Scientific tools dictate the boundaries of human knowledge, serving as the foundation for perceptions and explorations. In the era of Big Science, science are increasingly dependent on advanced analytical technologies and experimental platforms. Over the past decades, national and supranational entities have invested massive financial resources, collaborative networks, and collective intelligence to construct Big Science Facilities (BSFs) aimed at generating cutting edge knowledge. However, empirical evaluations of these machines actual performance in driving scientific innovation remain scarce. To address this gap, we collected 310,086 publications from 88 global BSFs and constructed a matched control dataset of approximately 3 million publications sharing the same last authors. Our analysis reveals that the utilization of BSFs has expanded significantly since 1950s. Crucially, publications supported by these facilities exhibit higher recombinant novelty and interdisciplinary integration. Furthermore, this improvement is most pronounced in non physical sciences domains traditionally peripheral to BSFs core focus indicating the emergence of a powerful intra facility knowledge spillover effect. By enriching the Facilitymetrics framework, our findings provide empirical evidence that BSFs act as vital engines for scientific discovery, offering policymakers essential metrics to justify infrastructural investments, while prompting the science of science community to reassess the profound impact of scientific tools on knowledge production

翻译：科学工具界定了人类知识的边界，是感知与探索的基础。在大科学时代，科学研究日益依赖于先进的分析技术与实验平台。过去数十年间，国家和超国家实体投入了大量财力、协作网络及集体智慧，建设旨在产生前沿知识的大科学设施（Big Science Facilities，BSFs）。然而，对这些机器在推动科学创新方面的实际表现进行的实证评估仍然匮乏。为弥补这一空白，我们收集了来自全球88个BSF的310,086篇论文，并构建了一个包含约300万篇具有相同末位作者论文的匹配对照数据集。我们的分析揭示，自20世纪50年代以来，BSF的利用规模显著扩张。至关重要的是，由这些设施支持的论文展现出更高的重组新颖性和跨学科整合程度。此外，这种提升在传统上属于BSF核心领域之外的物理科学领域最为显著，表明一种强大的设施内知识溢出效应正在显现。通过丰富设施计量学（Facilitymetrics）框架，我们的研究结果为BSF作为科学发现的关键引擎提供了实证证据，为政策制定者提供了证明基础设施投资合理性的关键指标，同时促使科学学界重新评估科学工具对知识生产产生的深远影响。