APPARATUS AND METHODS FOR DEPOSITING VARIABLE INTERFERENCE FILTERS

Abstract

Apparatus for depositing one or more variable interference filters onto one or more substrates comprises a vacuum chamber, at least one magnetron sputtering device and at least one movable mount for supporting the one or more substrates within the vacuum chamber. The at least one magnetron sputtering device is configured to sputter material from a sputtering target towards in the mount, thereby defining a sputtering zone within the vacuum chamber. At least one static sputtering mask is located between the sputtering target and the mount. The at least one static sputtering mask is configured such that, when each substrate is moved through the sputtering zone on the at least one movable mount, a layer of material having a non-uniform thickness is deposited on each said substrate.

Claims

1. Apparatus for depositing one or more variable interference filters onto one or more substrates, the apparatus comprising a vacuum chamber, at least one magnetron sputtering device and at least one movable mount for supporting the one or more substrates within the vacuum chamber, the at least one magnetron sputtering device being configured to sputter material from a sputtering target towards the mount, thereby defining a sputtering zone within the vacuum chamber, wherein at least one static sputtering mask is located between the sputtering target and the mount, the at least one static sputtering mask being configured such that, when each substrate is moved through the sputtering zone on the at least one movable mount, a layer of material having a non-uniform thickness is deposited on each said substrate.

2. The apparatus according to claim 1, wherein the at least one static sputtering mask is configured such that the layer of material deposited on each substrate varies in thickness along a first direction and is substantially uniform in thickness along a second direction substantially perpendicular to the first direction.

3. The apparatus according to claim 2, wherein the at least one static sputtering mask is configured such that the layer of material deposited on each substrate varies linearly in thickness along the first direction.

4. (canceled)

5. The apparatus according claim 1, wherein the at least one static sputtering mask comprises one or more apertures configured such that the layer of material deposited on each substrate has a non-uniform thickness.

6. The apparatus according to claim 5, wherein the one or more apertures are elongate and tapered.

7. (canceled)

8. The apparatus according to claim 1, wherein the at least one magnetron sputtering device is a direct current (DC) magnetron sputtering device.

9. The apparatus according to claim 6, wherein the DC magnetron sputtering device is a pulsed DC magnetron sputtering device.

10. (canceled)

11. The apparatus according to claim 1, comprising two or more magnetron sputtering devices, each said magnetron sputtering device being configured to sputter material from a corresponding sputtering target towards the mount, thereby defining one or more respective sputtering zones within the vacuum chamber, wherein a corresponding static sputtering mask is provided between each sputtering target and the mount, each said static sputtering mask being configured such that, when each substrate is moved through each sputtering zone on the at least one movable mount, a respective layer of material having a non-uniform thickness is deposited on said substrate.

12. The apparatus according to claim 1, wherein the apparatus comprises at least one plasma processing device configured to direct plasma-generated gas ions towards the mount, thereby defining a plasma treatment zone within the vacuum chamber.

13. The apparatus according to claim 9, wherein the plasma processing device is configured to direct oxygen ions towards the mount.

14. A method for depositing one or more variable interference filters onto one or more substrates, the method comprising: at least one magnetron sputtering device sputtering material from a sputtering target to thereby define a sputtering zone within a vacuum chamber; at least one mount moving the one or more substrates through the sputtering zone; and providing at least one static sputtering mask between the sputtering target and the mount, the at least one static sputtering mask being configured such that a layer of material having a non-uniform thickness is deposited on each substrate moved through the sputtering zone.

15. The method according to claim 14, wherein the at least one static sputtering mask is configured such that the layer of material deposited on each substrate varies in thickness along a first direction and is substantially uniform in thickness along a second direction substantially perpendicular to the first direction, wherein the at least one static sputtering mask is configured such that the layer of material deposited on each substrate varies linearly in thickness along the first direction.

16. (canceled)

17. (canceled)

18. The method according to claim 14, wherein the at least one static sputtering mask comprises one or more apertures configured such that the layer of material deposited on each substrate has a non-uniform thickness, and wherein the one or more apertures are elongate and tapered.

19. (canceled)

20. (canceled)

21. The method according to claim 14, wherein the at least one magnetron sputtering device is a direct current (DC) magnetron sputtering device.

22. (canceled)

23. (canceled)

24. The method according to claim 14, wherein two or more magnetron sputtering devices sputtering material from corresponding sputter targets towards the mount, thereby defining two or more respective sputtering zones within the vacuum chamber; the at least one mount moving the one or more substrates through each said sputtering zone; and providing at least one corresponding static sputtering mask between each sputtering target and the mount, each said static sputtering mask being configured such that a layer of material having a non-uniform thickness is deposited on each substrate moved through each sputtering zone.

25. The method according to claim 14, further comprising at least one plasma processing device directing gas ions generated by a plasma towards the mount, thereby defining a plasma treatment zone within the vacuum chamber.

26. The method according to claim 25, wherein the at least one plasma processing device is configured to direct oxygen ions towards the mount, to thereby form a layer of metal oxide on each substrate moved through the plasma treatment zone.

27. An optical device comprising a substrate and a variable interference filter deposited thereon by the method of claim 14, wherein the substrate comprises two said variable interference filters located symmetrically thereon.

28. (canceled)

29. A variable interference filter comprising at least two sloping regions, across which the thickness of the variable interference filter varies, and one or more lower-gradient regions therebetween, across which the rate of change of thickness is less than the rate of change of thickness across each of the at least two sloping regions.

30. (canceled)

31. (canceled)

32. An optical sensor comprising at least one light source, at least one variable interference filter according to claim 29 and at least one light detector, the detector being configured to receive light emitted by the at least one light source and transmitted through at least one lower-gradient region of the at least one variable interference filter.

Description

DESCRIPTION OF THE DRAWINGS

[0149] An example embodiment of the present invention will now be illustrated with reference to the following Figures in which:

[0150] FIG. 1 shows a schematic cross section through variable interference filter deposition apparatus;

[0151] FIG. 2 shows the deposition apparatus of FIG. 1 with a substrate positioned in front of a first pulsed DC magnetron sputtering source;

[0152] FIG. 3 shows the deposition apparatus of FIG. 1 with the substrate positioned in front of a microwave plasma source;

[0153] FIG. 4 shows the deposition apparatus of FIG. 1 with a substrate positioned in front of a second pulsed DC magnetron sputtering source;

[0154] FIG. 5 shows a cross-section (along a first direction) through a layered variable interference structure deposited using the deposition apparatus of FIG. 1;

[0155] FIG. 6 shows a cross-section (along a second direction perpendicular to the first direction) through the layered variable interference structure of FIG. 5;

[0156] FIG. 7 shows a band-pass variable interference filter formed from two layered variable interference structures as shown in FIGS. 5 and 6;

[0157] FIG. 8 shows schematically the path of atoms sputtered from a target towards a rotating substrate, a first atom following path A passing through a sputtering mask to reach the target and a second atom following path B being blocked by the sputtering mask;

[0158] FIG. 9 shows a quarter of an erosion track profile formed on a sputtering target in a planar, rectangular magnetron sputtering device, the upper half of the Figure showing the profile as a contour plot and the lower half of the Figure showing the profile as a surface plot;

[0159] FIG. 10 shows the simulated thickness variation across the length of a deposited thin-film for a particular sputtering mask configuration taking into account two possible yield peak angles;

[0160] FIG. 11 compares experimental ellipsometry measurements of the thickness variation along the length of a thin-film deposited using a sputtering mask (having the same configuration as that used in the simulations of FIG. 10) with the results shown in FIG. 10;

[0161] FIG. 12 shows the measured absorption spectra at four different locations along the length of an optical linear variable interference filter deposited using the apparatus of FIG. 1;

[0162] FIG. 13 shows schematically a static sputtering mask used in the deposition of the optical linear variable interference filter of FIG. 12;

[0163] FIG. 14 shows the measured transmission spectra at seven different locations along the length of a second optical linear variable interference filter deposited using the apparatus of FIG. 1; and

[0164] FIG. 15 shows the linear variation in transmission wavelength peak height for the linear variable interference filter of FIG. 14.

DETAILED DESCRIPTION OF A FIRST EXAMPLE EMBODIMENT

[0165] FIG. 1 shows schematically deposition apparatus 1 for depositing one or more variable interference filters. The apparatus 1 includes a rotatable drum mount 2 mounted on a substantially horizontal axle 3, first and second pulsed DC magnetron sputtering sources 4 and 6, each provided with respective adjustable sputtering masks 5 and 7, and a microwave plasma source 8, all provided within a vacuum chamber 9. The drum mount 2 has the shape of a dodecagonal prism, the axle 3 extending along the longitudinal axis of the prism and each external prism face 10 comprising a bracket for receiving a deposition substrate.

[0166] The first pulsed DC magnetron sputtering source 4 is set up such that, when turned on, it generates and confines a substantially rectangular, planar plasma adjacent to a titanium target, to thereby sputter titanium atoms towards the drum. The second pulsed DC magnetron sputtering source 6 is also set up such that, when turned on, it generates and confines a substantially rectangular, planar plasma adjacent to a silicon target, to thereby sputter silicon atoms towards the drum. The microwave plasma source 8 is supplied with a flow of oxygen gas to generate a flow of oxygen ions towards the drum.

[0167] In use, a glass substrate 11 is mounted on one of the flat external surfaces of the drum and the drum is rotated continuously about the axle 3 at a speed of 60 rpm. The vacuum chamber is evacuated and backfilled with argon or another inert gas. The first pulsed DC magnetron sputtering source 4 is switched on, as is the microwave plasma source 8. As shown in FIG. 2, the substrate 11 is first rotated past the first sputtering source 4, which sputters titanium atoms towards the drum. A portion of the sputtered material is blocked by the sputtering mask 5 but the remainder of the sputtered material is able to pass through the sputtering mask 5 to deposit a few monolayers of titanium onto the substrate.

[0168] As shown in FIG. 3, the substrate is subsequently rotated past the microwave plasma source 8. As the substrate passes the microwave plasma source, high-energy oxygen ions are directed onto the substrate which oxidise the previously deposited few monolayers of titanium, completely converting this into a layer of TiO.sub.2.

[0169] Each time the substrate passes the first sputtering source 4, a few more monolayers of titanium are deposited onto the existing layers of TiO.sub.2. Each time the substrate passes the plasma source 8, these monolayers of titanium are converted into TiO.sub.2. Over repeated passes of the substrate past the first sputtering source and the plasma source, the thickness of a layer of TiO.sub.2 is built up a few monolayers at a time until a desired thickness is achieved.

[0170] At this point the first sputtering source 4 is switched off and the second sputtering source 6 is switched on. As the substrate is rotated past the second sputtering source, a portion of the sputtered material is blocked by the sputtering mask 7 but the remainder of the sputtered material is able to pass through the sputtering mask 7 to deposit a few monolayers of silicon on top of the layer of TiO.sub.2. The substrate is subsequently rotated past the plasma source which oxidises the silicon monolayers to form SiO.sub.2. Over repeated passes of the substrate past the second sputtering source and the plasma source, the thickness of a layer of SiO.sub.2 is built up a few monolayers at a time until a desired thickness is achieved.

[0171] As rotation of the drum continues, the first and second sputtering sources are repeatedly switched on and off to deposit alternating layers of TiO.sub.2 and SiO.sub.2 onto the substrate.

[0172] The rotational speed of the drum is such that only a few monolayers of material are deposited during each pass through each of the titanium and silicon sputtering zones, which provides precise control over the layer thicknesses and ensures that full oxidation of the sputtered titanium and silicon is possible in the microwave plasma oxidising zone, resulting in the formation of stoichiometric oxide layers. By physically separating metal deposition and oxidation zones, it is easier to control the thickness of the resultant metal oxide layers. In particular, it is easier to control and reproduce a gradient in thickness through the resultant metal oxide layers when metal deposition and oxidation steps are separated.

[0173] An example deposited layered structure is shown in FIGS. 5 and 6. Two layers of TiO.sub.2, 12A and 12B, and two layers of SiO.sub.2, 13A and 13B, have been deposited on the substrate, each individual layer itself comprising several monolayers of TiO.sub.2 or SiO.sub.2 respectively. The shape of the sputtering masks 5 and 7 have been selected to ensure that the thickness of each sputtered layer is uniform along a direction perpendicular to the axle (i.e. the thickness is uniform around the circumference of the drum), as can be seen in FIG. 5, but is non-uniform along a direction parallel to the axle, as can be seen in FIG. 6.

[0174] Because the refractive index of TiO.sub.2 is greater than that of SiO.sub.2, this layered structure functions as an interference filter. Because the thicknesses of each layer of TiO.sub.2 and SiO.sub.2 vary along at least one dimension, the structure in fact functions as a variable interference filter. Depending on the layering pattern and the layer thicknesses, the deposited device may function as a high-pass interference filter or a low pass interference filter. Two such interference filters may be sequentially deposited on opposing faces of a substrate, or two such interference filters may be bonded together, to form a band-pass interference filter as shown in FIG. 7. The layer thickness for optical interference filters typically ranges from 50 nm to 200 nm.

[0175] In practice, one or more substrates may be mounted to each external, planar face of the drum, permitting a plurality of layered structures to be deposited at the same time. This significantly increases throughput compared to existing deposition technologies. The polygonal drum permits a high surface area of substrate to be coated for a given vacuum chamber volume.

[0176] The thickness variation of the layers of deposited material, and hence the spectral properties of the deposited interference filters, are controlled through the shape of the sputtering masks 5 and 7 as explained in more detail as follows.

Detailed Description of Sputtering Mask Design

[0177] The inventors have developed a model used to simulate the thickness distribution of a film of material sputtered onto a substrate from a particular target. The model takes into account (i) the sputtering yield distribution across the sputtering target surface, (ii) the angular distribution of sputtered material, (iii) masking shielding, and (iv) the thin film growth process on a rotating drum. With this model the appropriate sputtering mask design can be obtained for a desired spatial thickness distribution.

[0178] As shown in FIG. 8, an emission angle, , may be defined as the angle between the flux of atoms ejected from a target within the sputtering device and the normal to said target surface. Similarly, an incident angle, , is defined as the angle between the incident sputtered atom flux and the normal to the substrate surface. The path taken by sputtered atoms which reach the substrate through the sputtering mask is labelled A. The path taken by sputtered atoms which are blocked by the sputtering mask and therefore do not reach the substrate is labelled B.

[0179] The sputtering yield at the sputtering target is a function of the sputtered particle density, the applied magnetic field (which is used to confine the plasma adjacent to the sputtering target), the cathode voltage applied to the sputtering target and the gas pressure.

[0180] During deposition gas pressure and cathode voltage typically remain constant and can be assumed to be uniform across the whole target surface. Accordingly, it is not necessary to take into account gas pressure or cathode voltage when modelling the sputtering yield distribution.

[0181] The magnetic field, B.sub.tan, measured tangential to the target surface, is typically highly non-uniform. This is reflected in the characteristically non-uniform erosion track profile formed on the target surface during sputtering. It has been found that, between maximum and minimum critical magnetic fields, ionisation is a linear function of the logarithm of a dimensionless magnetic field term, , defined as

[00001]$=\sqrt{\frac{e}{2.Math.m}}.Math.\left(\frac{a}{V_{\mathrm{dis}}^{\frac{1}{2}}}\right).Math.B_{\tan }$

where e is the electron charge, m is the electron mass, is the distance from the centre of the target to the location at which the magnetic field is tangential to the target surface, and V.sub.dis is the cathode bias applied to the target. Accordingly, the relative yield distribution at the target surface can be obtained by measuring the tangential magnetic field B.sub.tan. The magnetic field can be measured using a magnetometer or gauss meter, or using a Hall probe.

[0182] Alternatively, the relative yield distribution at the target surface can be obtained by measuring the erosion track profile directly. This method allows the target erosion speed (which is also related to the magnetic field B.sub.tan) to be determined directly, and is more useful for accurate thickness distribution simulations.

[0183] FIG. 9 shows one quarter of an example erosion track profile of a target (the full erosion track being substantially rectangular in shape). The upper half of the Figure shows a contour plot of the erosion track profile, the lower half shows a surface plot. Units are in millimetres. Such measured profile data can be used to generate a two-dimensional data array of the relative sputtering yield on the target surface, referred to as the trackyield, which is indicative of the target erosion speed. For example, the trackyield is proportional to the depth of the erosion track profile which can be measured directly.

[0184] The angular distribution of atoms ejected from the target must also be taken into account. The emission angle distribution is itself a function of the angle of incidence of sputter flux (i.e. the flux of ions used to sputter material from the target). The inventors have found that for polycrystalline material targets (such as titanium), the sputter yield tends to increase with increased angular deviation from the normal direction until a maximum is reached, and then decreases again as the angle of incidence approaches glancing incidence. For crystalline materials, such as silicon, the behaviour is more complicated and different minima and maxima are observed in the sputter yield as a function of angle of incidence, depending on the particular crystalline structure of the target. Minima are typically observed where ion bombardment occurs parallel to close-packed crystal structure directions because, in such orientations, ions may be directed along open crystal channels with fewer collision events dramatic enough to eject atoms or molecules from the target surface.

[0185] When microwave plasma pulsed DC magnetron sputtering is used, ion bombardment typically occurs approximately normal angle to the target surface (indeed, by considering particle collisions and a turbulent electric field, a mean incident angle of 7.95 to the target surface normal has previously been calculated using Monte Carlo simulations of sputtering). Accordingly, the variation of bombardment ion incident angles can be taken to be negligible for microwave plasma pulsed DC magnetron sputtering simulations and only the general function of the angular distribution of ejected particles is needed for simulations of magnetron sputtering.

[0186] For a crystalline target, the angular distribution of sputtered particles is related to crystal structure of the target and has previously been determined experimentally for, for example, a (100) Ag crystalline target bombarded by 100 eV Hg ions as set out by P. Sigmund in Fundamental processes in sputtering of atoms and molecules (SPUT92): symposium on the occasion of the 250.sup.th anniversary of the Royal Danish Academy of Sciences and Letters, Copenhagen, 30 Aug.-4 Sep. 1992: invited reviews. 43, (Kongelige Danske videnskabernes selskab, 1993). A similar distribution can be assumed for sputtering from a crystalline Si target.

[0187] The angular distribution of sputtered particles from polycrystalline targets has also previously been determined experimentally. According to the experimental results, polycrystalline targets can be divided into two groups dependent on their maximum emission angles measured relative to the target surface normal: a first group of targets in which the maximum emission angle is 0 (i.e. parallel to the surface normal); and a second group of targets in which the maximum emission angle occurs at 30 to the surface normal. For the first group, the angular distribution of sputtered particles, (), varies according to ()=A cos.sup.n . For the second group, ()=A cos.sup.n B cos.sup.m . In general, the inventors have found that the angular distribution of sputtered particles from a polycrystalline target can be approximated as:

()=cos(+.sub.0).sup.n

where .sub.0 is the angle of maximum emission and n is an adjustable parameter. In practice, both .sub.0 and n can be fit to an experimentally determined thickness distribution for a given target with a known mask shape.

[0188] Each point on the sputter target can be considered as a point source of sputtered particles. Accordingly, the number of particles arriving on the substrate surface is proportional to

[00002]$\frac{1}{r^2},$

where r is the distance between a point (X.sub.t, Y.sub.t, Z.sub.t) on the target and a point (X.sub.s, Y.sub.s, Z.sub.s) on the substrate. A factor of cos must also be introduced to take into account the projection effect, where is the angle between the deposition beam and the substrate surface normal.

[0189] A function describing the shape of the mask can be used to carry out a geometrical analysis to determine whether sputtered particles can reach the substrate surface or whether they will be blocked by the mask. This result is called the passrate and takes the values 1 (i.e. particles reach the substrate) or 0 (i.e. particles are blocked by the mask).

[0190] Accordingly, the probability that a particle sputtered from the point (X.sub.t, Y.sub.t, Z.sub.t) on the target will successfully travel to the point (X.sub.s, Y.sub.s, Z.sub.s) the substrate is given by:

[00003]$P=\frac{{\cos \left(+{}_0\right)}^n.Math.\mathrm{trackyield}.Math.\mathrm{passrate}.Math.\cos .Math..Math.}{r^2}$

[0191] In order to carry out a calculation of P across the target, the target is divided into a fine mesh to allocate each point on the target a set of coordinates of (X.sub.t,Y.sub.t,Z.sub.t). At any one particular point on the substrate, (X.sub.s, Y.sub.s, Z.sub.s), the relative thickness of the coating deposited can be obtained from the sum of the probabilities of all target mesh elements.

[0192] Because the film thickness is uniform around the circumference of the rotating drum, it should only be necessary to calculate the film thickness along the centre line of the drum plate which is parallel to the axis about which the drum rotates, in order to optimize the mask design. However, because the drum is rotating, each on the centre line of the plate is moving around the circumference of the drum during deposition. As shown in FIG. 2, it is possible to construct a new set of substrate coordinates for the point X.sub.s at the centre line adjusted for rotation through an angle . The corrected coordinates of each point on the centre line are (X.sub.s, Y.sub.s0R sin , Z.sub.s0R cos ), where (X.sub.s, Y.sub.s0, Z.sub.s0) are the coordinates on the rotating drum axis. The relative thickness at each said point can be obtained by integrating with respect to between .sub.0 and .sub.0. The thickness T(X.sub.s) is then given by

[00004]$T\left(X_s\right)=\underset{}{.Math.}.Math.\underset{X_t.Math.Y_t.Math.Z_t}{.Math.}.Math.\frac{{\cos \left(+{}_0\right)}^n.Math.\mathrm{trackyield}.Math.\mathrm{passrate}.Math.\cos .Math..Math.}{r^2}$

[0193] This equation, combined with a predefined mask shape function, allows the deposited layer thickness to be simulated numerically.

[0194] FIG. 10 shows the simulated thickness distribution for a particular mask shape which results in the deposition of two regions varying linearly in thickness (between approximately 100 mm and 50 mm, and between approximately 50 mm and 100 mm, in position). The simulation was carried out using MathCAD software. FIG. 11 shows experimentally determined thickness measurements (using ellipsometry) from a real sample deposited using a mask having the same shape as that used in the simulations which generated the data shown in FIG. 10. The thickness simulation is shown to be highly accurate.

[0195] The effect of changing the mask shape on the deposited layer thickness distribution can be studied using this model. Accordingly, the mask shape can be optimised (e.g. numerically) in order to achieve a desired thickness variation.

[0196] The inventors have used this method to design a sputter mask for the deposition of a linear variable filter, where the layer thickness varies linearly along one dimension. FIG. 12 shows the superimposed experimental absorption spectra of a linear variable filter (deposited using the mask developed through this method) measured at four different locations along the length of the filter. Four distinct passbands are observed, having widths of approximately 30 nm at 700 nm wavelengths and 5 nm at 400 nm wavelengths. The peak position gradient is roughly 7 nm/mm, and the length of linear variable filter required to cover the entire visible spectrum is approximately 40 mm.

[0197] The sputtering mask 14, as shown in FIG. 13, used to deposit the linear variable filter has two substantially tapered apertures 15A and 15B through which two regions of linearly varying thickness may be deposited in use. A single mask may therefore be used to deposit two regions of linearly varying thickness on a single substrate.

[0198] FIG. 14 shows the superimposed experimental transmission spectra for an alternative linear variable filter deposited using a similar mask. Transmission was measured at seven different locations along the length of the filter. Seven distinct passbands are observed within the wavelength range of 450 nm to 900 nm. The bandwidth of each transmission band is 1.5% of the specified wavelength. The percentage of light transmitted at each particular wavelength peak is greater than 50%, although this varies with peak position.

[0199] FIG. 15 shows the measured spatial variation of the peak transmission wavelengths along the length of the filter. The variation is linear with a gradient of approximately 11.3 nm/mm.

[0200] Further variations and modifications may be made within the scope of the invention herein disclosed.