Broad-area laser diodes are the most efficient coherent light sources and are widely used today. The extraordinary efficiency, modulation easiness, availability to virtually any wavelength and compactness are the principle drives stimulating the development of light sources based on laser diodes. However, due to fundamental limitations of laser diode gain medium, the emission from laser diodes has a major drawback — the emission is not spatially coherent. In other words laser diode light is often seen as a light bulb emission that cannot be focused in a diffraction limited spot of λ/2 or easily transmitted as a narrow beam. Despite the fact that kilowatts of multimode power can be easily extracted from a laser diode array, the resulting single-lobe single-element power from a laser diode is always limited by a value of several watts. Many applications are waiting for a spatially coherent laser diode source offering power from 1 to 100W. The potential substitution of YAG and fiber lasers by a compact, direct single- mode laser diode source would bring significant advancements in a number of applications such as LIDAR/LADAR (Light Intensity Detection and Ranging/Light Amplification Detection and Ranging) systems; high-bit-rate, long-haul free space communication systems; industrial processing applications; and many more.
The problem is not new. Over several decades, multiple research groups conducted experiments to force a large-stripe laser diode to laze in a single-mode regime. Multiple solutions were proposed. However, mechanical instabilities and associated coupled cavity problems, difficulties to realize sufficiently low and long-term reliable AR coating, and strong energy coupling between lateral modes in single-mode operation were the major reasons why, until today, all these intra-cavity experiments were not converted into commercial products.
Our concept is based on the “no-feed-back combining” technology, which is a profoundly different architectural concept for a broad-area laser diode source. An external linear optical system (with no feedback to the laser) converts the laser’s multimode emission into a spatially-coherent (diffraction-limited) spot.
The key enabler of this method is the “rock-solid” stability of the far-field mode patterns of high-power broad-area lasers. The most common laser diode testing procedure is the observation of the spectrally resolved near-field. Typically, messy clouds of overlapping longitudinal and lateral modes are present. This very common and simple observation led many laser diode specialists to a hopeless conclusion of the impossibility of linear combining of broad-area laser diode modes.
The observation of the same pattern in the far-field plane, however, provides a completely different picture; instead of a messy cloud of modes, a well defined regular pattern is observed. These “parabolic” (Fig. 1) pictures of far-field stay valid up to the power levels of catastrophic degradation (up to 10 watts for 100μm wide stripe). All modes of broad-area laser diode emission follow the rectangular box-model description, and are perfectly stable (as a whole) with respect to current and temperature, and are perfectly distinguishable in wavelength or angle. Therefore, all the modes are convertible into one spatially-coherent spot by means of a simple linear optical device. Fig. 1 illustrates this dramatic change in mode identification as a measurement method switches from near-field to the far-field observations.
The experimental demonstration capitalizing these far-field properties of laser diode radiation and realizing the spatial mode multiplexing is done with the set-up shown in Figure 2.
The light from the broad-area laser diode passes through a spectral dispersive device composed from the lens f2 and holographic grating G. This spectrally dispersive element projects all modes of the laser into different spatial positions at the phase modulator plane m2. By creating the specifically adapted binary phase patterns at the phase modulator plane m2, all modes are converted in single-lobe radiation patterns (using Fresnel zone principles). Then all modes are recombined back with a backward pass through the same spectrally dispersive device. After mirror m4, the output beam has a single-lobe elliptical pattern. An additional cylindrical lens transforms the laser emission into a circular spot. Using an off-shelf grating and lens, 30% total transmission of the combining device was demonstrated, with 60% mode conversion efficiency (ratio of single-lobe output power to the total power in the output beam with plane mirror as the phase modulator). Figure 3 shows the far-field patterns (a) with the binary phase mask and (b) with just a plane mirror in the optical pass (mirror m2).
The specifically-adapted pattern of the binary mask is fabricated in the following manner. First, a high-reflectivity mirror is covered with a phase-inverting layer of transparent polymer. The mirror is placed into the spectral plane (mirror m2). Then the spatio-spectral pattern of laser diode is measured at the surface of this coated mirror with a high spatial resolution (10000 × 640 points). A scanning setup including CCD camera and a translation stage are used. Phase alternation points are then determined using the hypothesis of Hermit-Gaussian mode profile. Lastly, precise milling instructions are transmitted to the 3-D, 50nm-resolution, fast translation stage with a sharp milling tool. This milling tool removes a transparent layer from the reflective surface in the places corresponding to the bright or dark Fresnel zones (conjugated to the infinite points). This simple method of mask fabrication demonstrated sufficient speed and good resolution for the purpose of demonstrating the concept.
To the best of our knowledge, this is the first demonstration of lateral mode combining of broad-area semiconductor laser emission into a single-lobe diffraction limited spot, using an external optical device with no feedback to the laser medium. The proposed direct combining method accepts laser diodes of any wavelength and allows direct high-speed modulation. These properties are attractive features of laser sources for ranging, free-space communication and HDTV projection systems, for which no high-power-capable external modulators are currently available. The broadband nature of this method is also convenient for ultra-short pulsed laser systems.
At UT Arlington’s High-power Laser Diode Lab, we are currently working on a 10W, single-element prototype of a spatially-coherent, compact optical source.
First, is it theoretically possible to fold all laser diode modes in the same spatial spot? The answer is yes. According to Liouville’s theorem, the brightness of a spatially incoherent source at a given wavelength cannot be increased by an imaging optical system. For example, if a source consists of two spots of white light having the same spectral density distribution and random independent phases, it is impossible to shape them in one spot of double brightness. The resulting brightness will always be less than or equal to the individual brightness of the two beams. But, if the source consists of two spots of different spectral densities, it is possible to combine them into a beam of double spatial brightness (integrated over the spectrum). The spectral brightness, however, would not increase.
Two laser beams of different wavelengths can be easily combined into one beam of double power using a diffraction grating through a technique known as WDM combining. If two beams have the same wavelength and fixed mutual phase relationship, they can be reshaped into one beam with doubled spectral brightness. This technique is called coherent combining.
The demonstrated method of combining is in full agreement with Liouville’s principle. Each lateral mode of the laser diode taken individually can be reshaped into an ultimate spot of λ/2 (coherent combining). However, an optical device for this purpose must have a specific pattern for each lateral mode. Using the fact that each lateral mode has different frequency, it is possible to create an individual mask for each lateral mode (coherent reshaping) and subsequently combine all modes (incoherently) into a single spot of λ/2. The spatial brightness will be increased proportionally to the number of lateral modes present in the laser diode cavity (10-12 times).