JPS6320392B2 - - Google Patents

Description

[Detailed description of the invention]

The present invention relates to a semiconductor laser coupler for a single mode optical fiber that efficiently couples light from a semiconductor laser to a single mode optical fiber. Coupling methods for efficiently coupling the light from a semiconductor laser (LD) to a single mode optical fiber (SMF) can be roughly divided into (a) methods using only a micro cylinder or spherical lens with a diameter of about 20 μm, and (b) methods using only a single mode optical fiber (SMF). There are two methods: (c) a method in which a micro cylinder or spherical lens is provided at the incident position of the mode optical fiber; and (c) a method in which two types of lenses with different focal lengths are combined into a confocal structure (referred to as a confocal multiple lens system). Figure 1 shows each bonding system. That is, FIG. 1A shows the method (a), in which a semiconductor laser 1 and
A microlens 4 is arranged between the single mode optical fiber 2 and the light from the semiconductor laser 1 is transferred to the optical fiber 2.
The light is made incident on the incident end face of the core 3. Figure 1B
shows the method of (b), in which the input end of the optical fiber 2 is gradually tapered, and a microlens 5 is integrally formed at the tip. FIG. 1C shows the method (c), in which a lens 6 is placed between the semiconductor laser 1 and the optical fiber 2.
and 7 are arranged. In methods (a) and (b), the permissible amount of axis deviation in the direction perpendicular to the optical axis of the microlens or lens-equipped optical fiber is extremely small, which is a major problem in actually manufacturing a coupler. For example, in order to suppress the deterioration of coupling efficiency to within 1 dB from the maximum value, it is necessary to realize axis alignment with an accuracy of within ±0.6 μm. On the other hand, (c)
This method was devised to reduce the accuracy required for the microlenses mentioned above, and the fixing accuracy of the lens used is relaxed to ±40 μm, and it is actually possible to connect a semiconductor laser and a single mode optical fiber. Fabrication of combined semiconductor laser modules has become extremely easy. However, the fixing accuracy of optical fibers still requires strict precision of ±2 to 3 μm, and unless this precision is relaxed, it will not be possible to achieve a price reduction equivalent to that of semiconductor laser couplers for multimode optical fibers. Ta. This invention is a confocal compound lens system in which the fixing precision of the single mode optical fiber is significantly relaxed, and the manufacturing efficiency of the semiconductor laser coupler is significantly improved. Explain. First, a conventional confocal compound lens system will be explained. FIG. 2 shows a diagram of the principle of a confocal compound lens system. Arrow 8 in the figure indicates the position of the light emitting end face of the semiconductor laser, and arrow 9 indicates the position of the input end face of the single mode optical fiber. The first lens 6 is located approximately at the focal length of the first lens 6 from the end face position 8 of the semiconductor laser.
The second lens 7 (focal length f ₂ ) is separated by f ₁ .
The distance between the lens 6 and the first lens 6 is the sum of their focal lengths (f ₁ + f ₂ ), and the position 9 of the optical fiber entrance plane is
is separated from the second lens 7 by f ₂ . When such a lens system is adopted, the emission diameter (spot size) 2W ₀ of the semiconductor laser is expanded by the ratio of the focal length of the lens system (f ₂ /f ₁ ), and the spot size is The diameter of 2W ₂ is
Form an image of 2W ₂ = 2W ₀ × (f ₂ / f ₁ ). Therefore, when the radius of the guided light beam of the optical fiber is W _f , f ₂ /f ₁
and W _f /W ₀ should be approximately equal. In reality, the laser beam at the junction surface of the semiconductor laser 1 has a beam radius W _v at the beam waist in the direction perpendicular to the junction surface, and a beam radius W _p at the beam waist in the parallel direction. W ₀ with √ _v・_p as target beam radius
is defined. Since this beam conversion depends only on the focal length ratio of the lens system, if the spot size of the semiconductor laser W ₀ =1 μm and the spot size of the optical fiber W _f =5 μm, then the lens combination of f ₂ /f ₁ =5 , and there are no restrictions on the values of f ₁ and f ₂ themselves. For example, as the second lens 7, a commercially available converging rod lens (a cylindrical glass rod whose refractive index distribution decreases in proportion to the square of the distance x from the central axis and which has a lens effect) can be used. n
If (r)=n ₀ (1-α ² /2r ² ), then the focal length f of a rod lens of length L is f(L)=1/(n ₀ αsinαL),
The pitch length L _p is defined as L _p =2π/α. )(diameter
1.8 mm, approximately 1/5 pitch length), f ₂ = 2.2 mm. Therefore, the optimum value for the first lens 6 is f ₁ =440 μm. The first lens 6 is usually sapphire.
A spherical lens made of ruby, YAG or glass, etc.
Use one with an outer diameter of 0.7 to 1.0 mm. Next, the axis deviation characteristics of this conventional confocal compound lens system will be explained. In manufacturing a coupler that integrates a semiconductor laser and a single mode optical fiber, it is necessary to clarify the fixing precision of each lens and the fixing precision of the optical fiber. Table 1 shows the fixing accuracy of each optical element to suppress deterioration within 1 dB from the optimum value.

[Table] This value is the accuracy required when manufacturing the _coupler . The idea is to make adjustments by moving in the y and z directions. As shown in Table 1, the first lens 6 and the second lens 7 may be fixed with relatively loose precision, and the optimum coupling can then be achieved by adjusting the optical fibers. For semiconductor lasers, the first
The lens 6 is aligned, then the second lens 7 is aligned, and the optical fiber is matched with the fiber by beam conversion by this lens system, so the fixing accuracy of the optical fiber at that time is equivalent to the connection characteristics of single mode optical fibers. . Therefore, no matter what lens system is used, as long as adjustments are made in the x, y, and z directions at the entrance surface of the optical fiber, a fixed accuracy of ±2.5 μm in the x and y directions is required as shown in Table 1. Ru. By the way, the connection characteristic between Gaussian beams is expressed as η(x, θ)=exp(−x ² /W ² −π ² W ² θ ² /λ ² ) (1). x is the axis deviation perpendicular to the optical axis, θ
is the angular shift between the two beams, W is the spot size at the beam waist of the Gaussian beam, and λ is the wavelength. However, it was assumed that the positions of the two beam waists were the same. According to this formula, if W is made larger, the allowable error of x becomes larger. However, it can be seen that if W is made too large, the error in the angular deviation θ becomes severe. Currently, the spot size of single mode optical fiber is approximately W = 5 μm, so 1 dB deterioration is expected.
The values of x ₁ and θ ₁ are x ₁ =±2.4 μm and θ ₁ =±2.3 degrees, respectively. Normally, the angular deviation can be suppressed to within ±0.5 degrees, so if the 5 μm spot size is increased several times, it is expected that the fixing precision of the single mode optical fiber will become extremely loose. Here, we will consider an actual compound lens system.
Assuming that the spot size of the semiconductor laser is W ₀ = 1 μm, that of the optical fiber is W _f = 5 μm, f ₂ /f ₁ = 5, and a focusing rod lens with f ₂ = 2.2 mm is used as the second lens 7, f ₁ =440μm. In the confocal system, W ₀ is converted to W ₁ by the f ₁ lens using the relational expression W ₁ =λf ₁ /πW ₀ . This value is approximately W ₁ = 182μ
m, resulting in a nearly parallel beam with a divergence angle of 0.13 degrees. As shown in Table 1, even if the distance between the first lens 6 and the second lens 7 deviates significantly from the confocal condition (f ₁ + f ₂ ), the efficiency changes little. This is because the beam system is converted into a parallel beam system. Here, in order to facilitate the adjustment of the single mode optical fiber, consider the case where the converging rod lens of the second lens 7 is integrated with the single mode optical fiber.
In this case, when W is 182 μm in equation (1), the axis shift and angle shift for a 1 dB loss increase are x ₁ =
±87 μm, θ ₁ =±3.75 minutes. In this example, the angle deviation is extremely severe, and even a deviation of ±0.2 degrees will increase the loss by 10 dB. Furthermore, if the distance between the first lens 6 and the semiconductor laser deviates from the optimum value f ₁ =440 μm by ΔZ, the position converged through the second lens 7 will be shifted by ΔZ.
(f ₂ /f ₁ ) ² = 25 times of deviation is calculated, and ΔZ is considered to be at least ±10 μm, so the position of the beam waist from the output end face of the focusing rod lens 7 (Fig. 1, 9) is calculated. It is necessary to take into account the variation of ±250 μm. Therefore, high coupling efficiency cannot be expected even if the converging rod lens 7, which is the second lens, is directly integrated into a single mode optical fiber. FIG. 3 shows an embodiment of the present invention, and FIG. 4 shows its principle. Parts corresponding to those in FIGS. 1 and 2 are given the same reference numerals. This is a case where a spherical lens is used as the first lens 6, and in the present invention, two converging rod lenses 71 and 72, for example, are used as the second lens. The second lens 7 is a combination of converging rod lenses 71 and 72, and the length is selected so that a desired focal length _f2 can be obtained when the two lenses are brought into close contact with each other (for example, 1/4 pitch length). In the illustrated example, the total pitch length of the lenses 71 and 72 is 1/4, the first focusing rod lens 71 has a pitch length of Φ, and the second focusing rod lens 72 has a pitch length of (1/4-Φ). Further, the second focusing rod lens 72 is integrated with the single mode optical fiber 2 using an adhesive or the like in advance, and the optical fiber is aligned with the lens 72.
It is assumed that the optical fiber 2 and the optical fiber 2 are fixed in the state of a fiber with a lens. The optimum length of the second focusing rod lens 72 will now be considered. Since the total length of the converging rod lens 7 is determined by the desired _f2 , the length of the second converging rod lens 72 is determined by determining the length of the first converging rod lens 71. At this time, if the beam waist between both converging rod lenses 71 and 72 is determined, the coupling characteristic can be determined from equation (1). So, what is the value (W ₂₁ ) of the beam waist W ₁ = 182 μm converted by the ball lens 6 using only the lens 71?
Fig. 5 shows the results of calculations. Fifth
In the figure, the horizontal axis is the pitch length of the first and second focusing rod lenses. From this figure, in order to expand W ₂₁ to several times (2 to 6) the beam waist diameter W _f = 5 μm of the optical fiber, the first converging rod lens 71 is required.
It can be seen that the pitch length of can be selected from 1/36 pitch to 1/12 pitch. Note that the position of the beam waist W ₂₁ defined here does not actually exist between both converging rod lenses 71 and 72, but exists within the optical fiber as conceptually shown in FIG. 4. In this case, the coupling characteristics of the present invention adjusted between the lenses 71 and 72 are such that the size and position of the beam waist (W ₂₁ ) of the real image formed by the beam of the beam waist W ₁ passing only through the lens 71 are as follows.
This can be calculated based on how well the size and position match the size and position of the beam waist (W _f2 ) of a virtual image created when a beam of spot size W _f corresponding to the fiber is incident on the lens 72 from the right in the opposite direction.
The spot size W ₂₂ formed after a beam with a spot size W ₁ passes through the focusing rod lenses 71 and 72 and the position of the beam waist match the fiber spot size W _f and the position of the fiber entrance end face. If so, the above real image
Since W ₂₁ and the virtual image W _f2 match in size and position, the axis deviation characteristics of the lensed fiber shown in Fig. 3 can be calculated by substituting the spot size W = W ₂₁ = W _f2 into equation (1). You can judge by that. Next, if the distance between the semiconductor laser and the ball lens 6 deviates from the design value f ₁ and the position of W ₂₂ deviates from the fiber entrance end face, calculate how much compensation can be achieved by changing the distance between the lenses 71 and 72. The results are shown in Figure 6. When the horizontal axis shows the pitch length of the first focusing rod lens 71, and the vertical axis shows the distance d between the two focusing rod lenses 71 and 72, when the distance d is changed by Δd,
The output end surface of the second focusing rod lens 72 (the third
The amount of movement ΔD of the beam waist W ₂₂ from FIG. 82) is shown. Here, the amount of movement outward from the lens end face 82, ie, into the interior of the fiber, is precisely determined. It can be seen from FIG. 6 that as the distance d is increased, the beam waist moves inward from the output end face 82 of the second focusing rod. If the two converging rod lenses 71 and 72 are designed to focus an image on the exit end surface 82 of the second focusing rod lens with the distance d, then the beam waist W ₂₂ can be changed relative to the exit end surface 82 by increasing or decreasing the distance d. You can move it from side to side. Here, if the installation error of the ball lens 6 in the optical axis direction is f ₁ ±20 μm, the beam waist will vary in the optical axis direction by ±500 μm from the design value. Therefore, if the value of (ΔD/Δd) is too small, the second lensed fiber must be moved back and forth significantly. For example, if the first focusing rod lens length is set to 1/36 pitch length, ΔD/Δd is 0.03.
Therefore, in order to cover the variation (±0.5 mm) of the beam waist in the optical axis direction, the movable distance of the second lens-attached fiber must be increased to ±17 mm. For this reason, as the semiconductor laser coupler becomes larger, dimensional fluctuations due to thermal expansion of the coupler member become more noticeable, making it difficult to suppress deterioration due to axis deviation in the direction perpendicular to the optical axis of the optical system. Therefore, the value of ΔD/Δd is preferably 0.1 or more. On the other hand, from the discussion on the size of the beam waist W ₂₁ converted by the first converging rod lens 71, the length of the first converging rod lens ranges from 1/36 pitch to 1/36 pitch.
12 pitch is good, that is, the ratio of the lengths of the first and second focusing rod lenses 71 and 72 should be 1:8 to 1:8.
We have found that a ratio of 3:6 is better, but if we consider the value of ΔD/Δd, we can see that a pitch of about 1/18 is better. Here, the first focusing rod lens 71 is set to 1/18 pitch, and the second focusing rod lens 72 is set to 7/18 pitch.
When the pitch is 36, the axis deviation and angular deviation characteristics in the direction perpendicular to the optical axis are determined from W ₂₁ = 14.5 μm, x ₁ =
±6.96 μm, θ ₁ =±0.78 degree (where x ₁ and θ ₁ are axis deviation and angular deviation amount that give a 1 dB increase in loss). On the other hand, the allowable axis deviation amount in the optical axis direction is relaxed to the reciprocal of the value of ΔD/Δd, that is, 8.3 times. The large permissible axis deviation in the optical axis direction can be understood from the fact that when the first focusing rod lens 71 is short, the position of the beam waist does not change much even if the distance d between the focusing rod lenses is changed. Considering the dispersion in the distance of the ball lens 6 in the optical axis direction, the distance in which the fiber can be moved back and forth should be ±4.2 mm. From the above explanation, the division ratio of the converging rod lens 7 (ratio of the lengths of the first and second converging rod lenses)
It can be seen that the optimum value exists when it is smaller than 1, but the absolute value depends on the degree of axis deviation and angular deviation in fixing the fiber, and the degree of variation in the distance between the ball lens 6 and the semiconductor laser. It depends on how much you estimate the variation in the aperture position of the beam waist. The values shown here are just an example; in reality, when the angular misalignment is 0.78 degrees, which is severe compared to the axial misalignment amount of 7 μm, the first focusing rod lens 71 should be made longer than 1/18 pitch, and vice versa. In this case, you can shorten it. However, when shortening the length, the displacement of the beam waist in the optical axis direction cannot be compensated for unless the movable distance of the fiber in the optical axis direction is increased. Another feature of this invention is that reflected light at the coupling portion does not return to the semiconductor laser. For example, if the entrance surface of a fiber is in the air, Fresnel reflection cannot be avoided, and since the entrance surface of the fiber is most narrowed down to the beam waist W ₂₂ , the exit end surface of the semiconductor laser has the original beam waist W ₀ . narrowed down. When the second focusing rod lens 72 is integrated with the entrance surface of the fiber using an optical adhesive as shown in the embodiment of the present invention, the reflectance of the entrance end surface of the fiber is reduced to one-tenth or less. On the other hand, the reflecting surfaces newly added compared to the conventional confocal system (FIG. 1C), that is, the opposing surfaces 83 and 84 of the focusing rod lenses 71 and 72, pass through the first focusing rod lens 71 twice. Returning, the length of the first converging rod lens 71 is 1/18.
When it comes to pitch, the length is equivalent to 1/9 pitch.
The value of W ₂₁ becomes 7.7 μm (from FIG. 5), and the light enters the ball lens 6. The rate at which the returning light is coupled with the spot size W ₀ of the semiconductor laser 1 is W ₁ = 182 μm.
It is sufficient to calculate the efficiency of W ₂₁ =7.7 μm, which is about 21 dB, which is 10 dB smaller than the fiber entrance surface and can be almost ignored. The reflection from the incident surface 85 of the first converging rod lens 71 is a combination of spots with a spot size W ₁ =182 μm, and a 0.5 degree angular shift causes a loss of 64 dB. That is, the light from the semiconductor laser 1 is converted into parallel light by the ball lens 6, and then the light from the semiconductor laser 1 is converted into parallel light by the first converging rod lens 71.
When the optical axis enters the converging rod lens 7,
0.5 if it is 0.25 degrees off perpendicular to the plane of incidence of 1.
Since an angular shift of degrees occurs with respect to the reflected light, it is thought that in reality, the effect of suppressing the reflected light due to the angular shift of the degree shown above always occurs. In order to more reliably remove the reflected light, the incident surface 85 of the first converging rod lens 71 may be tilted about 1 degree out of 90 degrees with respect to the optical axis. FIG. 7 shows another embodiment of the invention. The semiconductor laser 1 is mounted on a heat sink 110, and a ball lens holder 111 with a ball lens 6 attached thereto is mounted on the heat sink 110.
The ball lens 6 is attached using a flux (AuSn, PbSn solder, etc.) so that the optical axis of the ball lens 6 and the center axis of the ball lens 6 match. The heat sink 110 is fixed to a support 114, and in order to ensure the reliability of the semiconductor laser 1, the semiconductor laser 1, the heat sink 110, and the holder 111 are covered with a cap 113 having a window 112 made of sapphire or the like and hermetically sealed. The convergent rod lens 7 constituting the second lens is attached to the outside of the sapphire window 112 using an optical adhesive for the first convergent rod lens 71. Here, the sapphire window 112 is inclined by about 1 degree with respect to the optical axis. The optical fiber has a fiber holding pipe 301 with a capillary 300 whose inner diameter is slightly larger than the outer diameter of the fiber, and the bare fiber part 21 and the fiber core part 22 are inserted and fixed by adhesive, and the end face of the capillary 300 is fixed with an optical fiber. The end face of the fiber is polished by polishing. Fiber holding pipe 30
1 is inserted into the fiber lens holder 302 together with the second focusing rod lens 72 and fixed with adhesive or the like. At this time, enough optical adhesive is applied to the second converging rod lens 72 and the entrance end face of the fiber to prevent air from entering. Here, the length of the first focusing rod lens 71 is set to about 1/18 pitch, and the length of the second focusing rod lens 72 is set to about 1/18 pitch.
The length is set to a value slightly shorter than 1/4 minus the length of the first focusing rod lens 71. Fiber alignment is achieved by sliding the fiber lens holder 302 inside the outer frame 303 in the optical axis direction, and by sliding the fiber lens holder 302 inside the outer frame 303 in the direction perpendicular to the optical axis with the outer frame 303 in contact with the surface of the cap 113. Adjust by moving 303 in a direction perpendicular to the optical axis. The cap 113 and the outer frame 303 may be fixed by laser welding, solder, or adhesive, but it is preferable that the cap 113 and the outer frame 303 can be fixed instantly by welding or the like. Normally, axis misalignment is a problem in welding, etc., but with the above design, an error of ±7 μm is allowed to suppress efficiency deterioration within 1 dB, so it can be realized relatively easily. In the optical axis direction, ± several hundred micrometers is permissible, so fixing by welding or soldering can be realized with almost no problem. In the above explanation, the length of the converging rod lens as the second lens 72 was limited to 0.25 pitch or less, but the focal length of the converging rod lens is 0.5 pitch or less.
Since it remains unchanged even if pitch is added, the first and second
It can be seen that the same effect can be obtained even if the lengths of the converging rod lenses are increased by an integer multiple of 0.5 pitch. Further, as a countermeasure against reflection, it is also possible to apply an antireflection film to the interface between each lens and the air. As explained above, in the present invention, the strict fixing precision required for a single mode optical fiber in a conventional coupler of a semiconductor laser and a single mode optical fiber can be made several times looser, so that the semiconductor laser for the single mode optical fiber can be Coupler production will become as easy as that for multimode fibers, and significant improvements in productivity, yield, reliability, and cost reduction can be expected. In particular, conventional couplers use adhesive to fix the fibers and wait for the fibers to harden while aligning, resulting in poor manufacturing efficiency and reliability problems. It has the great advantage of being able to be fixed with a highly reliable solder material using a heating method, laser welding method, or the like. Furthermore, since reflections at the coupling portion, which cause deterioration of transmission characteristics, can be suppressed to a low level, the present invention can also be applied to semiconductor laser couplers that require strict measures against reflections, such as analog transmission. Furthermore, as is clear from the above principles, the present invention can be applied to any coupling system for coupling with a single mode fiber. For example, in FIG. 7, the ball lens 6 and the first focusing rod lens 71 are used individually, but they may be considered as a combined lens and replaced with one lens. At this time, the function of one lens is to perform image conversion where the distance a between the semiconductor laser and the lens is greater than f, where the focal length is f, and the image magnification is given by f/(a-f). be. A convergent rod lens or a spherical lens can be considered as a single lens, but these lenses can reduce the spot size of a single mode fiber to 5 μm.
Instead of narrowing down to a spot size that is several times larger (the spot size of a single mode optical fiber with a second convergent rod lens), one lens is arranged so as to couple it to a spot size several times larger than that.

[Brief explanation of the drawing]

Figure 1 is a side view showing various examples of conventional coupling methods between a semiconductor laser and a single mode optical fiber;
The figure is a principle diagram explaining a conventional confocal compound lens system, and FIG. 3 is a configuration diagram showing an embodiment of the present invention.
Fig. 4 is a principle diagram of Fig. 3, and Fig. 5 shows the image formed by the first converging rod lens when the converging rod lens used as the second lens is divided to explain the axis misalignment characteristics in this invention. Figure 6 shows an example of calculating the value of the beam waist W ₂₁ , and shows the amount of movement (ΔD) of the beam waist W ₂₁ with respect to the change in the distance (Δd) between the first focusing rod lens and the second focusing rod lens. FIG. 7 is a diagram showing calculated values obtained by varying the length of the first focusing rod lens, and is a side sectional view illustrating an embodiment of the present invention. 1: Semiconductor laser, 2: Single mode optical fiber, 3: Core of optical fiber, 4: Micro cylinder or spherical lens, 5: Tapered tip spherical lens with fiber, 6:
1st lens (spherical lens), 7: 2nd lens (converging rod lens), 71: 1st converging rod lens, 72: 2nd converging rod lens, 11
0: Heat sink, 111: Ball lens holder, 1
12: Sapphire window, 113: semiconductor package cap, 300: capillary, 301: fiber holding pipe, 302: fiber lens holder, 303: outer frame.

Claims (1)

[Claims] 1. A single mode optical fiber semiconductor in which a semiconductor laser chip, a first lens, a second lens, and a single mode optical fiber are arranged in this order to couple the semiconductor laser and the single mode optical fiber. In the laser coupler, the ratio of the focal lengths of the second lens and the first lens is determined by the beam radius of the light beam guided through the single mode optical fiber in the directions perpendicular and parallel to the junction surface of the semiconductor laser. The second lens is selected to be approximately equal to the ratio of the geometric mean of the beam radius at the beam waist, and a focusing rod lens whose length is slightly shorter than (1/4 + 1/2 l) pitch is used. (1/36+1/2m) Pitch: (2/9+1/2n
) pitch to (1/12+1/2m) pitch: (1/6+1/2
n) Pitch (l, m, n are each positive integers including zero)
The one with the shorter focal length in the divided converging rod lens is fixed in the same housing as the optical fiber.
The pitch is the refractive index distribution of the focusing rod lens.
(r)=n ₀ (1-α ² /2r ² ) (r is the distance from the central axis), then the length is given by 2π/α).