As an important optical element, microlens array has the characteristics of small size, light weight and high integration, which has attracted a lot of attention. With the development of the semiconductor industry and the improvement of photolithography and micro-processing technology, a series of new microlens array manufacturing technologies have emerged since the 1980s. Since lens array devices are divided into refractive microlens arrays and diffractive microlens arrays, different methods have been developed in their manufacturing processes.

3.1 Manufacturing method of refractive microlenses

Since refractive microlens array devices are widely used in light gathering, collimation, large area array display, optical efficiency enhancement, optical computing, optical interconnection and Micro-scanning and other aspects are becoming more and more widely used, and its manufacturing processes and methods have been increasingly studied. So far, there have been many methods for preparing refractive microlens arrays, including photoresist thermal reflow method, laser direct writing method, microjet printing method, sol-gel method, reactive ion etching method, grayscale mask method, Hot pressing molding method, photosensitive glass thermoforming method, etc. The following mainly introduces several mainstream microlens array manufacturing methods.

(1) Photoresist thermal reflow technology

The photoresist thermal reflow method (melted photoresist method) was proposed by Poporie in 1988. The entire process can be divided into Three steps, see the picture below: 1. Expose the photoresist on the substrate under the cover of the mask. The exposure pattern is circular, rectangular or regular hexagonal; 2. Develop and clean the exposed photoresist. Residual material; 3. Place it on the heating platform and heat-melt it. Since this method has the advantages of simple process, low requirements on materials and equipment, stable and easy-to-control process parameters, and easy replication, it is widely used in the production of microlens arrays.

However, the microlens array produced using this technology also has many shortcomings: 1. Since the photoresist wets the substrate material, the adhesion between the photoresist and the substrate is certain when it is in a molten state. , then when the molten photoresist is finally formed, there is a wetting angle between the spherical profile of the microlens and the substrate, so that the edge of the microlens has a certain curvature and the middle part sinks; 2. Generally, the filling factor of the microlens array is not It will exceed 80%, and the photoresist is easy to adhere after melting. Once the adjacent molten photoresist comes into contact, it will not form the surface shape of the lens. Since the filling factor is not high, the incident light cannot be fully utilized and background noise will be caused; 3. Since the photoresist itself has poor mechanical and chemical properties and low optical properties, it is not suitable for use as the final microlens or lens. Other microstructured materials.

(2) Laser direct writing technology

At present, because the laser direct writing method is easy to operate and has the advantages of small size and high precision of the micro-optical components produced, it is used in micro-fine It is widely used in research and processing fields. Laser direct writing technology uses a laser beam with variable intensity to expose the photoresist coated on the surface of the substrate with varying doses, and then forms the required relief profile on the surface of the photoresist after development. The biggest advantage of laser direct writing is that binary optical devices with multiple phase orders or continuous phases can be written at one time after the device is positioned, thus avoiding the loss of absolute axis accuracy caused by multiple mask overlays. The process of making microlens arrays by laser direct writing can be divided into three steps:

Use CAD to design the exposure structure of the microlens array and transfer it to the system of the laser direct writing equipment; The resist substrate is placed on the direct writing platform, and the photoresist is laser-written; the exposed photoresist is developed and the residual material is cleaned, and finally a neatly arranged and uniform microlens array structure is obtained. Laser direct writing method is suitable for high-precision single parts and model production. After using laser direct writing to produce the prototype of the microlens array, electroforming technology in the molding process is used to convert the microlenses into metal models for mass production. Because the electroforming replication process ensures the shape of the final product, it enables large-scale production of microlens arrays. Using these advanced technologies, micro-unit structures are repeatedly produced to produce high-quality and low-cost microlens array components.

3.2 Method of making diffractive microlenses

Diffractive microlenses have the functions of condensing light energy, correcting aberrations and imaging, and are small in size, light in weight, highly integrated, and easy to copy It is widely used in infrared photodetectors, image recognition and processing, optical communications, laser medicine, space optics and many other fields. Its main production methods include binary optical technology, electron beam direct writing technology, and grayscale mask technology.

(1) Binary optical technology

In the mid-1980s, a research group led by Veldkamp of MIT Lincoln Laboratory in the United States was designing a new sensing system. He was the first to propose the concept of "binary optics". It is different from the traditional production method and uses the production method of integrated circuits. The mask used is binary, and the mask is layered in the form of binary coding. . Subsequently, binary optics rapidly developed into -I'-J technology, which was favored by academia and industry. Binary optics technology is very suitable for the production of diffractive microlens arrays, in which the boundaries of microlenses can be easily made neat and sharp. , the fill factor of the microlens array can reach 100%, and it has the advantages of light weight, low cost, easy miniaturization and arraying.

Binary optics adopts phase quantized binary coding and production sequence. The phase series formed in N process steps is increased from N+I to 2N, as shown in Figure 1.2, which greatly reduces the number of iterations of process steps and reduces the cost. Processing costs required to manufacture diffractive optical elements with high diffraction efficiency. The manufacturing process of binary optical step diffraction microlenses is based on mature microelectronics technology and is suitable for mass production.

When the number of phase steps increases, binary optical elements can also have high diffraction efficiency like continuous relief elements. When the number of phase steps are 2, 4, 8, and 16 respectively, the theoretical diffraction efficiencies are 41%, 81%, 95%, and 99% respectively. As the number of steps increases, the diffraction efficiency increases, the production difficulty also increases, and the alignment accuracy requirements are higher. In order to ensure high diffraction efficiency and manufacturing accuracy, multiple photolithography and etching processes are required to produce multi-phase step diffraction microlenses. In the photolithography process, there is such a relationship between the phase level number L of the binary optical element and the required number of masks N: L=2IV. Therefore, three and four masks are required to make 8-phase step and 16-phase step microlenses respectively. In actual production, three masks are generally used, and an eight-phase (or eight-step) diffractive microlens array is manufactured through three photolithography and three etching techniques, which can basically meet the requirements. The manufacturing process of the microlens array mainly includes the design and production of the mask, using photolithography technology to transfer the designed mask pattern to the photoresist, and using dry etching or wet etching technology to transfer the photoresist pattern. Transferred to the substrate surface with high fidelity, forming the desired relief structure.

(2) Electron beam direct writing technology

In order to avoid the error accumulation problem caused by multiple overlays, people have developed a variety of one-time forming processing technologies, such as diamond turning. , laser direct writing method, chemical deposition method, etc. The direct writing method is a more practical method and is divided into three types: electron beam direct writing, ion beam direct writing and laser beam direct writing. The use of electronic direct writing technology to produce micro-optical devices began in the early 1980s. The principle of electron beam direct writing is different from that of laser beam direct writing. Before direct writing, a conductive film (such as Au, In, O) must be pre-plated on the substrate. , etc.) to leak electrons during exposure. The resolution of electron beam direct writing is very high. The Department of Electrical Engineering of the University of California, Los Angeles, USA used electron beam direct writing technology to produce a 45um diameter microlens with a critical size of only 60nm. Electron beam direct writing is an important method for producing subwavelength structured microlenses.

(3) Grayscale mask technology

Grayscale mask technology uses a grayscale mask to achieve multi-step diffractive optical elements or relief patterns with continuous phase changes through one photolithography. , and then through etching (or thin film deposition), the pattern is transferred to the substrate with high fidelity, as shown in the figure below. This technology simplifies complex multiple photolithography and pattern transfers into one completion, without problems such as alignment errors in overlaying. It is suitable for mass production, shortens the production cycle and reduces costs. The key to grayscale mask technology is the production of grayscale masks. Two of the more commonly used methods are color-coded reticles and high-energy electron beam sensitive glass reticles. The former uses different colors to represent different gray levels. One color represents a gray level. The four-phase surface relief distribution is represented by four colors. The eight-phase relief surface distribution is represented by eight colors. Then it is represented by colors. The grayscale graphics are printed on the transparent film with a high-resolution color printer, and then the color film is transferred to the black and white transparent film through shrinkage. This forms a mask with different grayscale levels, which can be obtained through one exposure. Relief surface distribution structure of multi-phase steps. The resolution of this mask is low, and the phase profile beam of the device is directly limited by the color level of the printer. High energy electron beam sensitive glass mask (HEBS) utilizes its different sensitivity to electron beams of different energies to form a true grayscale mask with a step change or continuous change in transmittance. This kind of mask has high resolution, up to 500 gray levels, and the mask production process is simple and low-cost. Components made using HEBS support grade masks have advantages such as high resolution and diffraction efficiency that cannot be matched by other methods. As the support level increases. The relief distribution is approximately continuous, but the production of the support mask will become very difficult as its gray scale increases, and the production cost will also increase significantly.

The various microlens array manufacturing methods mentioned above are more suitable for manufacturing small batches of microlens arrays. However, if mass production of microlens arrays is required, the above method is inconvenient and costly, the overall production process is complex, and product uniformity is difficult to guarantee. Therefore, developing replication technology has become the key to reducing the cost of micro-optical devices and promoting J1 applications. Generally, making microstructures on the surface of photoresist has the following disadvantages:

1. The surface of the photoresist material is relatively rough, which can easily cause diffuse radiation and reduce the optical performance of the device;

2. , The photoresist material surface has low mechanical strength, is susceptible to wear and is not suitable for harsh environments.