but 90% of our orders are ruled. I know some folks in the industry swear by holographic, but at the end of the day, our customers care about performance, and ruled Echelles just perform better.<\/p>\n

Now, after coating, we have to test every single grating. This isn\u2019t like buying a laptop, where you turn it on and hope it works. We test each Echelle in a custom-built setup that shoots light of known wavelengths at it, measures how much light is diffracted into each order, and checks for any stray light or groove defects. We also check the \u201cefficiency\u201d (how much of the incoming light actually ends up in the desired order, not scattered) and the \u201cblaze wavelength\u201d (the specific wavelength where it\u2019s most efficient). For an Echelle made for a customer who\u2019s studying hydrogen-alpha lines, we need the blaze to be exactly at 656 nm\u2014if it\u2019s off even by 10 nm, their data is useless. That\u2019s why we spend so much time on testing. I once worked with an astronomer who was testing a new exoplanet, and his Echelle was off by 5 nm, so he missed the tiny spectral line that told him the planet had water. We re-made the grating in 5 days, and he was able to get the data he needed. That\u2019s the stuff that makes all the late nights and scrapped batches worth it.<\/p>\n

But wait, there\u2019s one step I almost forgot: mounting and packaging. Echelles are delicate\u2014scratch the coating, nick the grooves, and they\u2019re toast. So we mount them in custom holders that are designed to keep them perfectly flat, even when they\u2019re being used in a vacuum chamber on a satellite or in a harsh industrial environment. The holders are usually made of Invar, a metal that doesn\u2019t expand much with temperature, just like the glass. Then we package them in custom wooden boxes lined with foam that\u2019s static-dissipating, because static electricity can scratch the coating. We even add a desiccant packet to make sure no moisture gets in, no matter how long they\u2019re in shipping. Last year, we had a customer in Antarctica, and their shipping time was 6 weeks, so we added extra desiccant and double-boxed it. It arrived perfectly, and they sent us a photo of it next to their telescope in freezing temperatures. That\u2019s the customer love we live for.<\/p>\n

Now, let\u2019s talk about common mistakes we see (and make) that ruin Echelles. First, dust. A single dust particle on the ruling engine can nick the diamond tool, and on the finished grating, it can block a whole line of grooves. We have our production rooms filtered to 100-class cleanroom standards\u2014meaning no more than 100 dust particles per cubic foot of air. That\u2019s way stricter than most office buildings. Second, worn tools. We used to skip checking the diamond tool every batch, but after that chip mistake I told you about, now we check it every single time under a microscope. Third, bad testing. We used to test only a sample of each batch, but we had a customer get a grating that had a tiny groove defect on the edge, which ruined their experiment, so now we test every single one. Yeah, it takes longer, but it\u2019s worth it.<\/p>\n

So how does this compare to other gratings? Let\u2019s keep it simple. Regular gratings (like the ones for spectrometers you use in college) have high groove density and shallow blaze angles, designed for lower resolution. Echelles? Low groove density, steep blaze angle, high resolution, and they work over a wide wavelength range. That\u2019s why they\u2019re the go-to for any application that needs super precise spectrum data.<\/p>\n

<\/p>\n

If you\u2019re reading this and you\u2019re someone who needs an Echelle grating for your work\u2014whether you\u2019re building a new astronomical spectrograph, testing semiconductors, monitoring industrial emissions, or doing academic research\u2014you don\u2019t want a cheap, off-the-shelf one. You want one that\u2019s made with care, tested properly, and has that precise groove alignment. We\u2019ve been making these for years, and we\u2019ve worked with everyone from small university labs to NASA projects. We know how important your data is, and we take the manufacturing process super seriously\u2014because we know that a good Echelle isn\u2019t just a part, it\u2019s a tool that helps you make big discoveries.<\/p>\n

Plane Ruled Grating<\/a> If you\u2019re looking to order an Echelle grating, or just have questions about specs, groove densities, or materials, feel free to reach out. We can talk about your exact application, custom sizes, wavelength ranges\u2014whatever you need. We don\u2019t do one-size-fits-all, because every project is different.<\/p>\n

References<\/h2>\n