Travelling wave optical modulator

0001This invention relates to a travelling wave optical modulator and particularly, but not exclusively, to a modulator for broad-band operation. By optical is meant that part of the electro-magnetic spectrum which is generally known as the visible region together with those parts of the infra-red and ultra-violet regions at each end of the visible region which are capable of being transmitted by dielectric optical waveguides.

0002Guided-wave modulators which exploit the electro-optic effect are finding many applications in optical transmission and processing systems. Lithium niobate integrated optical devices are attractive, for example, as they require low drive voltages and can be fabricated by planar processes.

0003The electro-optic effect in semiconductor and dielectric waveguides is relatively weak, and so in order to operate efficiently at high modulation frequencies above one or two gigahertz a travelling wave structure, as opposed to a lumped electrode, is usually employed. That is, the electrode is designed as a transmission line matched to a microwave modulating input signal with the modulating signal fed colinearly with the propagating optical wave.

0004The bandwidth of a basic travelling wave device is limited by any mismatch between the velocities of the modulating signal propagating down the electrode and the optical signal propagating down the waveguide which causes "walk-off" of the two signals. If the velocities are equal the electrode length can be arbitrarily long allowing a low drive voltage to achieve the required phase change. However, some important materials such as lithium niobate have an inherent mismatch in these velocities so any attempt to reduce the drive voltage and power by increasing the length of the device will decrease the maximum possible drive frequency. One approach to overcoming this limitation, discussed in a paper entitled "Velocity-Matching Techniques for Integrated Optic Travelling Wave Switch/Modulators" by RC Alferness, S K Korotky and EAJ Marcatilli, IEEE Jul Qu Elec Vol QE-20 No. March 1984, is velocity matching by intermittent interaction. The electrode is split apart from the optical waveguide before or at a point where the velocity mismatch between the modulating signal and the optical signal would result in sufficient 'walk-off' of the signals to cause a polarity reversal of the applied field. The electrode is returned to rejoin the waveguide at a re-joining point, the electrode and waveguide path having lengths between the splitting and rejoining points designed so that the optical signal in the waveguide is advanced relative to the electrode signal sufficiently to once again be subject to induced phase shifts over the next section of coincident electrode and waveguide that are in phase with those induced in the previous section. This is repeated to obtain the total effect on the phase of the optical signal required.

0005A disadvantage of this known intermittent interaction, velocity-matching technique is that, although it provides a method of compensating the velocity-mismatch at a given predetermined, high frequency, it produces no substantial increase in bandwidth over that of a velocity-mismatch limited travelling-wave optical modulator and the accumulated phase shift generally displays a complicated dependence on the applied, modulating signal frequency. Further, it cannot operate with non-periodic modulating signals as the optical signal is modulated by successively advanced portions of the modulating signal as it progresses from one section to the next.

0006An alternative approach to providing intermittent action is to compensate for the walk-off by designing the optical and electrode paths such that the optical signal is retarded relative to the electrode signal to achieve the required phase matching for the subsequent interaction. Previously, implementations have relied on bulk crystals with parallel reflectors to reflect a freely propagating optical signal along a zig-zag path to pass intermittently within an interaction region formed by a shorter zig-zag or straight electrode; see US3,791,718 and US3,393,954 respectively. Such known intermittent interaction optical modulators are not, however, suited to forming fully integrated optical circuits, require careful preparation and handling of the reflective surfaces, and accurate alignment of the light input to ensure the optical signal propagates in the required interaction region.

0007According to the present invention a travelling wave optical modulator comprises a substrate, an optical waveguide defined within the substrate an optical path for an optical signal; a travelling wave electrode by which a modulated electric field of a predetermined frequency is applicable to at least two spaced portions of the optical waveguide; with one or more delay portions each being between a respective splitting point and rejoining point of the electrode and the waveguide, the optical waveguide being spatially separated from the electrode between said points, whereby the transit time between the splitting point and rejoining point is greater for an optical signal in the optical waveguide than for the electric field in the electrode.

0008With this arrangement of intermittent interaction the optical signal is delayed relative to the electrode signal to compensate for the accrued walk-off but, in contrast to known intermittent interaction devices, it is returned to rejoin the electrode so as to be modulated by the same electrode signal. The electrode signal therefore need not be periodic.

0009Embodiments of the invention will now be described, by way of example only, with reference to the accompanying drawing in which - <ul id="ul0001" list-style="none"><li>Figure 1 is a diagramatic plan view of a first embodiment of a travelling wave optical modulator according to the present invention having a proton exchanged optical waveguide;</li><li>Figure 2 is a diagramatic plan view of a second embodiment of the present invention having a proton exchanged optical waveguide;</li><li>Figure 3 is a diagramatic plan view of a third embodiment of the present invention in which the optical waveguide includes a reflective element near a substrate surface; and</li><li>Figure 4 is a diagramatic plan view of a fourth embodiment of the present invention in which the optical waveguide includes several reflective elements fabricated within the substrate.</li></ul>

0010Referring to Figure 1, a travelling wave optical modulator 2 comprises a Z - cut lithium niobate substrate 4 having at its top -ve Z face surface 6 a generally S-shaped proton-exchange optical waveguide 8 formed in a known manner. The waveguide terminates at ends 10 and 12 at polished end faces 14 and 16 respectively to permit ready connection to fibre optic waveguides (not shown) for the input of optical signals and the output of corresponding optical signals after modulation.

0011An earth plane 18 and hot electrode 20 of aluminium, fabricated by e-beam evaporation, photolithograph and plasma etching on top of a dielectric buffer layer (eg. Sio₂, 0.25µm thick) on the top surface 6 of the LiNbO₃ substrate 4 together form a travelling wave electrode which defines active modulating region 22 beneath the hot electrode 20. The corners of the hot electrode 20 are truncated on the outside of its right angle bends with the ratio of the length of the truncation edge to hot electrode width being approximately 1:3.

0012The optical waveguide 8 lies within the modulating region 22 at three equal length sections S₁, S₂ and S₃. Assuming an optical signal enters at end 10 it will be influenced by a modulating signal passing along the travelling wave electrode while in section S₁ until a first splitting point A₁. The optical waveguide is then apart from the active region 22 until it rejoins at a first rejoining point B₁. The above action is then repeated through the section S₂ and second splitting and rejoining points A₂ and B₂, respectively. Finally the optical signal will pass along section S₃ before a final splitting from the active region 22 to exit the device at waveguide end 12.

0013The lengths of the electrical waveguide and optical waveguide between the splitting and rejoining points A₁, and B₁ are such that the total transit times of the optical and electrical signals along the strips along their respective paths from Al to B1 are the same. The optical and electrical signals thereafter enter the strip S₂ in the same phase relationship as they entered strip S₁. Similarly the optical signal is delayed relative to the microwave signal between the second splitting and rejoining points A₂ and B₂ so they enter strip S₃ with the same initial phase relationship as they entered S₁ and S₂    The above referenced paper calculates that the walk-off between the optical and microwave signals after an interaction length ℓ is given by<maths id="math0001"><img file="EP0340043B1_D0001.tif" /></maths> where w = (l - No/Nm) No is the optical index; Nm is the microwave index; C is the speed of light; and f is the microwave frequency. In the device of Figure 1 the strips S₁, S₂ and S₃ are each of a length L such that<maths id="math0002"><img file="EP0340043B1_D0002.tif" /></maths> for an arbitrary design frequency, f_d so the optical and microwave are split apart at the splitting points just as the walk-off results in a polarity reversal of the applied field. The optical waveguide lengths between A₁ and B₁ is L(l-No/Nm) longer than the travelling waveguide between these points.

0014The particular advantage of using proton exchanged optical waveguides is that a radius of curvature is sufficiently small to allow the optical waveguide to be removed from and rejoined with the hot electrode within the boundaries of the substrate without the requirement for reflecting surfaces as in, for example, the third and fourth embodiments.

0015Other geometric arrangements of the waveguides are possible, as for example as is shown in Figure 2 showing a section of a travelling wave optical modulator having a straight hot electrode 19 and earth plane 21, the former overlying portions of a proton exchanged optical waveguide 23 on the surface of a Z-cut LiNbO₃ wafer 25. The principle of operation is as for the device shown in figure 1.

0016Referring now to Figure 3 there is shown a travelling waveguide optical modulator fabricated on the top surface (- Z face) 32 of a Z - cut lithium niobate substrate 34. A travelling wave electrode with an active region 36 is defined by an earth plane electrode 38 and hot electrode 40 as in the device of Figure 1. A titanium indiffused optical waveguide 42 formed in the surface 32 consist of two straight portions 44 and 46 optically coupled by a reactively etched mirror surface 48. The waveguide arms 44 and 46 terminate at ends 52 and 54 at polished faces 56 and 58, respectively, which faces are perpendicular to the waveguide arms 44 and 46. The hot electrode has wider portions 47 away from portions S₃ and S₄ to reduce microwave loss. The distance between the earth plane and hot electrode should be adjusted in a known manner as the hot electrode is changed in size to maintain a constant characteristic impedence.

0017A waveguide for operation at wavelengths from about 0.4µm to 4µm can be manufactured as follows. Titanium is deposited on the - Z face by e - beam evaporation to a thickness of between 40 and 160nm and a width of 4µm to 16µm. Diffusion is carried out at a diffussion temperature of 1050°C with O₂ flowing at a rate of 31/minute through a hot (56°C) water bath and bubbler for approximately 7 ¹/2 hrs. A silica film is then deposited by c.v.d. on the substrate to a thickness of about 0.25µm. The device is then polished to form the mirrors and end faces, 50, 56 and 58.

0018Aluminium 5µm thick is deposited and an electrode pattern formed using TF20 photoresist which is then plasma etched using CCl₄ etchant gas.

0019The waveguide 42 lies within the active modulating region 36 at two equal length sections S₃ and S₄ each of length L₂. In a similar fashion to the device of Figure 1, an optical signal enters the waveguide 42 at end 54 and travels through the active region S₃. It then splits away from the active region at the splitting point A₃ to be reflected by the surface 48 to the rejoining point A₄ from where it passes through the active region S₄ before exiting from end 52.

0020The analysis of this device is as for the one shown in Figure 1, namely that the optical signal is delayed relative to the microwave signal between A₃ and A₄ to compensate for the walk-off during passage along S₃. In this case for Nm about 4.3 and N_o about 2.2, the contained angle between arms 44 and 46 is 29°, L₂ 4mm, A₃ - A₄ for the optical waveguide = 8 mm and A₃ - A₄ for the travelling wave electrode is 2mm. The length L₂ is determined by f_d which in this instance is approximately 20 GH_z.

0021The rephasing will occur in the general case when <maths id="math0003"><math display="inline"><mrow><msub><mrow><mtext>([A₃ to A₄]</mtext></mrow><mrow><mtext>opt</mtext></mrow></msub><msub><mrow><mtext>+L₂)/V</mtext></mrow><mrow><mtext>o</mtext></mrow></msub><msub><mrow><mtext>=([A₃ to A₄]</mtext></mrow><mrow><mtext>elec</mtext></mrow></msub><mtext>+L₂)/Vµ</mtext></mrow></math><img file="EP0340043B1_D0003.tif" /></maths> In LiNbO₃ Vo/Vµ approximately 2 so replacing occurs in the Figure 2 embodiment if - <maths id="math0004"><math display="inline"><mrow><msub><mrow><mtext>[A₃ to A₄]</mtext></mrow><mrow><mtext>opt</mtext></mrow></msub><msub><mrow><mtext>-L₂/Vo =[A₃ to A₄]</mtext></mrow><mrow><mtext>elec</mtext></mrow></msub><mtext>/Vµ</mtext></mrow></math><img file="EP0340043B1_D0004.tif" /></maths> eg. <maths id="math0005"><math display="inline"><mrow><mtext>L₂ = 6mm</mtext></mrow></math><img file="EP0340043B1_D0005.tif" /></maths> , <maths id="math0006"><math display="inline"><mrow><msub><mrow><mtext>[A₃ to A₄]</mtext></mrow><mrow><mtext>opt</mtext></mrow></msub><mtext>= 10mm,</mtext></mrow></math><img file="EP0340043B1_D0006.tif" /></maths><maths id="math0007"><math display="inline"><mrow><msub><mrow><mtext>[A₃ to A₄]</mtext></mrow><mrow><mtext>elec</mtext></mrow></msub><mtext>= 2mm.</mtext></mrow></math><img file="EP0340043B1_D0007.tif" /></maths>

0022Referring now to Fig 4, a travelling wave optical modulator utilises reflective surfaces within, the volume of the substrate to define the optical waveguide delays. The modulator 60 is formed on a Z - cut lithium niobate substrate 62 from a Ti indiffused optical waveguide 64 having sections S₅, S₆ and S₇ within a linear active modulating region 66 defined by an earth plane electrode 68 and a hot electrode 70 with ends 72 and 74 at polished substrate surface 76. An optical signal entering the waveguide 64 at end 72 is reflected round the delay loops 76 and 78 between active sections S₅, S₆ and S₇ by reactive etched mirrored surfaces of which only one 80 is numbered for clarity.

0023The lengths of the active regions and electrode and waveguide paths between active regions are as in Figure 2 so the same analysis applies as to that embodiment. This device however has three active regions compared to the two of Figure 2.

0024The mirrors 80 and 48 can be formed by other methods for example ion beam milling.

0025The structure of Figure 3 can usefully be employed on x or y - cut LiNbO₃ with the appropriate lateral shift of the electrodes to define a field active region in the sections S₅, S₆ and S₇. To avoid leakey modes the waveguides are fabricated approximate parallel to the crystal Y and Z or X and Z axis. (± 5°).

0026The active region and optical delay sections can be repeated as often as desired as the relative positions of the microwave and optical signals is reset for the start of each successive active region S_i.

0027Devices can be formed according to the present invention in which the active region is not sufficiently long to cause full walk-off of the electrical and optical signals but some fraction thereof. In such cases the delay to be imposed on the optical signal will need to be smaller to compensate for the reduced walk-off.

0028Substrates other than LiNbO₃ can be employed in the production of the present invention for example LiTaO₃.

0029The invention finds application in many areas where electro-optic guidedwave devices are used e.g. phase modulators, Mach Zehnders and directional couplers.