1 Introduction

Optical beam deflectors are used in various applications such as laser printing, writing, and readout of optical data storage devices, data processing in optical telecommunications and laser pulse selection for power amplifiers and material processing [1]. Different technologies have been proposed and demonstrated to realize laser beam deflectors, such as galvanic mirrors and MEMS based deflectors [2], acousto-optic deflectors [3], and EO deflectors [4–6]. Among these technologies, the MEMS-based optical deflectors with moving parts are relatively slow and subject to wearing and eventual damage. The acousto-optical deflectors usually suffer from some limits such as low response speed, small aperture, and low resolution. EO deflectors can provide high speed beam scanning due to motionless components, so they offer a good candidate for practical applications [7–9].

For applications in laser radar for long-range applications, a high deflection speed is often demanded and high power laser beams are needed, therefore, beam scanning devices should be designed with apertures sufficiently large in order to operate below the threshold of optical damage. For over several decades, EO deflectors based on crystal materials have been extensively studied as alternative techniques for their advantages in response speed and free of wear-tear concerns. Among these crystals, LiNbO3 has some distinct advantages such as high response speed, availability of large volume materials, high damage threshold, and maturity in its growth technology as well as applications [10–12].

In this paper, we demonstrate a large aperture 2D EO digital beam deflector based on LiNbO3 crystal switch array. In the proposed deflector, multistage basic beam splitting units and a deflection prism array are cascaded to produce a 2D array of beam positions. Each basic beam splitting unit consists of an EO switch and a PBS. Design, construction details, and performance of the 3×3 beam array deflector with a aperture of 6.5 mm are experimentally investigated. The performances of the deflector including the transmission efficiency, scanning speed, half-wave voltage at 1.55 µm wavelength are measured. Analysis of the effects of propagation direction deviation from the optical axis is carried out and a remedial technique is proposed to improve the performance.

2 Principle

The schematic diagram of the proposed EO deflector is shown in Fig. 1. The deflector is composed of multistage cascaded basic beam splitting units and a prism array. The basic, controllable, binary beam splitting unit consists of an EO switch and a polarizing beam splitter (PBS). The EO switch element is used to rotate the direction of polarization of the light beam by 90°, and the subsequent change in the propagation direction of the beam is produced by a PBS. Each basic beam splitting unit can produce two resolved spots. The direction or position of the output beam can be controlled by applying two different voltages to the EO switch, as shown in Fig. 2. The electric field is applied in the x direction via deposited electrodes on the yz faces of the crystal bar. The LiNbO3 crystal is a z-cut bar, so the optical axis c is parallel to the direction of beam propagation. This simplifies the design of the optical system as good thermal stability is ensured automatically. The linearly polarized input beam can directly propagate when no electric field is applied. When the half-wave voltage V π is applied, thanks to the polarization modulation at the EO switch, the polarization direction of the output beam would be rotated by 90° with respect to the input beam and the propagation direction is also changed by the especial PBS.

Fig. 1
figure 1

Schematic diagram of the proposed 2D electro-optic deflector

Full size image
Fig. 2
figure 2

A basic beam-splitting unit consists of an EO switch and a PBS

Full size image

More than two output directions of the beam can be achieved by arranging two or more binary deflection units in series. A total of m-stages can produce 2m resolved spots, where the mth stage includes 2m−1 basic units. A two-dimensional N×N array deflection can be achieved by proper arrangement of the PBSs and application of half-wave voltages to EO switches. The scanning route and switch time can be controlled by a computer.

The scheme of beam steering involves the use of a prism array composed of N×N prisms to produce widely separated spots in the far field. The angular coverage of the far field is shown in Fig. 3. A small wedged prism is placed on the path of each resolvable beam and refracts the beam to an appointed angular position. The wedged prism has a normal incidence and an inclined output. The parameters satisfy the following relation:

$$ \theta\approx(n - 1) \cdot\alpha $$
(1)

where α is the wedge angle, n denotes the refractive index of the prism material, and θ is the deviation angle. The wedge angle α depends on the need of practical application.

Fig. 3
figure 3

The coverage of field of view with N×N array spots in the far field

Full size image

3 Experiment and results

The proposed two-dimensional EO deflector is experimentally demonstrated. The experimental setup is shown in Fig. 4. The proposed system can produce 3×3 rectangle array spots scanning with a field of view of 1.5∘×1.5∘. The laser used in the experiment was a 1.55 µm fiber-coupled semi-conductor diode CW laser. The 10 mW laser beam was first collimated into a beam with a waist diameter of 2.87 mm and then was changed into linearly polarized beam by a λ/2 wave plate. The beam coming out of the deflector is then split into two separate parts; one is reflected into an opto-detector, where the scanning speed of the deflector can be measured by a digital oscillograph; the other is displayed on a ground glass screen (GGS). The images of the deflected beams on the GGS were recorded and imaged simultaneously by an infrared CCD camera attached with a prefix lens with a focal length f=16 mm. The objective distance between the GGS and the lens of the CCD camera is 200 mm.

Fig. 4
figure 4

The experimental setup of the 2D digital deflector; CL: collimated lens, WP: λ/2 wave plate, D: deflector, BS: beam splitter, OD: opto-detector, DO: digital oscillograph, GGS: ground glass screen, CCD: charge-coupled device

Full size image

The deflector consists of eight beam splitting units and a prism array with nine small prisms, which is shown in Fig. 5 in detail. The 3×3 rectangle beam scanning array can be achieved by proper arrangement of the eight beam splitting units. Real-time control of the scanning route and time is realized by real-time control of the switch states of S1, S2, S3, S4, S5, S6, S7, and S8, which is programmable. The total size of the 3×3 prism array is 19.5 mm×19.5 mm. The wedge angles of four prisms corresponding to beams B1, B3, B5 and B7 in Fig. 7a and 7b are α≃2.05∘. The wedge angles corresponding to beams B2, B4, B6 and B8 are α≃1.45∘. The wedge angle of the central prism corresponding to beam B9 is α=0∘.

Fig. 5
figure 5

Schematic of the 3×3 rectangle beam array digital deflector; S1∼S8, LiNbO3 EO switches, PBS1∼PBS8, polarizing beam splitting

Full size image

In experiment, the half-wave voltage V π of LiNbO3 EO switch at 1.55 µm wavelength is measured first. V π can be approximately calculated from a theoretical formula [13],

$$ V_{\pi} = \frac{\lambda}{2n_{o}^{3}r_{22}} \cdot\frac{d}{l}$$
(2)

where r 22 and n o are the electro-optic coefficient and refractive index of LiNbO3 crystal, λ is the wavelength, l is the crystal length in the optical axis direction, and d is the aperture size at the incidence plane. These characteristic parameters r 22 and n o vary with λ and no direct values at 1.55 µm are available in reference. The half-wave voltage of LiNbO3 at this wavelength is measured to be V π ≈1950 V, as can be deduced from Fig. 6. The experimental results described should prove valuable for the design and optimization of EO devices and systems based on LiNbO3 crystal.

Fig. 6
figure 6

The half-wave voltage measurement results of LiNbO3 EO switch

Full size image

Then, for the case of nine scanning positions of 2D array, the deflected beams are displayed on a ground glass screen (GGS) and imaged the image plane of the lens attached with the infrared CCD camera. At the same time, the beam positions are recorded by the CCD when a particular scanning route is applied. And the transmission efficiency of the deflector is also measured. When a group of half-voltages are applied on the switches controlled by the computer, one of the nine beams can be deflected at a given time. Figure 7b gives the photographs of 3×3 array scanning spots along the designed route from B1 to B9 as shown in Fig. 7a. The scanning route and scanning time of the deflector can be programmable based on the need of experiment. The measured transmission efficiencies of B1∼B9 beams are respectively 79.6 %, 73.3 %, 69.3 %, 72.0 %, 67.8 %, 71.4 %, 66.3 %, 60.7 %, and 62.7 %. Experimental results show that the deflector can exactly scan at the pre-designed scanning route and duration time. But we also can see that the measured transmission efficiencies of the nine beams are different. This is because the nine beams pass through the deflector in different paths. The whole system of the deflector consists of multiple cascaded basic beam-splitting units. Beams B1–B7 have passed through three stages while B8 and B9 four. The optical axis of EO crystal in the first stage does not consist exactly with that of the second, third, and final stages. The transmission efficiencies are influenced by the deviation of the propagation direction from the optical axis owing to the machining errors of optical and mechanical elements. The machining errors can be diminished by optimizing design of the optical and mechanical elements. Energy losses at each interface can be reduced by antireflection coating.

Fig. 7a
figure 7

The scanning route of the deflector from B1 to B9

Full size image

To measure the scanning speed and test the deflection accuracy of the deflector, in our experimental setup an infrared photoelectric detector is placed behind one of the nine output beam to monitor its power and duration time at certain position. The measured signal is displayed on a digital oscilloscope. In our design, the duration time of each deflected beam is 30 ms and the interval time between two continuous deflected beams is 0 ms. The experimental result from a digital oscilloscope is clearly shown in Fig. 8. As can be seen from Fig. 8, a full scan of nine beams (one period) is 270.01 ms, and the duty cycle is 11.1 %, corresponding to the ratio of the duration time of each beam to the full scan time of one period is 1:9. So the measured scanning speed of the deflector is in good agreement with that of our design. The scanning time can be greatly shortened in practical applications owing to the high switch speed (1010/s) of LiNbO3 EO switch. From all the experimental results mentioned above, it can be deduced safely that the designed parameter of beam deflector agrees well with the measured results.

4 Errors analysis and discussion

The experiments are performed for the proposed 3×3 array EO deflector. But the transmission efficiencies of the nine beam spots are not high enough. A key reason of this result is the deviation of the propagation direction from the optical axis owing to the machining errors of optical and mechanical elements. In our experiment setup, the deflector consists of three or four stages of cascaded basic beam-splitting units. The optical axis of EO crystal in the first stage is not exactly consistent with that of the second, third, and final stages. This will lead to the rotation of the polarization direction and the change of refraction index with electrical field for a small angular variation between the wave normal and the optic axis in LiNbO3 crystal [14, 15].

As shown in Fig. 9, z-axis is optic axis, z′′ is the direction of wave normal, the ellipse with diagonals is the section perpendicular to the wave normal, in which the two directions of principal axes (x′′′, y′′′) a re the two polarization directions of the wave in crystals. Their indexes are n x′′′ and n y′′′. x′ and y′ are the electrically induced principal axes. According to the equation of index ellipsoid, the birefringence optic path difference and the rotation of polarization direction can be calculated for a small angular variation and the applied electric field E. The birefringence optic path difference:

(3)

And the rotation of polarization direction is given by

$$ \tan2\beta= \frac{2\gamma E\sin2\phi}{2\gamma E\cos2\phi-\varDelta \sin^2\theta}$$
(4)

where \(\varDelta = 1 / n_{e}^{2} - 1 / n_{o}^{2}\), n o and n e are the o- and e-refraction index in LiNbO3 crystals, respectively, l is the optical path length, γ is the electro-optic coefficient, θ is the deviation angle between the propagation direction and the optic axis, ϕ is the displacement angle between the wave normal and the induced principal axis x′, β is the rotation angle of polarization direction. For LiNbO3 crystals, the electrical field direction is in the x–y plane, γ=γ 22. If the beam propagates along the optic axis, θ=0, ϕ=0, then Eq. (3) can be simplified as \(\varGamma = n_{o}^{3}l\gamma_{22}El\).

To test the performance of the proposed deflector when the light deviates from the optical axis, an EO modulation experiment is carried out and the extinction ratio of the deflector is measured at different deviation angle. In experiment, the proposed deflector is fixed. When the angle of incidence of laser beam to the input of the deflector is changed, the extinction ratio is also changed. In order to show this clearly, we define the normalized extinction ratio ER as

$$ \mathit{ER} = \frac{\mathit{ER}_{a}}{\mathit{ER}_{0}},$$
(5)

where ER 0 and ER a are the measured extinction ratio at θ=0 and θ>0, respectively.

The extinct ratio of beam B1 in Figs. 7a, 7b is measured. When θ=0, the measured maximum extinct ratio is approximately 181, that is ER 0=181. Figure 10 shows the changes of the normalized extinction ratio ER with the deviation angle θ. As can be seen, the value of ER decreases rapidly with the increasing of θ. When θ is in the neighborhood of 0.1°, ER is approximately 50 %. So the effect of light propagating away from the optic axis has great influence on the performance of the proposed deflector. Therefore, some methods should be adopted to diminish the corresponding negative effect in practical applications.

Fig. 7b
figure 8

The beam positions recorded by an infrared CCD camera when the scanning route in Fig. 7a is applied

Full size image
Fig. 8
figure 9

Scanning speed and during time measuring of one beam position by oscilloscope tracing

Full size image
Fig. 9
figure 10

Index ellipsoid of LiNbO3 crystal for an electrical field applied

Full size image
Fig. 10
figure 11

The off-axis extinction ratio measurement results of the EO deflector

Full size image

Two effective methods are proposed in this section to diminish the influence of the off-axis propagation. First, the mechanical design supporting the cascaded basic splitting units should be optimized so as to avoid deviations from the direction of optic axis c. An improved design with full consideration to the off-axis problem is proposed in our final system which will certainly have much higher transmission efficiency. Second, new compact configurations of the splitting unit with less transmission and reflection interfaces should be adopted. For example, the PBS can be replaced by a crystal with high birefringence such as calcite or YVO4 crystal. A suggested schematic of a 1×4 beam deflector system is shown in Fig. 11. The basic beam splitting unit consists of an EO switch and a birefringent crystal.

Fig. 11
figure 12

A compact beam deflector based on EO switch and birefringent crystal

Full size image

5 Conclusion

A large aperture 2D EO deflector using multistage cascaded beam splitting units is proposed. A 3×3 beam deflector is fabricated and experimentally investigated. The scanning route and duration time of the deflector can be controlled by programming the voltages applied on the EO switches. The detailed performance of the proposed system is presented. Influence of the propagation direction deviation from the optical axis to the transmission efficiency is experimentally investigated and two measures are proposed to diminish this effect. Advantages such as large aperture, high response speed, and non-mechanical scanning make this device quite suitable for applications where fast scanning of high power laser beam is needed. A compact configuration with optimized mechanical design and improved beam splitting unit structure and, therefore, higher transmission efficiency is in development.