Spin-dependent photoluminescence in carbon-based quantum dots

TL;DR

This work demonstrates room-temperature spin-dependent photoluminescence in amino-acid derived carbon quantum dots (aa-CQDs), delivering MPL at about $10~\mathrm{mT}$ and optically detected ESR with $g \approx 2$, consistent with a radical-pair mechanism. The authors synthesize 19 aa-derived CQDs, perform comprehensive structural and optical characterizations, and implement MPL and ODMR measurements that reveal robust spin dynamics and a paramagnetic-sensing response. They show MPL can be quenched by RF-driven spin mixing and by paramagnetic species, and demonstrate a lock-in imaging approach that exploits MPL to suppress background fluorescence. These results position aa-CQDs as scalable, biocompatible optical-spin sensors with potential for in situ bio-imaging and sensing, and suggest avenues for performance optimization through chemical engineering and pulsed schemes.

Abstract

The ability to modulate the photoluminescence (PL) of nanomaterials via spin-related effects is vital for many emerging quantum technologies, with nanoscale quantum sensing and imaging being particular areas of focus. Carbon-based quantum dots (CQDs) are among the most common forms of luminescent nanomaterials, appealing due to their ease of synthesis, tunability through organic chemistry, high brightness, and natural biocompatibility. However, the observation of room temperature, spin-dependent PL has remained elusive. Here we report on the observation of PL modulation of CQDs by magnetic fields ($\sim 10$ mT) under ambient conditions. We synthesize a series of CQDs using 19 different amino acids, which have a range of PL emission spectra and exhibit a clear magneto-PL effect (up to $\sim 1$% change). Furthermore, an electron spin resonance is detected in the PL with a g-factor of g $\approx$ 2, suggesting a process similar to the radical pair mechanism is responsible. Finally, we show that the magneto-PL contrast decreases in the presence of paramagnetic species, which we attribute to an increase in magnetic noise-induced spin relaxation in the CQDs. Our work brings new functionalities to these commonly used and biocompatible luminescent nanoparticles, opening new opportunities for in situ quantum sensing and imaging of biological samples.

Figures (27)

• Figure 1: Synthesis of amino acid-derived carbon-based quantum dots.a Schematic showing direct conversion of amino acids to carbon-based quantum dots (CQD) via pyrolysis, and TEM images of two different amino acid dervied aa-CQDs. b High resolution C 1s XPS spectrum of the Gly-CQD sample. c Absorption (purple trace) and PL emission spectrum (blue trace) of Gly-CQDs in water, and PL spectrum of the same particles dried on glass (green trace). PL spectra were acquired using 405 nm excitation. d Photos of a few samples under 365 nm illumination. The bar graph indicates the photoluminescence intensity of all 19 samples (normalised to the brightest) in water under 405 nm excitation, collected through a 420 nm longpass filter. The samples are ordered by decreasing brightness.
• Figure 2: Magneto-photoluminescence (MPL) in aa-CQDs.a Schematic of the experiment, where aa-CQDs are placed on a printed circuit board (PCB) for RF control (used in Fig. \ref{['fig3']}), with an electromagnet positioned underneath for $B$-field control. The light excitation and collection is performed from above and separated with appropriate filters. b PL response to switching magnetic fields on and off for a few exemplar samples both dry (green) and in water(blue). The sampling rate was varied between 2 and 6.25 Hz depending on brightness, to obtain an acceptable signal-to-noise ratio. c Histogram of the MPL contrast observed for the 19 aa-CQDs samples dry (green) and in water (blue) form. The error bar is the standard deviation from multiple on-off cycles. MPL is confirmed in all but 3 samples for which the error bar is larger than the mean. d Example time traces of the PL when the magnetic field is turned on for a few dry (top panel, green) and in-water (bottom panel, blue) samples. Data for all samples can be found in the SI (Fig. S13). e Maximum MPL contrast as a function of the applied magnetic field for Gly-CQDs dry (green) and in water (blue) form. Solid lines are stretched-exponential fits.
• Figure 3: Optically detected magnetic resonance and spin-pair origin.a Simplified model with optical cycling between ground and excited states (GS and ES). Charge transfer from the excited state creates a weakly coupled spin pair. The singlet state (S) of the spin pair decays directly back to the GS, while the triplet state (T) decays to an intermediate long-lived state first. Singlet-triplet mixing occurs either at $B=0$ or under RF driving, favouring a rapid return to the GS, resulting in maximum PL. b PL change relative to the $B=0$ condition, as a function of magnetic field with an RF driving field applied at 980 MHz, measured for dry Gly-CQDs. A cancelling of the MPL effect is evident at $B\approx35$ mT corresponding to the ESR condition. The solid line is a fit with a stretched-exponential decay plus a Lorentzian peak. c Zoom-in on the ESR peak for different RF driving powers. The solid line is a Lorentzian fit. d ESR spectra taken by scanning the RF frequency at different magnetic field strengths from 30 to 170 mT. Inset: Relationship between the ESR position and the applied magnetic field, yielding a slope $g=2.01(1)$. e Contrast of the ESR peak ($C_{\rm RF}$) as a fraction of the MPL contrast at the same field ($C_{\rm M}$), i.e. the ratio $C_{\rm RF}/|C_{\rm M}|$, for the 16 aa-CQDs samples exhibiting clear MPL contrast, measured at $B=35$ mT. For most samples the ratio approaches 100% meaning the MPL effect is completely suppressed by the RF.
• Figure 4: Sensing paramagnetic molecules in solution.a Illustration of the paramagnetic sensing experiment, where gadoteric acid is added to a solution of CQDs causing a reduction in MPL contrast. b PL emission spectrum with 405 nm excitation of the control Tyr-CQD solution compared with 28 mg/ml of gadoteric acid, showing only a subtle difference. c MPL trace when switching $B$ between 0 and 46 mT every 5 s, for the Tyr-CQD control (blue) and with 28 mg/ml of gadoteric acid (purple). d MPL response (Tyr-CQD) to increasing magnetic fields for different gadoteric acid concentrations (solids lines are exponential fits). e MPL contrast at $B=46$ mT of several aa-CQDs as a function of gadoteric acid concentration (solids lines are exponential fits).
• Figure 5: Bar graphs showing the atomic concentrations of elements based on XPS analysis of aa-CQDs (top) compared to the elemental compositions of amino acid precursors (bottom).