JPH02219002A - Diffraction grating and its production - Google Patents

Abstract

PURPOSE:To enhance diffraction efficiency and to decrease the difference in efficiency by the polarization direction of incident light by having a substrate, a photosensitive layer provided on the substrate and a reflecting film to be provided on the photosensitive layer and specifying the grating groove shape and the wave length region and the disposition to be used, respectively. CONSTITUTION:The photosensitive layer 12 is provided on the substrate 11 and periodic grooves are inscribed on the photosensitive layer 12 by holographic exposing. The reflecting film 13 as thin as not to impair the shape inscribed on the interference exposing layer 12 is thereafter provided to constitute the diffraction grating. The parameters h, gamma, phi have the shape satisfying the range of equation II when the spacing between the gratings is designated as d, k=2pi/d, the coordinates perpendicular to the groove direction of the gratings as x and the shape eta (x) of the grating grooves is expressed by formula I. The wavelength region to be used is standardized by the inter-grid spacing d and is expressed by formula III. Further, the disposition to be used is executed in the range satisfying equation IV where the incident angle is designated as theta and the first order retroangle as thetaL. The diffraction efficiency in non- polarization is enhanced in this way and the difference in the diffraction efficiency by the polarization direction of the incident light is decreased.

Description

DETAILED DESCRIPTION OF THE INVENTION Field of Industrial Application The present invention relates to a diffraction grating used in spectroscopic equipment, optical communication equipment, etc., and a method for manufacturing the same.

BACKGROUND OF THE INVENTION Conventionally, diffraction gratings have been used as optical wavelength dispersion elements mainly in spectroscopic measurement equipment, and are used to select and extract a desired wavelength from a certain wavelength range.

However, it is known that the intensity of diffracted light, that is, the diffraction efficiency, of diffraction gratings is generally significantly affected by the wavelength, polarization direction, and angle of incidence of incident light.

Hereinafter, a conventional diffraction grating and its manufacturing method will be explained.

For example, B, BRIEF, R, PRAUIL, M,
Prydon and Day, Maystr: Microwave Verification of a Numerical Optimization Op Fourier Gratings, Applied Fixtures, 24. Nα2
．． P, 14 (1981) CP, B11ek, R,D
eleuil, M., Breidneand D., Ma.
ystre: Microwave verificat
ion of numerical optim
Ization of Fourior grati
ngs.

Applied Physics, 24. Nα2. P
, 147 (1981)], the efficiency of the Fourier grating can be improved by local theoretical optimization, since it is possible to produce a Fourier grating that synthesizes a profile by a combination of the fundamental frequency and the second-order vermomicks using holographic exposure technology. is thicker than the efficiency of the echelette grating for a wide wavelength range, and this theoretical prediction has been experimentally verified in the microwave region.

Here, in general, the Fourier grating is defined by the combination 'I (x)−h {sin(kx)+1 sin
(2Kx-φ)), and h, γ, and φ are parameters that determine the shape.

Note that the diffraction grating shape shown in the literature is h=0.42 in this case. r=0.286. Corresponds to -90°.

Next, the manufacturing method uses the conventional three-beam interference exposure method using coherent light, but the sensitivity characteristics of the photosensitive layer provided on the substrate depend on the amount of exposure and the thickness of the photosensitive layer after development. Manufactured within a range where the residual amount has a nearly linear relationship.

Figure 16 shows the relationship between the exposure amount and the remaining film thickness of the photosensitive layer after development, that is, the sensitivity curve. In this case, the photosensitive layer is a positive type photosensitive material, that is, the exposed area is removed by development. It shows the photosensitive material. 16a is the intensity distribution of exposure interference fringes, 16b is the cross-sectional shape of the grating after development, 1
6c is a sensitivity curve.

According to this, since the sensitivity curve is linear, if a Fourier grating is formed, for example, M.

Prydun, Niss, Jimachi Hanson, Elouinilsson & H, Arlen: Plased Holographic Gratings, Optica Acta, 2
6. Statement 11. P, 1427 (1979) (M, Br
eidne, S., Johansson, L-E Ni1.
sson and H, Ahlen: Blazed h
olographic gratings.

Optica Acta, 26. kll, P, 142
7 (1979)), two exposure steps and precise positioning were required.

Problem to be Solved by the Invention However, keeping in mind that the local optimization shown in this document is used in spectroscopic instruments, the difference between the angle of incidence on the diffraction grating and the angle at which the first-order diffracted light is reflected is 17
This is done on the condition that: This is, for example, λ/d=
1 means that the light is incident from an angle 8.5 degrees away from the Littrow angle.

When using a diffraction grating for optical communication, it is often used near the beam draw arrangement, and is not used while maintaining an angular difference of 17°. Therefore, the superiority of Fourier lattice based on local optimization in this case has not been quantitatively stated. This is the first problem.

In fact, although we have verified our theoretical predictions in the microwave region, the shape of the actual diffraction grating is quite different from the locally optimized shape. In general, the wavelength range used in optical communications is 0.8
The wavelength range is from μm to 1.55 μm band. From this, it cannot be concluded that the superiority of locally optimized grating shapes has been demonstrated in all wavelength ranges, only in the microwave region. This is the second problem.

Furthermore, the lattice shape obtained by calculation assumes that the reflective surface is a perfect conductor, and does not take into account the dielectric constant of the reflective surface.

Regarding polarization, assuming that the polarization direction perpendicular to the grating groove direction is the TM wave, and the polarization direction parallel to the grating groove direction is the TE wave, M.G., Moharram and T.K. Of America Knee, 31st Floor, LIP, 1780 (1986) CM,
G,Moharaa and γ,Ni,Gaylord
:Rigorous coupled-@ave an
lysis of metallic surface
-relief gratings, Journal
of Optical5ociety of A+ee
rica A, 3. Na11. P, 1780 (1
According to 986)), although the lattice shape is square wave-like,
Due to the dielectric constant of the metal on the surface of the diffraction grating, the efficiency changes depending on the angle of incidence compared to when it is a perfect conductor, especially in the wavelength range of several μm.
It is stated that the change in efficiency of TM waves is more rapid than that of TE waves in the following regions. This is the third problem.

Furthermore, when actually manufacturing a Fourier grating, there is a problem in that, as described above, two exposure steps and precise positioning must be performed using a large-scale apparatus.

In view of the above problems, the present invention provides a diffraction grating and a diffraction grating having a grating groove shape that has a high diffraction efficiency especially in a retro arrangement and reduces the difference in efficiency depending on the polarization direction of incident light, for use in optical communication. The present invention provides a method for manufacturing.

Means for Solving the Problems In order to achieve the above objects, the diffraction grating of the present invention has a substrate, a photosensitive layer provided on the substrate, and a reflective film provided on the photosensitive layer. x, grid spacing d, k=
2π/d, the grating groove shape η(x) is η(x) −h (sln(Kx)+ysin (2K
x-90")), the parameter h,
γ and φ satisfy the following ranges: 0.05≦γ≦0.32 0.26≦2 h/d≦0.52, and the wavelength range to be used is normalized by the lattice spacing to be 0.67≦λ/d≦1. .15, and the arrangement used satisfies θ −5° ≦θ≦θL+5°, where θL is the angle of incidence on the diffraction grating and the first-order cross-draw angle is θL.

Preferably r, 2h/d, and λ/4 are each 0.1
Any material that satisfies ≦γ≦0.25, 0.35≦2h/d≦0.45, 0.70≦λ/d≦1.1 is sufficient.

As for the manufacturing method, the photosensitive layer interferes within the range where the second derivative of the curve representing the exposure amount and the remaining film thickness of the photosensitive layer after development has a positive value and the sensitivity characteristic does not have a third-order inflection point. All you have to do is expose it to light.

The exposure intensity distribution is I (x), the sensitivity curve in the exposure range is approximated as a quadratic curve, and the quadratic and linear coefficients are respectively a
, b, then 1(x)=io+I, s In (kx), 2h/d"l, l b+2a l0 7=a II /2 l b+2 a I.

It is sufficient if the relational expression holds true.

Preferably, for example, in a positive photosensitive material, 1, -tl, < l b/2a In addition, in negative photosensitive materials, b<o, 10+l, >lb/2a It is sufficient if it satisfies the following.

Effect The present invention has a special grating shape of the diffraction grating with the above configuration, and has a diffraction efficiency of 85% or more for non-polarized light in the Littrow arrangement, and a difference in diffraction efficiency depending on the polarization direction of the incident light is within 10%. A diffraction grating can be obtained, and 1
Such a diffraction grating can be manufactured by multiple exposures.

EXAMPLE Hereinafter, a diffraction grating and a method for manufacturing the same according to an example of the present invention will be described with reference to the drawings.

FIG. 1 shows a cross-sectional view of a diffraction grating in one embodiment of the present invention. In Fig. 1, la is the substrate, 1b
is a photosensitive layer, and IC is a reflective film.

A photosensitive layer 1b is provided on a substrate la, and periodic grooves are imprinted on the photosensitive layer 1b by holographic exposure. after that,
A reflective film IC that is thin enough not to damage the shape engraved on the photosensitive layer 1b is provided to constitute one diffraction grating.

FIG. 2 is a perspective view of the diffraction grating. 21 is a substrate, 22 is a photosensitive layer, and 23 is a reflective film. In the figure, the substrate is flat and square, but the shape of the substrate is not limited to this.

FIG. 3 shows a conceptual diagram of a method for manufacturing a diffraction grating according to an embodiment of the present invention. In FIG. 3, 31 is the intensity distribution of the interference light to be exposed, 32 is the sensitivity curve of a positive photosensitive layer, that is, the exposed portion of the photosensitive layer is removed after development, and 33 is the sensitivity curve after development. It has a lattice shape.

In the case of the two-beam interference exposure method using coherent light, the intensity distribution of the light holographically exposed on the photosensitive layer is generally a trigonometric function, and if this is directly transferred to the photosensitive layer, it will have a sine wave shape. By imparting nonlinearity to the sensitivity characteristics, the lattice shape after development can be changed from a sinusoidal shape to a distorted shape. The figure shows one example for realizing T=0.2, 2h/d=0.435 in a Fourier lattice, and the intensity distribution is 1(x)=I. +I, s in (kx), l0 = 2.32 mJ, I, = 1.74 mJ.

d=0.348 μm In addition, 1/l is a curve obtained by quadratic approximation of the sensitivity curve showing the remaining thickness of the photosensitive layer in the exposed range of 1/10. =aE2+
bE+c and a=1.0. It shows the conditions where b--0,899, c"0.25. If it can be approximated to a quadratic curve in the exposed range, it does not have a third-order inflection point.

The intensity distribution 31 of the interference light is a trigonometric function, and the sensitivity curve 32 has a sensitivity characteristic in which the second-order differential coefficient is always a positive value within the exposure range. As the thickness increases, the change in the remaining thickness of the photosensitive layer becomes smaller. Therefore, when exposed with the intensity distribution 31, the grating shape 33 after development becomes a distorted shape from a sine wave in which the peaks of the grating grooves are thin and the valleys are round. A diffraction grating can be constructed by providing a reflective film as thin as possible without damaging the grating shape.Although the embodiments shown above are for positive-type photosensitive materials, negative-type photosensitive materials, that is, those that are not exposed to light, can be constructed. It can be easily inferred that the same thing can be said about a photosensitive material whose photosensitive layer is removed after development.Therefore, the numerical examples shown above are not limited thereto.

FIG. 4 shows the diffraction efficiency of the diffraction grating according to the first embodiment of the present invention. In Figure 4, the shape is γ = 0.05
．． λ/d=0.7. A curve 41 shown in the experiment is the diffraction efficiency for the TE wave component, and a curve 42 shown by a - dotted chain line is the diffraction efficiency for the TM wave component. A curve 43 shown by a broken line is the diffraction efficiency when unpolarized light is incident. Under the conditions of the diffraction grating beam draw arrangement, numerical calculations were performed by Y., Okuno & T., Mazda: Efficient Technique Forsa Numerical Solution of Diffraction by Afourie
Grating, Journal of the American Society of America, 4. Nn3. P, 46
5 (19B?) (Y.

okuno and γ, Matsuda: Effic
ient technique forthe nu+
weflcal 5 solution of diffr
action by aFoqrier graLln
g, Journal of 0pLical 5oci
etyof America A, 4. P, 465 (
The diffraction grating surface was made to be a perfect conductor using the method described in (1987)).

If you do the above, the diffraction efficiency 43 for non-polarized light will be 85%.
The above ranges are 0.26≦2 h/d≦0.46. In particular, the difference in diffraction efficiency depending on the polarization direction is 0.2
The polarization dependence is small within 10% when 6≦2 h/d≦0.44. Figure 5 shows the grating shape of the diffraction grating that was numerically calculated. 51 is a grid shape. In the figure, this corresponds to the shape at 2 h/d=0.383.

Due to the addition of the second harmonic, the shape is slightly more distorted than the sine wave groove shape.

FIG. 6 shows the diffraction efficiency of the diffraction grating according to the second embodiment of the present invention. In Figure 6, the shape is γ-0,32
, λ/d = 0.9.

In this case, the range in which the diffraction efficiency 63 for unpolarized light is 85% or more is 0.31≦2 h/d≦0.46. In this range, the difference in diffraction efficiency depending on the polarization direction is within 8%. FIG. 7 is a diagram showing the grating shape of the diffraction grating for which numerical calculations were performed, and 71 corresponds to the shape when 2 h/d =0.40. As the distortion from the sinusoidal shape increases, the range of high diffraction efficiency shifts toward the larger 2 h/d.

FIG. 8 shows a third embodiment. In FIG. 8, the shape is r = 0.1 and λ/d = 0.67.

In this case, the range in which the diffraction efficiency 83 for non-polarized light is 85% or more is 0.26≦2 h/d≦0.43. In this range, the difference in diffraction efficiency depending on the polarization direction is within 5%, and the polarization dependence is particularly small. FIG. 9 shows the grating shape of the diffraction grating that was numerically calculated.

In the figure, 91 corresponds to the shape at 2 h/d -0.35.

FIG. 10 shows a fourth embodiment.

In FIG. 10, the shape is γ-0.2. λ/d=0.8.

In this case, the range in which the diffraction efficiency 10c for non-polarized light is 85% or more is maintained at 2h/d, which is a fairly wide range of 0.29≦2/hd≦0.48. In this range, the difference in diffraction efficiency depending on the polarization direction is also within 10%.

FIG. 11 is a graph showing the wavelength dependence at 2 h/d =0.47 in FIG. 1O.

In such a case, the diffraction efficiency 11c for unpolarized light is 85%.
The above range is 0.78≦λ/d≦1.15, but in the range 0.88≦λ/d≦1.0, the difference in diffraction efficiency between both polarized lights becomes large. FIG. 12 shows the grating shape of the diffraction grating that was numerically calculated. In the figure, 12a corresponds to the shape when 2h/d=0.47.

FIG. 13 is a graph showing the wavelength dependence when the diffraction efficiencies for both polarized lights are equal in 2 h/d=0.435 in FIG. 10.

In this case, the range in which the diffraction efficiency 13c for non-polarized light is 85% or more is 0.70≦λ/d≦1.15, especially in a wide wavelength range that satisfies 0.73λ/d≦1.07. The difference in diffraction efficiency depending on the direction is within 10%.

FIG. 14 shows the results of measuring the dependence on the incident angle using a diffraction grating that was actually fabricated by providing a photosensitive layer on a glass substrate and performing one holographic exposure. Grating groove spacing is 1.3μm
The surface is vapor-deposited with gold having a reflectance of 99% or more, and a semiconductor laser beam in the 1.3 μm band is used as the measurement light. In Figure 0, the incident angle of 30° corresponds to the Littrow angle. The curve 14a is the TE wave, the curve 14b is the TMI, the curve 14c is the diffraction efficiency in unpolarized light, and Δ and the mouth mark are the TE wave component 14a, respectively.
.

This is the measurement point of the TM wave component 14b.

Generally, under the condition of λ/d=1, the diffraction efficiency increases in a first-order beam-draw arrangement, and decreases as the incident angle moves away from the draw angle. In the figure, the diffraction efficiency for both polarized lights is at its maximum in the draw arrangement, which has a high diffraction efficiency of 95% for non-polarized light. The diffraction efficiency at that time was almost completely inferior to the calculated value in FIG. As can be seen from the figure, for incident angles within 5 degrees from the Littrow angle, the diffraction efficiency is maintained at over 85% for unpolarized light, and the difference in efficiency due to polarization is small, within 10%. . 5 from the corner of Retro
``The angular difference between the incident angle and the reflection angle of the first-order diffracted light in the case of deviation corresponds to 10 degrees, and if the incident angle is farther from this, the diffraction efficiency 14b of the TM wave component will change due to the dielectric constant of the diffraction grating surface. It can be seen from the figure that the E-wave component 14a becomes sharply lower than that of the E-wave component 14a.1 For example, at an angle 10° away from the Littrow angle, the diffraction efficiency of both polarized lights differs by more than 20%, and therefore the diffraction efficiency of unpolarized light 14c also becomes lower. FIG. 15 is a photograph of the measured grating cross-sectional shape of the diffraction grating using an SEM. In this case, the grid shape is r = 0
．． 2, λ/d=1.0, 2h/d=0.435. Furthermore, parameter 2 h/d
, γ, and λ/d, many shapes can be considered by changing the combinations of , γ, and λ/d, so the present invention is not limited to the examples given above.

Effects of the Invention As described above, the present invention includes a substrate, a light-transmitting layer provided on the substrate, and a reflective film provided on the photosensitive layer, and the lattice groove spacing is set to d, k.
=2π/d, and the coordinate perpendicular to the grating groove direction is X, the grating groove shape η(x) is η(x)=h 1sin(x)+
y sin (2 to 90'), the parameters h, r, and φ are 0.26≦2 h
/d≦0.52 0.05≦γ≦0.32, and the wavelength range used is normalized to the lattice spacing d"il', and 0.67≦λ/d≦1.15 By setting the incident angle to θL and setting the Littrow angle of the first order to θL, the diffraction efficiency in non-polarized light can be 85% or more by arranging the incident angle in a range that satisfies θ −5′ ≦θ≦θL+5°. The difference in diffraction efficiency depending on the polarization direction of incident light can be kept within 10%.

In addition, the photosensitive layer has sensitivity characteristics in which the second-order differential value of the curve representing the exposure amount and the remaining film thickness of the photosensitive layer after development is a positive value, and there is no third-order inflection point, and the exposure interference Letting the coordinate perpendicular to the direction of the stripes be x, the stripe interval d, and k-2π/d, the exposure intensity distribution r(x) is 1(x)=. 4+11 s i
n (kx) When the curve of the sensitivity characteristic of the photosensitive layer is approximated to a quadratic curve within the exposure intensity range, the quadratic coefficient and the first-order coefficient are respectively a and b, 2h/d=I, |b+2al.

A diffraction grating having a similar effect can be realized by one exposure by manufacturing so as to satisfy r−a 11 /2 l b+2 a, Io.

[Brief explanation of the drawing]

Fig. 1 is a cross-sectional diagram of a diffraction grating in an embodiment of the present invention, Fig. 2 is a perspective view of the diffraction grating shown in Fig. 1, and Fig. 3 is a sensitivity curve in a positive type photosensitive layer and the resulting diffraction. A conceptual diagram showing the grating shape of the grating, FIG. 4, shows r=0.05 of the diffraction grating in the first embodiment of the present invention. r/d=0.67
, a calculation diagram of the diffraction efficiency for 2 h/d in the Littrow configuration, FIG. 5 shows r=0.05. 2 h/d=0.383
FIG. 6 is a diagram showing the grating shape of the diffraction grating in the second embodiment of the present invention, where γ=0.32. γ/d-0.9, calculation diagram of diffraction efficiency for 2 h/d in Littrow configuration, Figure 7 is a grating shape diagram at γ-0.32, 2h/d = 0.40, Figure 8 is a calculation diagram of the diffraction efficiency for 2 h/d of the diffraction grating in the third embodiment of the present invention with r = 0.1, λ/d -0,7, and Littrow configuration. .1,2 Grating shape diagram at h/d=0.35, Fig. 10 shows the 7= of the diffraction grating in the fourth embodiment of the present invention.
0.32. λ/d=0.8, 2 in Retro-configuration
A calculation diagram of diffraction efficiency for h/d pairing, FIG. 11 shows r=0.
2,2h/d=0.47, calculation diagram of diffraction efficiency for λ/4 in Littrow configuration, Figure 12 is γ-0.2,2h/
The lattice shape diagram at d=0.47, Fig. 13 is γ=0.2,
2h/d=0.435, calculation diagram of diffraction efficiency for λ/4 in Littrow configuration, FIG. 14 is γ-〇, 2. λ/d=
1.0,2 h/d = 0.435, Figure 15 is an SEM photograph of the measured diffraction grating, Figure 16 is the conventional one. FIG. 2 is a conceptual diagram showing a sensitivity curve in a positive photosensitive layer in a method for manufacturing a diffraction grating and a grating shape of a diffraction grating formed thereby. la, 21... board, lb, 22...
Photosensitive layer, lc, 23... Reflective film, 31...
... Intensity distribution of exposure interference fringes, 32 ... Sensitivity curve, 33 ... Grid shape after development, 41, 61, 8
1. lOa, 11a, 13a. 14a...Diffraction efficiency of TE wave component, 42, 62
゜82, lob, llb, 13b, 14b...
Diffraction efficiency of TM wave component, 43, 63, 83 10c11
c, 13c, 14c... Diffraction efficiency with non-polarized light. Name of agent Patent attorney Shigetaka Awano 1 person Kazutomi 13-Kaishaku 2h/l:l Fig. 111)%Ω-
F1a lattice shape Otsu 1 2h/cl Fig. 2h/d Fig. 10 2A /d Diffraction effect of TEiff component 4c- wfif&*, Diffraction table Fig. 14 Fig. 15 Figure λ Intrajective θ (de+3]

Claims (1)

Scope of Claims: (1) A substrate, a photosensitive layer provided on the substrate, and a reflective film provided on the photosensitive layer, the grating groove spacing being d, k=2π
/d, and the coordinate perpendicular to the grating groove direction is x, and the grating groove shape η(x) is η(x)=h{sin(kx)+γsin(2kx−9
0゜)}, the parameters h, γ, φ are
It has a shape that satisfies the range of 0.26≦2h/d≦0.52 0.05≦γ≦0.32, and the wavelength range to be used is normalized by the grating spacing d, and 0.67≦λ/d≦ 1.15, and furthermore, the diffraction grating (2) is arranged in a range that satisfies θ_L−5°≦θ≦θ_L+5°, where the incident angle is θ and the first-order Littrow angle is θ_L. It has a substrate, a photosensitive layer provided on the substrate, and a reflective film provided on the photosensitive layer, and the grating grooves are arranged such that the grating groove interval is d, k=2π/d, and the coordinate perpendicular to the grating groove direction is x. Shape η(x)
But, η(x)=h{sin(kx)+γsin(2kx−9
0゜)}, the parameters h, γ, φ are
It has a shape that satisfies the range of 0.35≦2h/d≦0.45 0.10≦γ≦0.25, and the wavelength range to be used is normalized by the grating spacing d, and 0.67≦λ/d≦ 1.15, and the diffraction grating is used in a range that satisfies θ_L−5°≦θ≦θ_L+5°, where the incident angle is θ and the first-order Littrow angle is θ_L. (3) The photosensitive layer has a substrate, a photosensitive layer provided on the substrate, and a reflective film provided on the photosensitive layer, and the photosensitive layer has a curve representing the amount of exposure and the residual rate of film thickness of the photosensitive layer after development. A diffraction grating characterized in that interference exposure is performed in a range having a sensitivity characteristic in which a second-order differential value of is a positive value and a third-order inflection point does not exist. (4) The wavelength λ to be used is normalized by the grating spacing d, and 0.
67≦λ/d≦1.15, and the arrangement used is performed within a range that satisfies θ_L−5°≦θ≦θ_L+5°, where the incident angle is θ and the first-order Littrow angle is θ_L. Diffraction grating according to item (3). (5) When the coordinate perpendicular to the direction of the interference fringes to be exposed is x, the fringe interval is d, and k = 2π/d, the exposure intensity distribution I(x) is: I(x) = I_0+I_1sin(kx) Sensitivity of the photosensitive layer When the characteristic curve is approximated to a quadratic curve within the exposure intensity range, the quadratic coefficient and first-order coefficient are a and b, respectively, and the grating shape η(x) of the diffraction grating is expressed as η(x)=h {sin(kx)+γsin(2kx-9
0°)}, then 2h/dI_1, |b+2aI_0| γ=aI_1/2|b+2aI_0| The diffraction grating according to claim (3), wherein the following is satisfied. (6) The diffraction grating according to claim (5), wherein a positive photosensitive material is used for the photosensitive layer, and the following is satisfied: I_1+I_0<|b/2a|. (7) The diffraction grating according to claim (5), wherein a negative photosensitive material is used for the photosensitive layer, and when b<0, I_0+I_1>|b+2a| is satisfied. (8) A photosensitive layer is provided on a substrate, and the second derivative of the curve representing the exposure amount and the remaining film thickness of the photosensitive layer after development is a positive value, and the third inflection point is a positive value. A method for producing a diffraction grating, which comprises performing interference exposure in a range having non-existent sensitivity characteristics and providing a reflective film thereon. (9) Assuming that the coordinate perpendicular to the direction of the interference fringes to be exposed is x, the fringe interval is d, and k = 2π/d, the exposure intensity distribution I(x) is: I(x) = I_0+I_1sin(kx) Sensitivity of the photosensitive layer When the characteristic curve is approximated to a quadratic curve within the exposure intensity range, the quadratic and first order coefficients are a and b, respectively, and the grating shape of the diffraction grating is η(x), η(x)= h{sin(kx)+γsin(2kx-9
0°)}, then 2h/d=I_1|b+2aI_0| γ=aI_1/2|b+2aI_0| The method for manufacturing a diffraction grating according to claim 8, wherein the following is satisfied. (10) The method for manufacturing a diffraction grating according to claim (9), characterized in that a positive photosensitive material is used for the photosensitive layer, and I_0+I_1<|b/2a| is satisfied. (11) The method for manufacturing a diffraction grating according to claim (9), wherein a negative photosensitive material is used for the photosensitive layer, and when b<0, I_0+I_1>|b/2a| is satisfied.