Non-contact scanning probe station for advanced wafer scale testing of photonic integrated circuits

集成光子制造正在迅速成为一个成熟的数十亿磅全球产业，因为光子学正在支持与英国在人工智能和Web 5.0，未来移动，医疗保健和未来传感器方面的工业战略相一致的广泛应用。测试和测量是工业和研究中大多数制造工作流程的关键部分，因为它有可能在最早的阶段，当然也是在任何昂贵的封装和集成步骤之前，发现偏差。晶圆上测试提供了一个早期的机会，可以通过防止故障设备在处理工作流程中继续，从而避免不必要的资源、工具和能源浪费，从而潜在地节省大量成本。重要的是，晶圆测试允许反馈由制造过程中的故障和超过制造公差的较小漂移引起的与原始设计的任何偏差。半导体行业领先的IRDS路线图2020已经确定了光子学在下一代计算中的重要性，但也指出了现有测试工具中的关键瓶颈，以解决光子芯片的良率，可变性，精度和可调谐性方面的挑战。制造缺陷是目前实现可靠的大批量光子制造的主要限制因素之一。随着新的光子探针扩展到半导体电子制造中常用的商用晶圆探针的出现，现在可以获得一系列复杂的端到端表征。这是一系列测试的理想选择，这些测试验证性能，并根据预期输出验证整个电路响应，以及识别导致系统级故障的关键离群值。然而，目前这一代的工具缺乏提取光子电路内部发生的信息的能力。随着集成电路变得越来越复杂，电路中缺少中间探测点成为一个越来越紧迫的问题。事实上，这个问题最近在其他研究项目中得到了解决，其中小组提出了电路中的可擦除输出耦合器作为对中间探测点进行更深入测试的一种选择。然而，正如我们在导致该项目的一些原则性研究证明中所示，一种更普遍的方法是可以实现的。事实证明，这些光子电路中使用的半导体对短波长的紫外光有反应，实际上反应足够灵敏，以至于设备中一个微小的微观点的照明在电路的输出端产生可追踪的信号。通过设备扫描这个点，我们可以建立一个详细的图像，光在时间和空间上的位置。我们甚至可以在波长中解析这张图，以建立一个远远超出商业探针台能力的器件性能的完整图像。虽然到目前为止这仍然是一个基本的研究课题，但我们在这里提出将这种方法作为晶圆级光子学测试的通用工具。为此，我们需要使这些技术更快，更强大，更可靠地用于制造工作流程，并与不同平台上最终用户的实际要求保持一致。因此，该项目的大部分内容都集中在开发这种仪器上，使用开源Python数据采集框架进行操作，以实现互操作性和用户定制，并为每个平台生成一组令人信服的测试和演示程序，这些测试和演示程序将用于利用该平台在不同应用领域的功能。在项目结束时，我们希望我们已经开发出一个独立的仪器，将在世界各地的广泛的研究和制造环境中使用。

Integrated photonics manufacturing is rapidly becoming a mature multi-billion pound global industry as photonics is underpinning an very wide range of applications aligned with UK's industrial strategy in AI and Web 5.0, the Future of Mobility, Healthcare, and Future Sensors. Test and measurement forms a critical part of most fabrication workflows both in industry and in research, for its potential of picking up deviations at the earliest stage possible and certainly before any expensive packaging and integration steps. On-wafer testing provides an early opportunity which can potentially save significant costs by preventing malfunctioning devices to continue in the processing workflow thus avoiding unnecessary waste of resources, tooling and energy. Importantly, wafer testing allows to feed back any deviations from the original design caused by failures in the fabrication process and smaller drifts exceeding the manufacturing tolerances. The semiconductor industry's leading IRDS Roadmap 2020 has identified the importance of photonics in next generation computing but has pointed out critical bottlenecks in available testing tools for addressing challenges in yield, variability, precision, and tunability of photonic chips. Fabrication imperfections are currently amongst the main limiting factors for achieving reliable high-volume photonics manufacturing.With the appearance of new photonic probe extensions to commercially available wafer probers commonly used in semiconductor electronics manufacturing, a range of sophisticated end-to-end characterisations is now available. This is ideal for a range of tests verifying performance and validating the entire circuit response against the expected output and identifying the critical outliers which are responsible for failure at the systems level. However the current generation of tools lack the capability of extracting information on what happens inside the photonic circuit. As integrated circuits become more and more complex, the lack of intermediate probe points in the circuit becomes an ever more pressing issue. Indeed this issue was addressed recently in other research projects, where groups have proposed erasable output couplers in the circuit as an option for more in-depth testing of intermediate probe points. However a more general approach is within reach as shown by us in a number of proof of principle studies leading to this project. It turns out that the semiconductors used in these photonic circuits are responsive to short-wavelength UV light, in fact responsive enough that illumination of a small microscopic point in the device gives rise to a traceable signal at the output of the circuit. By scanning this spot through the device, we can build up a detailed image of where the light is in both time and space. We can even resolve this map in wavelength, to build up a complete picture of device performance far beyond the capabilities of the commercial probe stations.While this so far has remained a basic research topic, we propose here to push this approach forward as a versatile tool for wafer-scale photonics testing. For this we need to make the techniques much faster, robust and reliable for use in a manufacturing workflow, and aligned with the actual requirements of the end users on different platforms. The majority of the project is therefore focused on developing this instrumentation, operating this with an open source Python data acquisition framework for interoperability and user customization, and generate a convincing set of tests and demonstrators for each platform that will be used to leverage the capabilities of this platform for different application areas. At the end of the project we expect that we have developed a self-contained instrument that will find use in a wide range of research and manufacturing environments around the world.