BACKGROUND OF THE INVENTION

The invention relates to a digitally controlled light intensity modulator, constructed with integrated optics.

It is known to construct light modulators controlled by an analog voltage, from a two-arm interferometer in which there is created, by an electro-optical effect, a difference in the propagation speed between the two arms, producing an electrically controlled phase shift. The emerging light intensity varies with the phase shift and is therefore modulated by the analog voltage inducing the electro-optical effect. The response of such a modulator is sinusoidal but may be linearized by adding an optoelectronic feedback loop. This improvement is described in French patent application No. 79.08 372 filed by the applicant.

SUMMARY OF THE INVENTION

The invention aims at using the same optical structure to construct a multiple-control signal modulator, these signals being moreover digital, i.e. having two voltage states. A digitally controlled modulator finds in fact numerous applications in the telecommunication and data processing fields. For suh a modulator to be compatible with all existing electronic circuits, the control voltages must be low, preferably of the same order of size as the voltages used in TTL logic (0-5 volts).

By choosing the position of the electrodes and the crystalline orientation of the substrate in which the interferometer is constructed, it can be arranged for the electric field created between the electrodes to act on the speeds of propagation of the waves only in the portions of the arms situated opposite these electrodes. By juxtaposing along the interferometer several sets of electrodes to which control voltages are respectively applied, the corresponding positions of the interferometer may be acted on independently, with negligible interaction between the different sets of electrodes. To obtain digital control, the light intensity response must depend on the respective weights of the different control bits. For that, the invention uses either sets of electrodes of different lengths, in a geometrical progression in conformity with the successive powers of 2, or sets of electrodes of the same length, the control bits being applied to voltage dividers whose voltage ratios conform to the successive powers of 2. These two processes enable successive phase shifts to be obtained whose values, when the control bits are at state 0, conform to the successive powers of 2. Furthermore, it is possible to obtain adjustment of the zero (the output intensity is minimum when all the bits are at 0) by means of polarizing electrodes to which an adjustable voltage is applied.

DESCRIPTION OF THE DRAWINGS

The invention will be better understood from the following description and accompanying drawings.

FIG. 1 shows a two-arm interferometer of a known type.

FIG. 2 shows the response curve of the interferometer of FIG. 1.

FIG. 3 is a diagram of the modulator in accordance with the invention.

FIG. 4 shows the response curve of the modulator of FIG. 3.

FIG. 5 shows one embodiment of the invention.

FIGS. 6, 7 and 8 are diagrams of different embodiments of the invention.

FIG. 9 is a diagram of a digital-analog converter in accordance with the invention.

DESCRIPTION OF THE PREFERRED EMBODIMENTS.

FIG. 1 shows a two-arm interferometer of a known type able to serve to modulate light intensity. This interferometer comprises a monomode wave-guide branch 1 into which is coupled a light wave of intensity I_E. This branch 1 forks at a fork B₁ into two parallel arms 2 and 3 forming monomode guides which join up again at a fork B₂ to extend into monomode wave-guide branch 4 where an intensity I_S may be recuperated. The directions of propagation of the light are the same in the four guides. Electrodes are disposed on each side of each of guides 2 and 3, for example a central electrode E_M, an electrode E₂ on the other side of arm 2 and to which a voltage V₂ is applied and an electrode E₃ on the other side of arm 3 and to which a voltage V₃ is applied, electrode E_M being connected to ground. The arrangement of the electrodes chosen for the figure, i.e. not overlapping the guides but being adjacent thereto, causes in each of guides 2 and 3 electric fields perpendicular to the direction of propagation of the light waves and in the plane of the figure. In this case, the mode of propagation used is a TE mode. The whole of the device is constructed at the surface of a substrate having electro-optical properties, such as lithium niobate (LiNbO₃), lithium tantalate (LiTaO₃), gallium arsenide (GaAs) . . . . It is for example possible to obtain guides by diffusing titanium in a substrate by conventional photolithography processes. In the case of the figure where the electric fields applied to the guides are in the plane of the figure, it is preferable for the axis c of the lithium niobate to be also in this plane and perpendicular to the direction of propagation common to guides 1 to 4, so as to present the maximum electro-optical effect. If voltages V₂ and V₃ are of the same sign, the electric fields applied to arms 2 and 3 are in opposite directions, these fields cause variations in the speeds of propagation of the light waves which are opposite in sign for each of the two arms. The difference Δβ between the speeds of propagation in the two arms is proportional to the sum of the amplitudes of the electric fields applied, i.e. practically proportional to the sum of the voltages V. There is obtained, at the position of fork B₂, a phase shift Δφ between the two waves from arms 2 and 3 which were originally in phase: Δφ=Δβ×L where L is the length of action of the electric fields, i.e., in the case of the figure, the length common to the electrodes.

The intensity I_S recuperated in guide 4 varies as a function of phase shift Δφ according to a sinusoidal law whose variation curve is shown in FIG. 2. The intensity I₃ varies between a minimum I_m and a maximum I_M, the modulation rate I_M /I_m depending on the coupling losses and on the dissymmetry of the device. When voltage V results from an analog modulation, intensity I_S follows the modulation, provided that the phase shift remains between 2 n π and (2n+1)π, so as to keep a constant curve slope. This modulation is not linear, except in the vicinity of inflection points (Δφ=(2n+1)π/2).

The properties described above of the two-arm interferometer are used in the device of the invention of which the operating diagram is shown in FIG. 3. It is a multiple-control voltage modulator. The modulator of FIG. 3 is digitally controlled: the control voltages are binary signals having two states: state 0 where the voltage is zero and state 1 where the voltage is at a predetermined value v common to all the signals. In the device of the invention, n sets of electrodes are used, n being the number of control voltage inputs. In FIG. 3, n=3. Between the two arms 2 and 3 of the interferometer, there is disposed a central electrode E_M connected to ground. On each side of the arms are disposed symmetric pairs of electrodes: E₂₀ and E₃₀, E₂₁ and E₃₁, E₂₂ and E₃₂. The electrodes of each pair are connected together and are subjected to binary signals: respectively v₀, v₁, v₂. The signals form a control word from a modulation source. Conventionally, v₀ is the bit with the smallest weight and v₂ the bit with the largest weight. To obtain modulation of the output intensity I_S which takes into account these different weights, the pairs of electrodes have lengths in geometric progression: the length of electrodes E₂₀ and E₃₀ being equal to l₀, the length of electrodes E_2n and E_3n receiving a signal v_n is equal to l_n =2ⁿ l₀ (thus l₁ =2l₀ and l₂ =4l₀). Since the phase shift produced by the electro-optical effect between the waves circulating in arms 2 and 3 is, for an equal control voltage, proportional to the length of action, it can be seen that the effect of the different bits on the phase shift depends directly on their weight. Furthermore, the successive phase shifts introduced by the successive voltages v₀, v₁, . . . cumulate and the total phase shift obtained at the position of fork B₂ is representative of the analog value corresponding to the word formed by the different input bits.

In the embodiment shown, it is arranged for the mode of action of the input voltages to be such that the total phase shift Δφ is equal to π when all the bits are at state 0 and equal to 2π when all the bits are at state 1. Thus, when the analog value corresponding to the input word runs through the whole range of possible values, the output intensity I_S increases uniformly from its minimum value I_m to its maximum value I_M. The different values of I_S corresponding to the different possible words are obtained from the curve of FIG. 4 which originates from that of FIG. 2. As abscissa are shown the cumulated phase shift Δφ and the voltage values v₀, v₁, v₂ which lead thereto. As ordinates are shown the corresponding values of the output intensity. To satisfy the first condition: Δφ=π when all the control voltages are zero, it is necessary to introduce this phase shift by π between the two arms. In fact, in the absence of any voltage, the output intensity is maximum. To this end, an additional polarization electrode E_P is provided subjected to a constant voltage V_P and of length l_P, so that k l_P V_P =π. This electrode E_P may be situated, for example, on the same side as electrodes E₂₀, E₂₁, E₂₂, facing a portion of electrode E_M. To satisfy the second condition: Δφ=2π when all the control bits are in state 1, the values of l₀ and v must be suitably chosen. The total phase shift Δφ holds good in the general case with n electrodes:

Δφ=k[lP VP +2l0 v0 +2l1 v1 + . . . +2ln-1 Sn-1 ]

with:

klP VP =π,v0 =v1 =v2 =v,l1 =2l0 and l2 =4l0,

we obtain

Δφ=π+2kl0 v(1+2+ . . . +2n-1)=π+2kl0 v(2n -1).

If v is for example fixed by the electronic modulation means, the value of l₀ is thus determined allowing Δφ=2π to be obtained.

FIG. 5 shows a first embodiment of the invention. The modulator is formed from a substrate 5 made from lithium niobate. The guides 1, 2, 3, 4 are formed by photolithography by localized diffusion of titanium on one of the faces Σ of the substrate. The orientation of axis c of the substrate has been chosen perpendicular to the surface Σ. The total length of arms 2 and 3 is chosen so as to be greater than the sum l₀ +l₁ + . . . l_n-1. The light energy coupled to guide 1 may for example come from an optical fiber 11, the coupling being effected directly on the edge of the substrate. Similarly, the emerging wave may be recuperated by an optical fiber 12 at the other end of the substrate. Because of the orientation chosen for the lithium niobate, the electrodes are disposed so that the electric fields created in guides 2 and 3 are parallel to the axis c so perpendicular to surface Σ. In this case, the propagation mode used in the guides is a TM mode. For that, it is necessary for one of the electrodes of each pair E_2i -E_M and E_3i -E_M to overlap respectively guides 2 and 3. It has been chosen for E_M to overlap guide 2 and E₃₀, . . . , E_3n-1 to overlap guide 3.

Thus, the voltages applied to the pairs of electrodes E_2i E_3i with respect to electrode E_M connected to ground cause in guides 2 and 3 electric fields in opposite directions whose effects on the phase shift are added together. The binary signals v₀, . . . v_n-1 originate in a modulating source 6. The polarizing voltage V_P comes from an adjustable DC voltage source 7. Electrode E_P is placed on the slanting part of electrode 2, close to fork B₁.

The device shown was constructed for a 6-bit control with guides having a length of 2 microns, a distance between arms of 60 microns and a fork angle of 1°. Thus, the direction of propagation in the slanting part is practically the same as in guides 1, 2, 3, 4.

A length l₀ =0.3 mm was chosen, so that the total length of action of the control electrodes is (2⁶ -1) l₀ =18.9 mm. It has been possible to measure that, when all the control bits are in state 1, by choosing v=1 V, a phase shift range is obtained equal to π, which is the result sought. To have a polarizing voltage of the order of 5 V, the length l_P must be of the order of ##EQU1## By finely adjusting the value of V_P, when all the bits are in state 0, a very low minimum value of I_m is obtained, since the modulation rate measured I_M /I_m is of the order of 95%. The voltage values (v=1 V, V_P =5 V) are compatible with electronic circuits constructed according to the TTL technology.

Instead of seeking a phase shift range equal to π, which leads with the device of FIG. 4 to a non-linear distribution of the emerging intensity, this range may be reduced so as to use the part of the variation curve situated on each side of the inflection point over a more reduced range. If it is desired to keep the same values for v and V_P, with the same characteristics of the device; length l_P must be increased and the value of l₀ decreased. For example, to divide the range of Δφ (instead of running through the range π-2π, the range 5π/4-7π/4 is run through and the range of I is divided by √2), it must be arranged for:

klP VP =5π/4.

As in the preceding case, we had k l_P V_P =π with l_P =4 mm, k and V_P being kept, we have l_P =5 mm. Furthermore it is necessary for 2 k l₀ v (2ⁿ -1)=π/2. As in the preceding case, we had 2 k l₀ v (2ⁿ -1)=π with l₀ =0.3 mm, k, v and n being kept, we have l₀ =0.15 mm. This mode of modulation presents the particularity that the emerging intensity is not zero when all the input bits are in state 0.

There exists another means, preferable to the preceding one, for obtaining practically linear distribution of the emerging intensity, this time over practically the whole of the phase shift range π-2π. This means, described in itself in French patent application filed by the applicant on Apr. 3, 1979, under the number 79.08 372 which relates to a modulator having a single modulating voltage input, consists in using a feedback loop so as to superimpose on the modulating voltage a voltage proportional to the emerging intensity. The use of this process in a digitally controlled modulator is illustrated, in two possible variations, by the diagrams of FIGS. 6 and 7.

According to the variation of FIG. 6, the control bits and the polarizing voltage are applied between the grounded electrode E_M and a series of electrodes disposed on one side of an interferometer: E_P, E₂₀, E₂₁ . . . whose lengths increase in accordance with the powers of 2. The radiation emerging from the interferometer is intercepted by a separator means 8, which may be formed in different ways described in the abovementioned patent application. A fraction of the radiation is applied to a photodetector 9 which is also described in the same patent application, which supplies a feedback voltage V_R. This is applied between electrode E_M and a second set of electrodes E₃₀, E₃₁, . . . having respectively the same length as electrodes E₂₀, E₂₁ . . . .

The effects of electrodes E_2i and E_3i on the speed of propagation of light in each branch of the interferometer are thus proportional to the voltages which are respectively applied thereto, with the same coefficient of proportionality.

Instead of separately applying voltages v₀, v₁ . . . on the one hand and voltage V_R, on the other hand, to different sets of electrodes, they may be applied, as shown in the diagram of FIG. 7, to the same set, for example E₃₀, E₃₁ . . . through summing circuits 50, 51, . . . having two inputs one of which receives voltage V_R and the other bits v₀, v₁, . . . . The set of electrodes E₂₀, E₂₁, . . . being omitted, they are replaced by a single polarizing electrode V_P.

Many other embodiments including a feedback loop or not are possible, by applying the control bits and possibly the feedback voltage to one or two sets of electrodes, which have lengths increasing in conformity with the powers of 2. When a single set of electrodes is used for the input of the control bits, the products l_i v_i must be multiplied by 2 with respect to the case where the control bits are applied to two sets of electrodes and acting therefore on both arms of the interferometer.

To modulate the light intensity with control words formed of n bits, it is not indispensable to use electrodes of unequal length. FIG. 8 illustrates an embodiment in which all the control electrodes have the same length. Since the phase shift between the waves propagated in the two arms is proportional to the voltage x electrode length product, it is necessary, to achieve the different weights, to effect the weighting on the voltages instead of the lengths. This is why the modulator shown in FIG. 8 comprises, besides the interferometer associated with the polarizing electrode E_P receiving voltage V_P and electrodes E₂₀, E₂₁, . . . connected respectively to electrodes E₃₀, E₃₁, . . . voltage dividers 60, 61, 62, 63 receiving respectively the voltage bits 0 and v and delivering respectively voltages V₀, V₁, V₂, V₃ with ##EQU2## where r is a ratio of division chosen as a function of v and of the value of the lengths of the control electrodes. The operation of the modulator of FIG. 8 is identical to that of FIG. 3. A digitally controlled modulator may be constructed with electrodes of the same length by using all the above-described variations, and particularly by associating therewith a feedback loop for linearizing its response.

The modulator of the invention is particularly applicable to optical transmission systems by means of optical fibers and to the optical processing of data. The low control voltages which it requires, its large passband and its reduced dimensions make it advantageous with respect to the devices already known. One interesting application of this modulator is digital-analog conversion. In effect, since, by suitably choosing the parameters, an increasing emerging intensity may be obtained with a control word which can be associated with an increasing analog value, it is possible to use this property to recuperate an analog voltage whose value will be practically proportional to this analog value. One of the possible means consists in placing a photodetector on the output guide 4 of the interferometer. A diagram of a converter is shown in FIG. 9. It is strongly recommended to use a feedback loop for ensuring linear viariation of the intensity so good quality conversion. At the output of a photodetector 100, a voltage V_a is recuperated corresponding to the analog value of the input word.

It is apparent that within the scope of the invention, modifications and different arrangements can be made other than are here disclosed. The present disclosure is merely illustrative with the invention comprehending all variations thereof.