Share this post:
Photograph of a GaP-on-insulator chip with integrated devices being measured with optical fibers. The green glow is the third-harmonic light generated while pumping one of the ring resonators with a laser.
In the paper “Integrated gallium phosphide nonlinear photonics”, recently published in the peer-reviewed journal Nature Photonics, we report on the development of high-performance photonic devices made of the crystalline semiconductor gallium phosphide. This work represents a breakthrough in the manipulation of light with semiconductor materials integrated on a chip. It opens the door to a multitude of applications that could have significant impact on information technology and the future of computing.
Gallium phosphide (GaP) has been an important material in photonics – the science and technology of light – since the 1960s, forming the basis for a range of light-emitting devices. Despite this early start, the lack of methods to fabricate complex GaP structures on a chip has prevented the development of more sophisticated devices, such as photonic integrated circuits. Recently, our team at IBM Research – Zurich invented a scalable and manufacturable solution for integrating high-quality GaP on the same wafers as used in the electronics industry. Together with colleagues from the École Polytechnique Fédérale de Lausanne (EPFL), we have now exploited this capability to create exceptional on-chip photonic devices, heralding a new era in which GaP can be integrated with other building blocks employed in computing hardware. We expect the addition of GaP to the photonics toolkit to have a major impact on applications as diverse as telecommunications, sensing, astronomy and quantum computing.
On-chip frequency comb generation with GaP
In our paper, we demonstrate the capabilities of the integrated GaP platform by engineering waveguide resonators producing optical frequency combs. A frequency comb is a light source with a spectrum consisting of a series of equally spaced narrow lines. Such a spectrum corresponds to a regular train of ultrashort light pulses having a fixed repetition rate. Based on work going back to the late 1970s, the inventors of frequency combs were awarded the Nobel Prize in physics in 2005.
Optical frequency combs are used today as optical ‘rulers’ (a method of precisely measuring optical frequencies to create, for example, ultraprecise optical clocks), in high-resolution spectroscopy, and as a link between microwave and optical signals. The scientific instruments necessary to generate frequency combs can be bulky and expensive, filling an optics laboratory. Integrated photonic devices offer an attractive alternative, as they can be operated at low power, fabricated at low cost, and combined with electronic devices.
Scanning electron microscope image of a GaP-on-insulator waveguide ring resonator on a silicon chip.
But the materials previously used to generate such frequency combs usually either do not operate at low-power or cannot be integrated on chips because they are not compatible with the established fabrication techniques. We have overcome these challenges with our GaP platform. We generate broadband (>100 nm) Kerr frequency combs in the telecommunications C-band with a threshold power as low as 3 mW. Because of the strong second-order nonlinearity of GaP, we also simultaneously form frequency combs at double the frequency, close to the visible spectrum, and for certain devices, we observe efficient Raman lasing. Propagation loss in these devices is only 1.2 dB/cm – a remarkably low value for such an immature technology and comparable to state-of-the-art silicon-on-insulator waveguides.
What’s so special about GaP?
GaP possesses an attractive combination of a large refractive index (n > 3 for vacuum wavelengths up to 4 μm) and a large electronic bandgap (2.26 eV). The former allows light to be confined to a small volume; the latter implies a wide transparency window. There are few materials which exhibit these inherently conflicting properties, as there is typically a tradeoff between index of refraction and bandgap. GaP offers a unique possibility of creating devices with strong light confinement (small mode volumes), transparency into the visible (λvac > 550 nm) and enhanced light-matter interaction. Importantly, two-photon absorption at the typical data communication wavelengths of 1310 nm and 1550 nm is dramatically diminished in comparison to silicon photonics. Consequently, high intensities can be used, as it often occurs in nanophotonic devices. In addition, GaP features a high second- and third-order nonlinear susceptibility, enabling efficient three- and four-wave mixing, the nonlinear optical processes in which we are interested.
Numerous applications on the horizon
Besides frequency comb generation, our GaP devices efficiently double and triple the frequency of laser light, providing a means to do on-chip wavelength conversion. We expect the nonlinear processes can be extended to create a supercontinuum, a broad spectrum of spatially coherent light that can be used for sensing, optical communications, and sophisticated scientific measurements such as optical coherence tomography for medical analysis of biological tissues. Importantly, our fabrication process is compatible with CMOS electronics and independent of the underlying substrate stack. Therefore, the GaP devices can be monolithically integrated with other more established photonic technologies, such as silicon or indium-phosphide photonics, or even on a CMOS electronics chip, to realize complex hybrid devices. One possibility is a fully integrated electro-optic modulator for high-speed optical interconnects as used in data centers and supercomputers. Beyond such classical applications, the second-order optical nonlinearity of GaP could be leveraged to create devices coupling optical and microwave fields at the level of individual photons. Such devices would serve as quantum-coherent transducers for connecting superconducting quantum computers with fiber-optic cables. As a whole, our paper showcases the unique advantages of integrated GaP photonics and signals the emergence of a mature new platform for nonlinear photonics.
This work was carried out in a collaboration between IBM Research – Zurich and the École Polytechnique Fédérale de Lausanne (EPFL) and supported by the European Union’s Horizon 2020 Programme for Research and Innovation under grant agreements No. 722923 (Marie Skłodowska-Curie H2020-ETN OMT) and No. 732894 (FET Proactive HOT). Figure 2