How do science and scientific collaboration respond to global health crises? Do these crises reinforce, accelerate or disrupt normal trends in the production of science? This project analyzes CORD-19, a growing dataset of over 200,000 scientific articles published about coronavirus/Covid-19 over the years, and Dimensions, a global bibliographic database, to study the impacts of Covid-19 on global trends in scientific production and collaboration. Drawing on theories and methods from sociolinguistics, natural language processing, and network science, we hypothesize that the Covid-19 pandemic has operated as an accelerator of long-term trends in science, including the growth in collaboration, interdisciplinarity, and cross-domain cooperation (e.g., between academic and health care institutions). We examine different types of scientific convergence produced by Covid-19, including convergence in collaboration between different disciplines, institutions, institutional domains, and countries; and convergence in terms of topics addressed and language used in coronavirus literature over time. We consider the effects of both the 2002-2004 SARS epidemic and the current pandemic, comparing collaboration and convergence patterns pre, during and post each of these two global health crises.