The links between optical fibers are divided into permanent links and active links. All connections come with pins, locating grooves or precision ferrules.

The connection process is to align the fiber axis, and then fix it, and some direct fusion. In the connection, connection errors caused by mechanical processing, technology, etc. and differences in optical fiber structure parameters will cause connection loss.

Between the connectors, the inherent loss caused by the mismatch of the structural parameters of the optical fiber is the internal loss, and the light is emitted from the fiber of radius a₁ to the a₂ fiber, a₂<a₁, the loss IL_a can indicate that the loss such as the A gap is the external loss, and the core diameter is 2a , the dislocation d, the end-face dislocation loss is expressed as

The loss caused by the interplane gap Z of the air refractive index n0 can be expressed as

Fiber coupling is to transmit the light emitted by the light source to the fiber to the maximum extent, which involves the spatial distribution of the light source radiation, the light emitting area of the light source, and the light receiving and transmission characteristics of the fiber.

Fiber coupling methods include direct coupling and indirect coupling.
Direct coupling is to place one or more optical fibers on the receiving end face directly close to the emitting surface of the light source.
Indirect coupling refers to the coupling method using optical elements such as lenses between the light source and the optical fiber.

The essence of the coupling of the optical fiber device is the matching of the laser output light field and the optical fiber receiving mode field, and the coupling efficiency is the expression of the matching degree of the two mode fields.

From the theoretical point of view of the coupling mode, the coupling between the semiconductor laser and the optical fiber is actually the mode matching of the two. The output light field of a semiconductor laser is an elliptically symmetric Gaussian distribution. If a flat-end fiber is used for direct coupling, the coupling efficiency is very low, about ten percent, due to the mode mismatch between the two.

Therefore, it is necessary to perform optical shaping of the outgoing optical field to achieve the matching of the optical mode field and the optical fiber, and reduce the loss caused by the coupling between the two. There are currently three types of pre-coupling shaping techniques: single-lens coupling, lens group coupling, and fiber-optic microlens coupling.

Single-lens coupling is to add a lens between the transmitting laser and the receiving end face of the fiber, as shown in Figure 1.

Fig. 1 Schematic diagram of single lens collimation

A lens is used to shape the mode field of the beam emitted from the semiconductor laser to match the mode field of the fiber.

The advantages of single lenses are their large working distance, easy adjustment, simple assembly and loss insensitivity. At present, self-focusing lenses, ball lenses and cylindrical lenses are commonly used in the semiconductor laser industry.

The lens group is two or more lenses for realizing the matching of the mode field of the output beam of the optical fiber and the semiconductor laser. Doing so can avoid the problem that the single lens is difficult to calibrate with low coupling tolerance due to the limitation of optical characteristics, and can improve the coupling efficiency and tolerance.

The advantages of optical fiber microlens shaping are low packaging cost and simple optical path. It is essentially that the optical fiber has the required spherical, hemispherical or cone angle microlenses on the end face of the optical fiber through optical etching, chemical etching or mechanical polishing, which greatly improves the coupling efficiency. However, this process is very complicated, the cost is high, and the scope of use is small.

From the above three methods, it can be seen that choosing between a single lens and a lens group is the most common shaping method that can easily achieve the purpose. Lens and D-shaped lens (cylindrical lens) collimation is the most economical and suitable choice.

It has been briefly described above that for semiconductor lasers, the full angle of divergence in the slow axis of LD is small (about 7°), while the full angle of divergence in the direction of the fast axis of LD is large (about 23°), which is larger than the numerical aperture of the coupling fiber itself, so in the collimating the fast axis direction requires the use of a cylindrical lens with a small R (curvature) range (as shown in the figure).

According to the analysis of the characteristics of the semiconductor laser, from the design aspect, the R and N of the collimating lens are optimized, then for the final result, the beam quality and spot requirements can meet the fiber coupling process.

Here we use ZEMAX to simulate the use of microlenses to collimate the fast-axis beam and slow-axis beam emitted by the laser, respectively. The width of the collimated beam can be basically controlled at about 5μ, and the center spot is basically 1μ.

According to the introduction, the single luminous point collimation experiment uses cylindrical lens collimation, mainly for the following reasons:
(1) In order to meet the coupling efficiency, only the fast axis can be compressed before the cylindrical lens is used for coupling.
(2) The commonly used cylindrical lens material is quartz glass, which can save costs.

Due to the large divergence angle of the semiconductor laser in the direction of the fast axis, if it is not collimated and shaped, then the light spot at 100mm in the far field has a large width in the Y-axis direction, the beam is divergent, and the energy density is extremely low. If a cylindrical lens is used to collimate the fast axis, the beam spot width on the Y-axis of the collimated beam is much smaller than that before the collimation, and the energy density of the beam spot is greatly improved, as shown in Figure 2.

Fig. 2 Simulation diagrams before and after calibration (left before calibration, right after calibration)

Since various other errors may occur during the experimental operation, the compressed light spot will be inconsistent with the theory, and the problem of the light spot caused by the collimation error is analyzed.

(1) On-axis error of cylindrical lens
In the process of the collimation experiment, the cylindrical lens becomes the collimating mirror, and the adjustment of its position can obtain the most suitable position in the axial direction. In the experiment, we give the maximum offset of the lens as 10μ and 10μ, respectively. Then, when the cylindrical lens is far away from or close to the light-emitting surface, the spot width in the Y-axis direction will increase, and the energy density will decrease accordingly, resulting in low coupling efficiency, as shown in Figure 3.

Fig. 3 Spot characteristics of axial collimation error

(2) The error between the vertical and the optical axis of the cylindrical lens
Cylindrical lenses will be displaced in the direction of the optical axis and will also be offset in the direction perpendicular to the optical axis.
When the cylindrical lens is offset in the direction perpendicular to the optical axis, the shape of the test far-field spot will change because it deviates from the predetermined position. The moving direction of the cylindrical lens affects the moving direction of the spot, and their movement are synchronous and consistent.
According to the difference of the up and down movement of the lens, the light spot shows different concave and convex shapes, as shown in Figure 4.
It is relatively easy to adjust the error on the vertical axis of the cylindrical lens in the collimation process, that is, it is necessary to move back and forth through the light screen to observe the shape change of the light spot until the position of the light spot no longer changes to determine the position of the light screen.

Fig. 4 Vertical axis simulation diagram (the left is the light path diagram, the right is the light spot diagram)

(3) When the cylindrical collimating lens and the P-N junction cannot be parallel to each other, there will be an included angle between the two light spots. The larger the angle deviation of the cylindrical lens, the larger the angle of the light spot.

Fig. 5 Simulation diagram of lens and P-N junction not parallel

In addition to the non-parallel P-N junction and the non-parallel with the light-emitting surface, the non-parallel to the light-emitting surface will cause uneven light intensity in the far field, and there will be phantom light parts. The light screen with scale lines can solve these two problems. In the case of the angle shift during the experimental operation, visually check whether the scale line of the light screen is parallel to the shape of the light spot. When the light spot is parallel to the scale line, it means that the angle of the cylindrical collimating lens is in the most suitable position.

Of course, the light spot in actual operation will not be the single case of the above-mentioned various errors, and each error may appear individually or at the same time, which requires us to correct and adjust the fixture one by one.

When the laser beam is coupled with the fiber array, if a higher coupling efficiency is required, the outgoing laser beam needs to be fully coupled and matched with the receiving end face of the fiber, which means that the diameter of the laser beam is required to be no larger than the core of the fiber in terms of physical size. The diameter of the laser beam should be matched with the numerical aperture of the fiber, so that the total reflection condition of the laser beam transmitted in the fiber can be satisfied.

d_laser laser spot diameter and d_core are the core diameter, g_laser is the laser divergence full angle

The laser beam parameter product (BBP) is defined as:

According to the Helmholtz invariant, if there is no aberration and diaphragm, the beam parameter product of a laser is a constant value, that is, the beam parameter product will not change due to changes in the optical system.

We can use the formula (IL_a – IL_b) to select the numerical aperture of the selected fiber and the size of the core respectively, to ensure the efficient coupling of the system and ensure the beam quality. In addition, the thermal effect caused by the efficient operation of the laser during the pumping process will be a comprehensive consideration when selecting the optical fiber to change the beam quality to the coupling conditions.

We can use the structure of the cylindrical lens to collimate the divergence angle of the laser beam in the radial direction. We know that the full divergence angle of the light emitted from the semiconductor laser array parallel to the junction direction is about 10°, and the divergence full angle perpendicular to the junction direction is about 40°. The collimating lens (usually replaced by a glass rod) is placed in front of the LD along the direction parallel to the junction with the help of an optical adjustment frame, as shown in Figure 6.

Fig. 6 Schematic diagram of the coupling optical path of the optical fiber cylindrical lens

The effective radius of the cylindrical lens is R, and the distance from the cylindrical lens to the laser light-emitting surface is Z.

The laser beam in the radial direction of the cylindrical collimating lens is collimated after repeated refraction by the effective receiving surface in the diameter direction of the cylindrical lens, which satisfies the numerical aperture angle fiber and effectively improves the coupling efficiency.

The figure below uses the adjustment of the cylindrical lens radius R and the distance z from the laser light-emitting surface to the cylindrical lens to make the effective numerical aperture of the fiber suitable for system coupling.

Fig. 7 Equivalent relationship diagram of cylindrical lens coupling and receiving

When Z/R=0.16, all laser beams with divergence angle less than 42° can be coupled into the fiber array.

While studying the coupling method, we must know the coupling characteristics of all the fibers in the fiber array, which helps us to improve the study of the array coupling method. We can analyze the coupling efficiency of the LD to the fiber array by tracing rays, and the numerical aperture (NA) of the fiber is related to the number of propagating modes. When studying the butt coupling between multimode fiber and LD, according to the knowledge of multimode fiber and transmission theory, when the divergence angle of the incident laser beam is smaller than the receiving aperture angle of the coupling end fiber, the coupled beam is transmitted stably and the loss is very small. The angle of the incident beam entering the coupling end face can be determined by the light field characteristics of the laser output beam.

The intensity distribution of the LD beam in the output light field can be expressed as the following function:

among

ω_0x, ω_0y are the half-height widths of the Gaussian distribution of the beam, and A(Z) is a constant related to Z.

The far-field divergence angle of the Gaussian distribution of the laser beam at the point where the intensity is l/e2 is:

The radius of curvature of the beam in the light field can be transformed into:

It can be known from the characteristics of Gaussian beams that in the transmission of light waves, the Gaussian state distribution of the vibration amplitude and intensity of the center of curvature of the beam in the longitudinal cross-section is changed, and its equipotential phase surface is always spherical.

Knowing the structural properties of the multimode fiber, when the numerical aperture is given, when the maximum incident angle light is less than the maximum numerical aperture angle θ_c, the light beam can be coupled into the fiber, at this time θ_c=sin^-1 (NA/n₀) . Take n₀=1, then θ_c=sin^-1NA, when the condition of θ<θ_c is satisfied, all light can be coupled into the fiber. The optical power in the four quadrants of the optical field is the same, and the total power of the laser beam field on the z=s plane is:

Usually, the constant coefficients are omitted, so that

then

According to Fresnel’s law of reflection, the transmittance T(z, y) varies with the incident angle of the light. The optical power within the effective aperture angle of the fiber is:

Similarly p0=px0/Py0, the coupling efficiency is:

Here, the coupling of the semiconductor laser array to the optical fiber array mainly refers to the coupling of the emitting light source of the laser array and the end face of the optical fiber array one by one. At the same time, the reasons for the loss of key accuracy errors and coupling efficiency in the alignment process are analyzed.

The key to such a coupling method lies in how to control and ensure that the laser output light is parallel to the axial direction of each end face of the fiber array and the optical axis coincides. It is commonly used to photoetch V grooves on Si wafers with a tolerance accuracy of 0.5μ. The spacing error of the V grooves is matched with the physical size tolerance of the interval between the laser light-emitting areas. Cover a flat or concentric V-shaped silicon wafer.

Optical devices such as beam collimation and coupling can be placed between the laser array and the fiber array, as shown in Figure 8.

Fig. 8 Flow chart of fiber array coupling

Taking the semiconductor laser array as an example, in addition to using a cylindrical lens (usually a fiber with the cladding removed), the beam perpendicular to the junction direction of the semiconductor laser array is collimated and compressed at the fast axis divergence angle, and then directly coupled. Of course, there are also people. The method of designing and simulating micro-lenses cuts and rearranges the output light of the semiconductor laser array to achieve the purpose, which is difficult to achieve in engineering.

The cylindrical lens is collimated with a fiber array, and the coupling requires the selected collimator and the array fiber core diameter to match, that is, the output of the beam is the same as the receiving aperture, and the laser, the collimator and the coupling fiber are required. The optical axial position is precise and stable in order to obtain the ideal coupling efficiency.

At present, in the coupling technology of laser and optical fiber, the machining accuracy of mechanical parts affects the mutual position error of the hardware of the optical coupling material itself, such as the accuracy of the device. At the same time, the cooperation between the fixture and the device will also be affected by the machining accuracy of the machine Produces such as angle error and so on.

Therefore, we can divide the combined effect of mechanical error on coupling into horizontal, vertical and angular errors according to the mechanism. These coupling errors are calculated and analyzed separately below.

The first one is also the most common loss in fiber coupling, that is, the loss caused by the lateral displacement error d. The so-called lateral displacement error refers to the fact that the axis of the fiber core and the central direction of the receiving spot are not completely coincident. A displacement d perpendicular to the optical axis is generated, resulting in an error, as shown in Figure 9

Fig. 9 Schematic diagram of lateral displacement error analysis

In the figure, the fiber core radius is set as R, and ω is the laser spot radius after focusing. When a beam of light is focused and its beam divergence angle meets the coupling conditions, if there is a lateral position shift d between the receiving end face and the outgoing light in the direction of the optical axis, some of the beams that are not fully compressed and focused will escape. out of the fiber end face, resulting in coupling loss.

When the entire beam meets the coupling conditions, the light energy on both sides of the coupling end face is proportional to the overlapping area of the fiber core end face and the exit light face, so the coupling efficiency η affected by the lateral displacement offset d can also be

In the offset efficiency tolerance formula,

We design a multimode fiber with a 150μ core, so R=0.075mm, ω=0.075mm, and the influence curve of d on the coupling efficiency is calculated as shown in the figure below. It can be seen from Table 1 that the influence of the error of d on the coupling efficiency is very obvious, so in the assembly process of the device, the control of d is an important factor in the control of the performance parameters of the device.

Table 1 Influence of lateral offset on coupling efficiency

Another kind of loss is the loss caused by the angle offset error θ. The angle offset error means that the optical axis of the incident beam and the optical axis of the receiving fiber are not in a straight line, and they have an included angle θ, which causes the coupling of the fiber to generate error. See Figure 10 for a diagram of the angular offset error:

Fig. 10 Schematic diagram of angular offset error

It can be seen from the angular offset error diagram that the angular offset error will cause the divergence angle of the outgoing beam and the numerical aperture of the receiving fiber to not meet the coupling requirements, that is, the two do not meet the coupling conditions (θlaser<2arc sin(NA )), so that the optical fiber cannot receive light energy beyond the maximum acceptance angle.

D.Marcuse has a conclusion that can be extended: In a multimode step fiber, the energy loss caused by the angular offset error θ is:

ω is the light-emitting radius of the focused spot, and n2 is the refractive index of the fiber cladding.

Table 2 Influence of offset angle on coupling efficiency

It can be seen from the calculated data that the angle between the optical axis of the receiving fiber and the optical axis of the outgoing beam has little effect on the coupling efficiency within the range of less than 0.05 radians (about 3°).

The third type of coupling loss is the coupling loss caused by the longitudinal position offset error s. The longitudinal position offset error means that the distance between the outgoing spot and the fiber receiving end face in the direction of the optical axis of the incident beam cannot be effectively coupled, resulting in Therefore, the spot diameter of the laser beam is larger than the end face diameter of the receiving fiber (d_laser>d_core), resulting in coupling loss.

When studying the coupling device, there are two methods for the precise distance between the focused spot and the fiber end face:
First, when the coupling efficiency of the coupling device is tested, the reading of the optical power meter can timely know whether the current distance is the best;
Second, the optimal longitudinal distance can be simulated using ZEMAX for different laser focus spots when the structure of the studied coupling device is fixed.

From the analysis of the simulated data, it can be concluded that the coupling efficiency has a significant effect only when the longitudinal distance exceeds 0.1 mm.

Through the analysis of these three kinds of mechanical alignment errors, the root cause of the error source is obtained from the data, and the importance of the error brought by the mechanical precision error to the coupling is understood.

Therefore, in the mechanical design of the coupling device, the influence of these three coupling alignment errors on the system should be considered, especially for the lateral displacement error that has a large influence on the coupling efficiency, a process guarantee is required, which is also one of the main considerations in this paper. question.

Of course, the end-face coupling of the large-core diameter multimode silica fiber is experimentally studied. The results show that the processing of the fiber end face and the reduction of the lateral alignment error are the key factors to reduce the splice loss of the fiber end face. The high grinding and polishing quality of the fiber end face can be The diffusion of light at the coupling interface is significantly reduced, and the angle between the fiber axes can also be reduced. The deviation between the theoretical and experimental values is mainly due to the fact that the light intensity caused by the actual fiber structure is a Gaussian distribution, and the theoretical analysis is based on a uniform distribution.

This chapter focuses on the characteristics of laser collimation coupling, the solution to the coupling efficiency problem caused by some mechanical errors in the process of fiber coupling laser, and the fiber coupling device involved in this paper is given to effectively avoid the mechanical error band in the coupling. To couple the conditions of the problem:
High-power semiconductor lasers use a single lens or a D-type lens (cylindrical lens) for beam collimation before the device is coupled;
The mechanical error analysis process shows that the lateral deviation is the mechanical error that has the greatest impact on the coupling efficiency, and it needs to be precisely controlled within 0.1mm in the coupling of optical fiber devices.