The quantum spin properties of nitrogen-vacancy defects in diamond have diverse applications including quantum computing and communications 1 , but nanodiamonds also have attractive properties for in vitro biosensing, including brightness 2 , low cost 3 , and selective manipulation of their emission 4 . Nanoparticle-based biosensors are vital for early disease detection, however, often lack the required sensitivity. Here we investigated fluorescent nanodiamonds as an ultra-sensitive label for in vitro diagnostics, using a microwave field to modulate emission intensity 5 , and frequency-domain analysis 6 to separate the signal from background autofluorescence 7 , which typically limits sensitivity.We focused on the common, low-cost lateral flow format as an exemplar, achieving detection limits of 8.2 × 10 −19 M for a biotin-avidin model, 10 5 -fold more sensitive than gold nanoparticles; and a use-case demonstration of single-copy detection of HIV-1 RNA with a short 10-minute isothermal amplification step, including a pilot using a clinical plasma sample with an extraction step. This ultra-sensitive quantum-diagnostics platform is applicable to numerous diagnostic test formats and diseases with the potential to transform early diagnosis, benefiting patients and populations.Rapid point-of-care tests have transformed access to disease testing in a variety of community settings, including clinics, pharmacies and the home 30 . Among the most common tests worldwide are paper microfluidic lateral flow assays (LFAs), with 276 million sold in 2017 for malaria alone 31 . LFAs satisfy many of the REASSURED criteria 32 for diagnostics, however, despite widespread use they are still limited by inadequate sensitivity to detect the low levels of biomarkers necessary for early disease detection.Fluorescent markers can be highly sensitive, but are practically limited by background fluorescence from the sample, substrate, or readout technique. In the case of nitrocellulose substrates used in LFAs, there is a significant background autofluorescence 7 , which inherently limits sensitivity. Various methods have been reported to reduce this effect, such as membrane modification to reduce background fluorescence 33 , exciting in the nearinfrared range and using upconverting nanoparticles 34 , and time-gated detection using longpersistent phosphors 35 to separate background fluorescence, which has a shorter lifetime.These methods have shown ∼10-fold improvements in sensitivity over gold nanoparticles, limited by relatively low brightness.Here we show the use of FNDs as a fluorescent label in an LFA format as a demonstrator of their first use for in vitro diagnostics, taking advantage of their high brightness and selective modulation. The use of a narrowband resonator allows for the lowpower generation of microwave-frequency electromagnetic fields, suitable for a point-ofcare device, to efficiently separate the signal from the background in the frequency domain by lock-in 6 detection. We aimed, after characterisation, functionalisation,...

Recent advances in the development of hybrid organic-inorganic lead halide perovskite (LHP) nanocrystals (NCs) have demonstrated their versatility and potential application in photovoltaics and as light sources through compositional tuning of optical properties. That said, due to their compositional complexity, the targeted synthesis of mixed-cation and/or mixed-halide LHP NCs still represents an immense challenge for traditional batch-scale chemistry. To address this limitation, we herein report the integration of a high-throughput segmented-flow microfluidic reactor and a self-optimizing algorithm for the synthesis of NCs with defined emission properties. The algorithm, named Multiparametric Automated Regression Kriging Interpolation and Adaptive Sampling (MARIA), iteratively computes optimal sampling points at each stage of an experimental sequence to reach a target emission peak wavelength based on spectroscopic measurements. We demonstrate the efficacy of the method through the synthesis of multinary LHP NCs, (Cs/FA)Pb(I/Br) (FA = formamidinium) and (Rb/Cs/FA)Pb(I/Br) NCs, using MARIA to rapidly identify reagent concentrations that yield user-defined photoluminescence peak wavelengths in the green-red spectral region. The procedure returns a robust model around a target output in far fewer measurements than systematic screening of parametric space and additionally enables the prediction of other spectral properties, such as, full-width at half-maximum and intensity, for conditions yielding NCs with similar emission peak wavelength.

A two-stage microreactor enables the quantitative evaluation of a kinetic model of nanocrystal nucleation and growth.