Short-Wave Infrared Reaches 5.32 Megapixels – This Band Is Moving From “Too Expensive to Use” to “Affordable to Use”

Hangzhou City, Zhejiang Province, China – November 11, 2026 – Short-Wave Infrared Reaches 5.32 Megapixels—This Band Is Moving From “Too Expensive to Use” to “Affordable to Use”

Over the past month, three things happened in the short-wave infrared band. They are independent of each other, but they point in the same direction.

One is that device specifications moved upward. A short-wave infrared image sensor for industrial equipment pushed pixel count to 5.32 megapixels, pixel pitch down to 3.45 micrometers, and spectral response from visible light all the way to 1.7 micrometers. With a copper-copper interconnect stacking structure, the overall device size can be made more compact.

Another is a change of technical route at the mechanism level. A domestic team developed a neuromorphic short-wave infrared event and radiation dual-mode imaging chip, putting event-driven acquisition and conventional grayscale imaging on the same device. The event channel outputs only sparse data on radiation changes, while the radiation channel acquires full grayscale on demand, with both paths synchronized from the same source.

The third happened at the material end. Short-wave infrared modules based on the colloidal quantum dot route have entered the supply chains for photovoltaic panel inspection, machine vision, and semiconductor inspection, with million-pixel products advancing toward volume production readiness.

Cost is the core selling point of this route. Compared with traditional indium gallium arsenide solutions, quantum dots offer a reduction of more than an order of magnitude; chip-level targets are even more aggressive, with some routes aiming for cost levels around one percent of InGaAs. The center of gravity for complete system prices is shifting from the previous hundreds of thousands of yuan toward the thousands. Here it is important to distinguish the basis of comparison—the reduction figures for chip, module, and complete system are not the same number.

Viewed together, these three developments show that short-wave infrared is moving from a laboratory band that “performs well but is unaffordable” to a product category that is affordable and can be deployed in volume. The impact of this shift on upstream optical components is deeper than it appears on the surface.

1. Why short-wave infrared has always been expensive

To understand this round of change, one must first understand where the high cost of short-wave infrared comes from. In the context of receiving ends such as InGaAs, short-wave infrared usually refers to 0.9 to 1.7 micrometers, a range that sits right outside silicon’s response boundary. Broader definitions also include 0.7 to 2.5 micrometers; this article uses the former.

Silicon’s bandgap is about 1.12 electron volts, which corresponds to a cutoff wavelength of 1.1 micrometers in terms of photon energy. When photon wavelength is longer than this value, the energy is insufficient to excite electrons from the valence band to the conduction band, and silicon produces no photocurrent. That is why CMOS image sensors used in phones and security cameras top out around 1 micrometer—beyond that, silicon cannot see.

The material that can cover this band is mainly indium gallium arsenide. By adjusting the ratio of indium to gallium, the bandgap can be tuned within a certain range, and the composition lattice-matched to an indium phosphide substrate responds exactly to 1.7 micrometers. The problem is that the indium phosphide substrate itself is expensive, epitaxial growth requires strict lattice matching, chip size is limited by the substrate, and both yield and capacity struggle to scale.

This cost structure propagates to the complete system: a short-wave infrared camera used to sell for hundreds of thousands of yuan. Price determined that its applications could only concentrate in a few cost-insensitive scenarios, such as scientific research, high-end industrial inspection, and certain specialized fields. Small installation volumes in turn failed to amortize R&D and production line investment, creating a cycle.

2. Which link does the quantum dot route change?

Colloidal quantum dots take a different path. They do not rely on epitaxial growth but instead synthesize nanoscale semiconductor particles by chemical methods, tuning the bandgap by controlling particle size. This is the quantum confinement effect: the smaller the particle, the larger the energy level spacing, and the more the absorption edge shifts toward shorter wavelengths.

But size is not the only variable. The actual response of short-wave infrared quantum dots is also significantly affected by surface ligands, shell structure, and interface defects, which change the effective bandgap and carrier recombination pathways. Size and surface chemistry must be coordinated to stably push the response edge to around 1.7 micrometers.

The key advantage of this route is its compatibility with CMOS processes. Quantum dots can be spin-coated or spray-coated onto already-fabricated silicon readout circuits, without the need for specialized substrates and epitaxial equipment like indium phosphide, and without flip-chip bonding alignment steps. Production lines can leverage existing semiconductor capacity, dramatically reducing fixed asset investment.

Another advantage is spectral tunability. With the same process, switching to a batch of quantum dots of different sizes yields devices with different response bands. Extending further, quantum dots of different sizes can even be arranged in zones to create a structure that simultaneously acquires multiple bands on a single chip. This is difficult to achieve in traditional epitaxial routes.

Of course, the quantum dot route has its own thresholds. Material stability, batch consistency, and interface defects during integration with silicon-based circuits are all hurdles to clear between lab and volume production. The current state is that it has already entered the supply chain for some industrial scenarios, but high-resolution, high-frame-rate cutting-edge products are still dominated by traditional routes, and the two will coexist for a long time.

3. Once costs come down, the questions optical components must answer change

The situation for optical components here is somewhat delicate. Lower complete-system costs mean higher installation volumes, and higher installation volumes mean batch consistency requirements shift from “being able to pick out the good ones” to “every piece must be good.” This shift is harder for filter manufacturers to meet than a simple performance specification increase.

Another change is the rising complexity of band combinations. In the past, short-wave infrared equipment was single-band, and optical components only needed to target one band. Now multi-spectral, multi-band simultaneous acquisition is becoming possible, and the same device may need three or four filters with different center wavelengths. This requires them to be made on the same substrate size and within the same coating process, with their relative positions and transmittance relationships controlled.

There is also size. As pixels shrink, the image plane for the same pixel count can be made smaller, and with stacked structures, the overall optical cavity has room for compression.

Once optical component sizes come down, the truly difficult part is not coating small pieces themselves. Coating uniformity mainly depends on the matching between substrate size and chamber geometry; small substrates are not necessarily harder than large ones. What is difficult is maintaining the relative relationships of multi-band coating stacks at small sizes: when several filters are installed in the same device, the center wavelength spacing between them must not drift due to process fluctuations.

4. What is special about short-wave infrared filters?

Filters for the short-wave infrared band are not the same as long-wave infrared; both the methods and the difficulties differ.

The material system is the first layer of difference, and the substrates usable for short-wave infrared hardly overlap with those for long-wave infrared. Germanium’s transmission window starts at 2 micrometers and is opaque across the entire 0.9 to 1.7 micrometer range, so it cannot be used here. Silicon has strong absorption below 1.2 micrometers and only becomes transparent above that, so it can only cover the latter half. The real workhorses for short-wave infrared are fused silica, sapphire, and fluorides, with zinc selenide also usable. This material change is not simply taking a long-wave design and changing some numbers—the coating stack must be recalculated.

Atmospheric absorption bands are the second layer to address. Short-wave infrared has a strong water vapor absorption band centered near 1.38 micrometers. When making bandpass filters, if the passband covers this position, the signal from outdoor long-distance passive imaging will be noticeably attenuated, with performance dropping quickly.

When using water vapor for identification, the common approach is not to press a single channel directly onto the absorption band for imaging, but to use a differential ratio between an absorption channel and an adjacent window channel, such as 1.38 micrometers versus 1.24 or 1.6 micrometers. Active-source absorption sensing is a different matter—in that case, measuring on the absorption band is valid. The two cannot be conflated.

Ambient light suppression is the third layer. The receiving surface of short-wave infrared can simultaneously see visible light and near-infrared. Outdoor sunlight has abundant energy in this band, and industrial site lighting, hot workpieces, and display screens also contribute. If out-of-band suppression is insufficient, ambient light directly raises the baseline and drowns the target signal. Filters for such systems typically require out-of-band suppression to OD4 or higher, with the blocking range covering the entire visible region.

One more point easily overlooked: when the substrate changes, the coating material system changes too. Silicon is a transparent window above 1.2 micrometers but strongly absorbs from 0.9 to 1.1 micrometers, so windows for the shorter wavelengths can only use fused silica or sapphire. The coating materials themselves overlap—some oxides and fluorides work in both bands—but the refractive index combinations and coating stacks must be redesigned and validated for the short-wave band. Long-wave solutions cannot be directly transplanted.

5. The fastest demand growth is not in consumer electronics

The reason short-wave infrared is being pushed forward at this particular moment also includes demand-side drivers. From public market analysis, defect inspection in semiconductor and photovoltaic fields is the fastest-growing segment, with annual growth on the order of thirty percent. The reason is not hard to understand: hidden cracks inside silicon wafers, metal impurities, and micro-cracks in solar cells cannot be seen under visible light, but under short-wave infrared they show high transmission and clear contrast.

This scenario has special requirements for filters. It does not need extremely narrow bandwidth but does need excellent uniformity. What is being inspected is the imaging quality across an entire wafer. Non-uniform filter transmittance introduces illumination gradients on the imaging plane, increasing residual error after flat-field correction and the risk of misjudgment in edge regions.

Such systems usually have flat-field correction that can calibrate out part of the non-uniformity. So the problem is not “certain misjudgment” but how much margin remains after correction. This is why such applications list coating thickness uniformity as a core metric.

Another segment is machine vision and food sorting. Different materials show more obvious reflectance differences in the short-wave infrared band than in visible light, and moisture, sugar content, and internal damage can be identified through multi-band ratios. Such equipment is cost-sensitive; once complete-system prices come down, installation volumes rise quickly, and demand for optical components is about volume rather than precision.

These two types of demand point in different directions: one requires extreme uniformity, the other requires controllable cost. For upstream manufacturers, this means a single process cannot serve both.

MULTI IR’s near-infrared narrowband products cover the 850 nanometer to 1653 nanometer range, which includes several common short-wave infrared bands. Center wavelength tolerance is controlled to plus or minus 2 nanometers, and out-of-band suppression reaches OD4 or higher.

Substrates in this band are mostly fused silica and sapphire, with refractive indices much lower than germanium and silicon. The design space for anti-reflection coatings is narrower, and the number of layers must be kept within a limited range. The selection logic for substrates and coating materials is a different set of considerations from long-wave infrared and cannot be copied directly.

MULTI IR’s experience concentrates on two things. One is coating thickness uniformity on large substrates: applications like wafer inspection demand high full-piece consistency, which relies not on one batch being good but on long-term stability of distribution within the coating chamber. The other is the relative relationships of multi-band combinations: when multiple filters are placed in the same device, the center wavelength spacing and transmittance balance between them must be designed in advance.

Technology route transitions are never completed overnight. InGaAs will not disappear, and quantum dots will not immediately take over the entire market. The two routes will coexist in their respective scenarios for many years. But once the price center shifts downward, the entire industry’s attention turns from “can it be made” to “can it be made consistently.” For those in optical components, this shift is the part that truly requires advance preparation.

What short-wave infrared will ultimately become is too early to conclude. What is certain is that when a device goes from hundreds of thousands of yuan to a few thousand, everything around it will be repriced, optical components included. Whoever can maintain consistency when prices come down stays on this chain.

MULTI IR – National-level “Little Giant” Specialized and Sophisticated Enterprise. Over 10,000 types of infrared sensitive components in stock. Top 3 globally in comprehensive strength. Product portfolio covers infrared filters, optical coatings, infrared sensors, and other core categories. Lead drafter of the infrared filter industry standard.

Website: www.mirhz.com | Global Site: www.miroptech.com

About Us

Founded in 2007, Hangzhou MULTI IR Technology Co., Ltd. is an optoelectronic technology enterprise integrating R&D, production, and sales. Its products are widely applied in aerospace, medical care, AR/VR, display imaging, photography, and other fields, steadily holding the position of the world’s largest spot supplier of optical components.

Media Contact
Company Name: HANGZHOU MULTI IR TECHNOLOGY CO., LTD.
Contact Person: Media Relations
Email: Send Email
Country: China
Website: https://www.miroptech.com/

 

Press Release Distributed by ABNewswire.com

To view the original version on ABNewswire visit: Short-Wave Infrared Reaches 5.32 Megapixels – This Band Is Moving From “Too Expensive to Use” to “Affordable to Use”