High-precision chemical quantum sensing in flowing monodisperse microdroplets

Abstract

A microfluidic device includes an aqueous phase of analyte and nanodiamonds in an aqueous inlet, a first oil inlet, a first junction fluidically connected to the aqueous inlet and the first oil inlet, the first junction configured to cause droplets of the analyte and quantum sensors to become suspended in oil from the first oil inlet, a second junction fluidically connected to the first junction and a second oil inlet, the second junction configured to control spacing between the droplets, a microwave source adjacent the second junction operable to apply microwaves to the droplets at a frequency greater than a frequency of a flow of the droplets, and a microscope arranged to illumination the droplets and sense signals from the droplets while the microwaves are applied to perform noise-suppressed quantum sensing. A composition of matter includes a microdroplet of fluid containing an analyte and quantum sensors.

Description

Patent Application U Cal No. BK-2024-119-2-PCT MN No. 407869-0201 HIGH-PRECISION CHEMICAL QUANTUM SENSING IN FLOWING MONODISPERSE MICRODROPLETS STATEMENT REGARDING FEDERALLY SPONSORED RESEARCH AND DEVELOPMENT

[0001] This invention was made with Government support under grant number DE-AC02-O5CH11231, DE-SC00230265, and DE-SC0022441 awarded by the Department of Energy. The Government has certain rights in this invention.TECHNICAL FIELD

[0002] This disclosure relates to high-precision chemical detection, more particularly to quantum sensing with droplet microfluidics.BACKGROUND

[0003] Quantum sensing is rapidly reshaping the ability to discern chemical processes with high sensitivity and spatial resolution, with the potential to impact a range of disciplines from synthesis to bioengineering. Specifically, sensors based on nitrogen-vacancy (NV) defects in diamond translate optically addressable electronic spin state information into detectable fluorescence signals in a manner that is sensitive to the local chemical environment. This has led to diverse applications, including in-cell thermometry and reactive oxygen species detection, high-sensitivity lateral flow assays, and nuclear magnetic resonance (NMR) measurements in picoliter volumes, marking an exciting leap in precision measurement technologies.

[0004] Traditionally, quantum sensing for chemical analysis has relied on single crystals hosting shallow NV centers. High throughput analysis faces challenges due to the small (mm-scale) sizes of these crystals, their substantial cost, and the need for precise crystal orientation. Only a fraction of the diamond crystal is used for sensing, and analyte moleculesare required to flow over its surface, necessitating complex integration of microfluidic structures directly onto the diamond.

[0005] Nanodiamond (ND)-based sensing offers a compelling alternative as they are inherently deployable and can yield spatially selective sensing in or near targeted volumes of interest. NDs are also low-cost, non-reactive, and bio-inert, and for particles of less than 40 nm, a significant proportion of their NV centers can interact with external analytes. Advances in nanodiamond chemistry have facilitated surface functionalization to control surface charge, hydrophilicity, or hydrophobicity, and for targeting proteins or cellular organelles.

[0006] Despite these advantages, ND-based sensing is fraught with challenges. Significant heterogeneity in particle size, shape, and NV center coherence times leads to large statistical errors in multi-particle analysis. Additionally, fluorescence fluctuations arise from different particle orientations, and spatial variations in the materials in which the particles are targeted. Overcoming these significant challenges is essential to achieving high-precision quantum sensing.BRIEF DESCRIPTION OF THE DRAWINGS

[0007] FIGs. 1 A-1D shows an embodiment of a microfluid device in accordance with the embodiments, and accompanying photos, Brightfield images and graphs.

[0008] FIGs. 2A-2B shows optically detected magnetic resonance (ODMR) measurement results in droplets.

[0009] FIGs. 3A-3G shows results of double modulation lock-in detection.

[0010] FIGs. 4A-4F shows results of analysis of stability metrics for in-flow droplet measurement.

[0011] FIGs. 5A-5F shows results of sensing paramagnetic species in flowing microdroplets.

[0012] FIGs. 6A-6C show embodiments of droplets containing surface functionalized magnetic beads and quantum sensors.

[0013] FIG. 7 shows examples of driven particle tracks and Brownian motion tracks.DETAILED DESCRIPTION OF THE EMBODIMENTS

[0014] The embodiments herein directly address these challenges by deploying nanodiamonds within flowing, monodisperse, picoliter-volume microdroplets that host analyte molecules. Rapid movement of the ND particles within the droplets effectively averages out their heterogeneity and ensures close interaction with the analytes. The embodiments take advantage of stable and controllable flow afforded by droplet microfluidics to develop a new method for background-free quantum sensing at high-throughput. Additionally, the dynamic nature of the flowing droplets yields exceptional measurement stability, resistant to experimental variations and temperature shifts. The embodiments herein demonstrate this stability over greater than 103measurement and across ~ 105droplets, greatly surpassing the typical stability in conventional quantum sensing experiments. Additionally, the ND sensor volumes required are minuscule, costing only about $0.63 for an hour of analysis over hundreds of thousands of droplets.

[0015] The embodiments involve a new platform technology fusing quantum sensing and droplet microfluidics and is marked by several novel aspects. Picoliter-scale droplets can function as microscopic confinement chambers, encapsulating diverse analytes, ranging from single cells to chemical reaction products, and can stably accommodate a broader range of concentrations than bulk solutions. The droplets are precisely controllable in terms of diameter, charge, and environmental conditions, and their movement under flow enhances sensor-analyte mixing. This approach is also amenable to digital control techniques for droplet “arithmetic,” including mixing, collisions, and sorting, further enhancing their application in quantum sensing.

[0016] The embodiments involve a novel flow-based method for high-precision chemical detection that integrates quantum sensing with droplet microfluidics. The process deploysnanodiamond (ND) particles hosting fluorescent nitrogen vacancy (NV) defect centers as quantum sensors in rapidly flowing, monodisperse, picoliter-volume microdroplets containing analyte molecules. ND motion within these microcompartments facilitates close sensor-analyte interaction and mitigates particle heterogeneity. One should note that NDs only comprise one example of quantum sensors that may be encapsulated in the droplets. The below discussion focuses on NDs as quantum sensors for ease of discussion, but no limitation to that particular type of quantum sensor is intended nor should any be implied.

[0017] Microdroplet flow rates are rapid (up to 4cm / s) and with minimal drift. Pairing this controlled flow with microwave control of NV electronic spins, the embodiments introduce a new noise-suppressed mode of Optically Detected Magnetic Resonance (ODMR) that is sensitive to chemical analytes while resilient against experimental variations, achieving detection of analyte-induced signals at an unprecedented level of a few hundredths of a percent of the ND fluorescence.

[0018] The embodiments demonstrate the application to detecting paramagnetic ions in droplets with simultaneously low limit-of-detection and low analyte volumes, in a manner significantly better than existing technologies. This is combined with exceptional measurement stability over 103seconds and across hundreds of thousands of droplets, while utilizing minimal sensor volumes and incurring low ND costs (<$0.70 for an hour of operation). Additionally, the embodiments demonstrate using these droplets as microconfinement chambers by co-encapsulating ND quantum sensors with a variety of analytes, including single cells. This versatility suggests wide-ranging applications, including singlecell metabolomics and real-time intracellular measurements from bioreactors. The embodiments pave the way for portable, high-sensitivity, amplification-free, chemical assays with high throughput; introduces a new chemical imaging tool for probing chemical reactionswithin microenvironments; and establishes the foundation for developing movable, arrayed quantum sensors through droplet microfluidics.

[0019] The microfluidics platform 10, schematically depicted in FIG. 1A, features a device that produces phase-separated, monodisperse, droplets in a water-in-oil emulsion, varying in diameter in range 10-150 pm and volume 300 fL - 500 pL. In one embodiment, devices are made from PDMS (polydimethylsiloxane) via soft lithography and bonded to a glass cover slip. Each device supports multiple devices, each hosting diverse microfluidic structures.

[0020] Droplets are formed by constricting an aqueous phase comprised of quantum sensors with an analyte of interest with oil channels from the oil inlets 12 and 14 using a fluidfocusing geometry at junction 18, as depicted in FIG. 1A, through an orifice 22 ~ 30pm orifice. This particular embodiment uses two oil inlets, but more or less may be used. Also in this particular embodiment, the quantum sensors, in the form of NDs, enter the system through the ND and analyte inlet in water at 16. Again, other fluids may be used that allow control of the droplets and their flow as discussed in more detail further. This creates a stable stream of uniformly spaced droplets. Droplet size and formation rate are controlled by the orifice size and water / oil flow rates. The process achieves droplet speeds in excess of 4 cm / s. The fabrication process ensures error of approximately less than or equal to 1pm error, guaranteeing high reproducibility across chips.

[0021] Additionally, the chips host a second junction 20 in FIG. 1A, which enables oil injection to precisely control droplet spacing. In one embodiment, the spacing change from 154 pm to 360 pm. This spacing is maintained constant and droplet flow is highly uniform downstream, with consistent speed, uniformity, and stability over several hours. The droplets are analyzed over a microscope objective 32 at 34 and microwave (MW) coil 30 for imaging and quantum sensing measurements. The ODFR has a laser that illuminates the sample in the objective area 22, with laser light 36. The objective 34 of the ODFR occurs before flowingthrough a circuitous path 24 that induces intradroplet mixing and extends the channel for downstream analysis of stationary droplets. Lastly, they are directed into a collection chamber 26, allowing for the simultaneous storage and further examination of over a hundred droplets. The collection chamber and the circuitous path are optional. The collection chamber may exist without the circuitous path and connected directly to the second junction. The material then may be removed through the collection outlet 28.

[0022] FIG. IB shows a closer view of the junction 18. The quantum sensors and analyte enter the junction 18, the aqueous phase and the oil do not mix, resulting in droplets of the quantum sensors and analyte separated from each other. FIG. 1C shows the second junction 20 in which the droplets are further separated from each other to increase their flow.

[0023] Diamond particles are incorporated into droplets in flow using an aqueous suspension of carboxylated (hydrophilic) NDs, which host ~ 1-3 ppm NV centers, as the dispersed phase. Zeta potential measurements reveal a surface charge sufficient to confer hydrophilicity and colloidal stability. The fluorescent NDs are thereby completely encapsulated within the droplet, as shown in FIG. 1C. An inverted configuration, with NDs in oil droplets surrounded by an aqueous phase, can instead be achieved by coating ND surfaces with polymer chains.

[0024] Fluorescence imaging allows visualization of microdroplets loaded with NDs of varying sizes. In one embodiment the microdroplets has sizes of 40 nm, 100 nm, and 3 pm sized particles. The NDs feature a heterogeneous size distribution (±30nm) characterized by dynamic light scattering (DLS) spectroscopy. In one embodiment, 40 nm NDs achieve a well-dispersed distribution within the droplet, occupying 0.01% of its volume with minimal evident aggregation. In one embodiment, each droplet contains around 106particles and a total diamond mass of less than 100 pg. This is significantly lower than that employed in single-crystal diamond microfluidic-channel methods. The 100 nm particles remain well dispersed in solution and the 3 pm particles, while even brighter, tend to settle at the dropletbottom for stationary droplets, indicating a trade-off between fluorescence intensity and colloidal stability.

[0025] The larger particles remain stationary in a static droplet but can be induced to sample the droplet volume upon motion. For smaller particles however, Brownian motion is much more pronounced, allowing the NDs to traverse large swathes of the droplet volume. In In one embodiment, the trajectories of individual 100 nm ND particles were measured within a droplet over 30 s. FIG. ID shows a schematic of a droplet 40 containing one or more quantum sensors such as 42, and an analyte of interest 44.

[0026] FIG. IE presents a histogram of total displacement for these particles, based on tracking 200 ND trajectories for the same interval. They traverse distances greater than 5 pm, several fold larger than their diameter. The long tail in the distribution points to anomalous diffusion reminiscent of Levy flight processes. Overall, these large excursions promote interaction with droplet-confined analytes.

[0027] Microdroplets can also serve as picoliter-scale containers capable of co-encapsulating entities. Diamond aggregation, influenced by ions in the yeast growth medium, can be managed by altering the medium or modifying the diamond surface functionality. The NDs can be targeted to the yeast cells via surface functionalization with Concanavalin-A, a protein with affinity to the cell surface.

[0028] Chemical sensing utilizes the NV center’s electronic spin sensitivity to its environment and the ability to convert this into optical signals via its spin-state-dependent fluorescence. This involves the measurement of NV center optically detected magnetic resonance (ODMR) spectrum, where ND photoluminescence (PL) is monitored as MW excitation is swept in frequency across NV spin transitions (FIG. 2A). At Earth’s field, the PL exhibits a characteristic dip at ~2.87 GHz, signifying a shift in spin population from the brighter ms= 0 state to the dimmer ms= ±1 states on -resonance.

[0029] FIG. 2A shows the first-reported ODMR measurements from ND particles in a single microdroplet, here held static. PL here is at the sub-nW level and is detected using a multipixel photon counter. These measurements are inherently noisy due to background, including PDMS autofluorescence, and fluctuations stemming from Brownian motion and particle reorientation.

[0030] To enhance measurement signal-to-noise ratio (SNR), the process utilizes MW lock-in detection by amplitude modulating the applied MWs at 1 kHz. This lock-in frequency is chosen from a balance between rates of NV repumping and MW-driven population redistribution. The result, shown in FIG. 2B, is an order of magnitude increase in SNR making the strain splitting near 2.87 GHz clearly visible.

[0031] The ODMR contrast, marked C in FIG. 2A, quantifies the ODMR dip relative to off-resonance PL. At fixed laser and MW powers, C serves as a proxy for the NV electronic Ti, which is influenced by dipolar interactions with analytes in the droplet. Low limit of detection (LOD) sensing necessitates the ability to detect minute changes in C. While the lock-in strategy in FIG. 2B yields an increase in SNR, its contrast Clock remains susceptible to non-analyte specific factors like fluorescence variations from background and particle motion the absence of an off-resonance reference in this measurement. This poses challenges for chemical sensing.

[0032] A mitigating strategy involves continuously measuring NV PL and normalizing it to the resonant ODMR lock-in signal in FIG. 2B, creating a ratio-metric measurement shown in FIG. 2C. Here, the ODMR signal from flowing droplets with MWs at 2.86 GHz over 90 seconds is normalized to the total PL measured at each instant. The normalized contrast, still referred to as C for convenience, in FIG. 2C has an average of 2.4% (dashed line); a 1% contrast window is shown here for clarity. Nonetheless, noise from PL measurements still affects C. Sensitive detection of analytes is limited by ability to resolve small fluctuationsabout this baseline contrast C. The discussion refers to this measurement precision as AC, and by convention it is reported it as a percentage of the base C level, rather than in absolute units. Indeed, C itself varies based on sample and experiment conditions, but the percent error AC allows for a standardized comparison across different experiments. From a histogram of the data (FIG. 2C(i)), one can estimate AC ~13% in this case, setting a bound on the quantitative sensing at low LOD.

[0033] To improve measurement precision and render ODMR contrast immune to background fluctuations, the process employs a strategy utilizing droplet flow, as illustrated in FIG. 3 A. Droplets, uniform in size and nanodiamond content, move at a controlled velocity (v) and sequentially enter the analysis region, where an optical spot, roughly equal to the droplet diameter, illuminates them. The spacing between droplets is vf^1, where fo is the rate at which droplets are analyzed in region a in FIG. 1 A, and v ranges from Imm / s - 4cm / s in experiments. FIG. 3A also illustrates mixing due to flow.

[0034] Simultaneously with flow, droplets are subject to MWs amplitude modulated at JMN =1 kHz (FIG. 3B) at 2.866 GHz under continuous laser illumination, introducing two distinct modulations to the PL: fc> and fuw (FIG. 3B). By arranging , / ir » / b, each droplet’s signal contains multiple MW modulation cycles. Locking into these frequencies allows us to filter out background noise, a method the discussion refers to as “double lock-in” detection.

[0035] Since the method rests on exploiting droplet flow, the discussion first highlights some of its features. FIG. 3C illustrates bright-field and fluorescence images of diamond-filled droplets in motion, complementing the stationary images in FIG. 1C. Flow exhibits remarkable regular modulation at fo due to droplet monodispersity, as evidenced by FIG. 3D displaying measured photoluminescence over 500 s and greater than 104droplets. Individual Is windows are shown here, and dashed rails highlight droplet stability, see also FIG. 4.

[0036] FIG. 3E zooms into a representative 140 ms window, showing modulation in the PL both from droplet flow and from the MWs. Solid lines are a fit to the data. The distinct timescales of both droplet and MW modulations are evident. This is clearer in the inset, FIG.3E(i), which focuses on a 30 ms window, illustrating oscillations spaced by f^w (Ims), with amplitude at -0.5% of raw PL (shaded grey region), consistent with the ODMR contrast in FIG. 2C.

[0037] The PL double modulation can be expressed as the functional form,S(t) = [m(t) + g(t)cos(2nfot ] • [1 - C(t)cos(2nfMwt + (p]+b(t) (1) where g(t) represents the droplet modulation profile influenced by droplet shape and separation. Microfluidic control enables adjustable profiles, from sinusoidal to square-like. C represents the ODMR contrast, with time dependence included to account for possible longterm drift across numerous droplets. m(t) reflects the baseline ND contribution to PL, becoming more prominent as inter-droplet spacing decreases (and controllable via junction J , see FIG. IB(ii)). Finally, b(t) captures PL noise from non-ND contributions.

[0038] FIG. 3F displays the Fourier transform amplitude J7of the PL time series from FIG.3E, measured over 15s of droplet flow, after subtracting its mean value. Data here is shown on a logarithmic scale for clarity. It features a distinct Fourier peak at fo =29 Hz with a narrow linewidth ( b — I Hz), reflecting minimal drift in flow rates (see FIG. 4). Solid lines here are Lorentzian fits. Square-like modulation leads to secondary harmonics at multiples of while background signal and autofluorescence occur near zero frequency and can be excluded. Expanding Eq. (1) uncovers frequency bands linked to MW modulation and their combinations with flow, at )MW and )MW ± fo. This is shown in the yellow shaded region, FIG.3F(i). It is evident the peak intensities here are - 1% that of the droplet modulation, a reflection of contrast C. FIG. 3F(ii) shows the same frequency window in a linear scale for clarity.

[0039] The ODMR contrast C(t) can now be calculated from the ratio of FT peak intensities at the MW and droplet frequencies. Within fixed time windows AC, this can be expressed as:\ _ F(f MW.ti)+ ?(fMW+ fp,td+ fp.ti) , 1(2)1 iJ~ ?tfp,ti) ' Zo(ti) Here, T (f ti) denotes the Fourier intensity at frequency / for a time interval bounded by tt and ti - h, averaging the PL over several droplets. Dividing the MW-associated FT peak intensities by the droplet frequency FT peak intensity yields a ratiometric ODMR contrast, factoring in the droplet PL. Eq. (2) also includes normalization proportional to droplet frequency to counteract minor frequency drifts that impact the baseline ND-dependent PL(m / )), providing a consistent contrast metric irrespective of droplet flow rate.

[0040] FIG. 3F additionally highlights that the noise profile diminishes oc 1 / suggesting that higher droplet rates could lead to lower noise while enabling greater analysis throughput. To demonstrate the versatility and control over droplet modulation in experiments, FIG. 3G shows smoothed PL profiles at various ? rates. At the upper limit ( ? =346.5Hz), more than a million droplets can be analyzed per hour (FIG. 4C), the droplets flowing here at a rapid velocity, v>5cm / s.

[0041] FIG. 4A depicts the result of Eq. (2) applied to 40 nm particles in droplets flowing at fD =34Hz, measured over a long period (T =104s). Data here is sampled every 100 ms (corresponding to roughly 3 droplets), and the top axis quantifies the droplet count. The ODMR contrast, C=5.6%, is marked by the dashed horizontal line in FIG. 4A. Data is displayed on an identical 1% contrast window to draw a comparison to FIG. 2C. Right panel (FIG. 4A(i)) shows this as a histogram, overlaid with the analogous histogram from FIG. 2C using the analog lock-in for clarity. The histogram linewidth in the case of FIG. 4A narrows significantly to AC=2% highlighting the enhanced measurement precision. FIG. 4A also illustrates the inherent stability in the measurement of contrast C, here over more than 2 hours and 250,000 droplets.

[0042] To delve deeper into the observed stability in FIG. 4 A, one can analyze the data as a time-domain spectrogram in FIG. 4B. This entails taking a Fourier transform of the PL over small windows A / =0.7 s, equivalent to 20 droplets, for an hour, and tracking the resulting spectra over 104,400 droplets (top axis). Colors in FIG. 4B represent Fourier spectral intensity, and the vertical position indicates frequency. The discussion focuses on frequency windows near fo and f w, using two distinct color bars for clarity. The resulting horizontal band-like patterns underscore the remarkable stability of the data. The left panel, FIG. 4B(i), shows the integrated signal across these bands. The narrow linewidths, evident even over this extended period, reflect the system stability.

[0043] Another perspective on stability is provided through an Allan deviation analysis, applied to the data in FIG. 4B and depicted in FIG. 4C. Allan deviation, cZZ(t) , provides insight into how measurement precision of C (ultimately related to analyte LOD) may improve with longer averaging time C or increasing number of droplets. Remarkably, the experimentally measured cZZ(t) closely aligns with the theoretically expected oc 1 , trend (dashed line) beyond 103s of averaging, yielding a greater than 30-fold reduction in AC (blue arrow in FIG. 4C). At the lowest point, this corresponds to detecting PL changes to a fewhundredth of a percent. FIG. 4C marks a notable improvement over previous studies, where equivalent scaling oc 1 , is challenging to obtain, and only achievable through highly sophisticated compensation strategies. One may attribute this enhanced stability to the immunity of the confined aqueous droplet volume to temperature drifts, their exposure to laser illumination only for short instants ( ^1~ 33ms), as well as compensation for laser power fluctuations and ND loading variations by the ratiometric scheme in Eq. (2).

[0044] Variability in FT intensities of the band centered at f w provides a convenient means to estimate an upper bound on the interdroplet ND variation. FIG. 4D shows histogram of the T (fMW, ti) amplitude binned every 7s (-200 droplets). When extrapolated to 103s following a-i / 2 scaling, one obtains an ND variation across droplets lower than 0.23%. In absolute terms, this corresponds to a very low droplet-to-droplet variation of .2300 NDs over the base level of ~ IM NDs per droplet.

[0045] Now, exploring the impact of potential drift mechanisms, in FIG. 4E-F, the discussion turns to assessing how laser power and ND concentration variations affect ODMR contrast C. First, in FIG. 4E the process adjusts ND concentration over a large range in the flowing droplets, using a 10-way valve to load successive samples containing different concentrations while maintaining fixed laser and MW powers. Red points in FIG. 4E show the resulting change in droplet PL, while blue points show the measured ODMR contrast C in percentage units. Despite a significant, ~500%-increase in ND concentration, ODMR contrast variation is less than AC ~11%. Given that inter-droplet ND concentration variation is <0.23% (FIG.4D), ND number variations minimally impact contrast.

[0046] Increasing laser power, with MW power held constant (FIG. 4F) affects C due to an interplay of NV center repolarization and MW-driven population shifts. However, as FIG. 4F indicates, at operational power (~100 mW), a 10% laser power variation alters ODMR contrast by only about AC=4%. In reality, laser drift is under 1%, and this leads to an insignificant effect on AC. Overall, FIG. 4E-F underscores the method’s resilience to common experimental fluctuations.

[0047] One may note that the double lock-in scheme could be performed through alternate approaches. MW modulation could be replaced by magnetic modulation, with potential advantages of a higher modulation depth (>10%). However, it lacks the precise frequency control and long-term stability of MW lock-in. The latter also benefits from lower Mf noise due to its higher operating frequency. While laser modulation, such as with an optical chopper, is a possible alternate method to fo droplet modulation, droplet flow, as introduced here, has distinct advantages: (1) immunity to autofluorescence, (2)averaging effects across droplets to reduce heterogeneity, (3) suitability for high-throughput analysis in a flowing geometry with (4) stable fo modulation for several hours, and (5) improved thermal stability of flowing versus static droplets, ensuring reliability over longtime analyses.

[0048] Leveraging the enhanced precision above, experiments detected chemical analytes in flowing droplets, starting with Gadolinium (Gd3+) ions in GdCh hexahydrate as a model system. The sensing mechanism, depicted in FIG. 5A(i) inset, relies on changes in the NV center 7i relaxation time from spin noise of paramagnetic species, affecting the ms= ± 1 population and, thus, ODMR contrast. This effect is concentration-dependent, allowing for quantitative analyte estimation. Enhanced precision via FIG. 3-FIG. 4 enables the detection of subtle contrast changes, significantly improving limits of detection (LODs).

[0049] FIG. 5A demonstrates this through Gd3+concentration titration in droplets, with each sample averaged for «4 min. Samples are automatically loaded using a 10-way valve controlled by a customized program, showcasing the potential for automated analysis. The small error bar in the measured points, AC~1% (i.e., 1% of the base contrast C), reflects the high measurement precision. FIG. 5A highlights the dynamic range over which this system can detect gadolinium ions, ranging from a 500 nM to 20 pM, above which the ionic strength of the solution causes aggregation of the carboxylated diamond particles. A separate experiment investigated the sensitivity at the low end loading 100 nM, 150 nM, and 300 nM samples in FIG. 5B. The experiment achieved a LOD of 100 nM over 2 min. of averaging.

[0050] While Gd3+has spin 5=7 / 2, most practical applications target single unpaired electronic spins (5=1 / 2). One experiment employed 4-Hydroxy-2,2,6,6-tetramethylpiperidine-1-oxyl (TEMPOL) as a second model (FIG. 5C(i)). TEMPOL is a stable radical probe that could serve as a proxy for other paramagnetic analytes, including reactive oxygen species (ROS), crucial for instance for metabolic studies in cells. FIG. 5C shows the ability todifferentiate between 500 nM and 5 pm concentrations. One estimate is an LOD of 2 pm over 1 min. of measurement.

[0051] FIG. 5D highlights how the droplet-based sensing of the embodiments requires a very small quantity of NDs. The red line shows sample volume used for varying measurement times - 1 droplet, 1 min of averaging, up to 1 hr. ND costs, at $50 / mg according to Adamas Nanotechnologies, are remarkably low, costing less than 30 for 1000 s of analysis involving ~3 X 104droplets. This benefits also from the NDs’ native hydrophilicity, which negates the need for surface treatments. The methods of the embodiments therefore not only reduce costs dramatically compared to traditional single-crystal diamond approaches requiring costly growth and surface modifications but also enhances portability. Combined with the data from FIG. 4, this highlights the capability for long-term, stable, and precise analyses at significantly reduced costs.

[0052] Finally, FIG. 5E shows an overview of the sensing technology landscape, identifying the niche filled by droplet-based quantum sensing. The experiments focused attention here on gadolinium and TEMPOL detection, and comparisons to widely used methods for ROS and spin trap detection. One can assess electron paramagnetic resonance (EPR), fluorescence, and NDs-based sensing across reported LODs and analyte sample volume. An ideal chemical sensing platform would occupy the bottom left comer of this plot. Though direct comparisons are challenging due to the diversity of implementations in each technique, the results show regions of applicability of each method, and representative references (marked points).

[0053] The various references are [1] Alert, R. D. ; et. Al. Lap Chip 2022, 22 (24), 4831-4840; [2] Nesmelov YE.; Journal of Mag. Res. 2004; 167(1): 138-146; [3] Radu, V.; et. Al. ACS Sensors 2020, 5 (3), 03-710; [4] Weil JA, Bolton JR. John Wiley & Sons; 2007 et. Al. ACS Sensors 2020, [5] (12), 3862-3869; [6] Sigaeva A.; et. al., Small. 2022; 18(44); [7]Stromquist, J et. al. Biophys J. 2010; 99(11):3821-3830; [8] Steinert, S.; et. al. Nat. Commun.2013, 4 (1), 1-6; [9] Grant ES.; et al.; Phys. Rev. Appl. 2023; 20(3).

[0054] Droplet-based sensing as described here occupies the orange region; and is projected to move along the red-dashed arrow with averaging following FIG. 4C. FIG. 5E illustrates that the method already provides significant improvements over existing technologies.

[0055] In addition to droplets containing the quantum sensors and the analyte of interest, the droplets could also act as targeted material sensors. For example, as shown in FIGs. 6A-6C show embodiments of materials that could be used to determine the presence of absence of another material.

[0056] In FIG. 6A the droplet contains a quantum sensor 54, such as a nanodiamond.Magnetic beads such as 50, have surface functionalization that is designed to attract the quantum sensors. FIG. 6B shows a closer view of a magnetic bead. The magnetic beads do not fluoresce, but when they are in contact with the particle of interest, they align in the presence of a low magnetic field, as shown in FIG 6C. The signal applied in this embodiment comprises the low magnetic field instead of a microwave signal. The motion of the quantum sensor particles in the droplets, inheriting the behavior from the beads, changes from Brownian to what is referred to as driven motion. FIG. 7 shows Brownian motion in the upper part of the image, and driven motion in the lower part of the figure. Driven motion allows detection of the targeted compound. The presence of the targeted compound therefore causes the nanodiamonds in a manner that can be easily discerned. The sample can then be identified, so measurement in this instance comprises detecting the presence of the analyte. The system of FIG. 1A may include a high-speed camera 38 to capture the motion.

[0057] The embodiments here combine droplet microfluidics with quantum sensing and introduces many new innovations, including (1) deploying ND quantum sensors within droplets, utilizing confinement and flow to (2) facilitate continuous analysis with highprecision, capable of detecting contrast changes by AC~2%, (3) with high stability across multiple hours and hundreds of thousands of droplets, (4) all while utilizing minuscule sensor volumes and entailing incredibly low ND costs.

[0058] All features disclosed in the specification, including the claims, abstract, and drawings, and all the steps in any method or process disclosed, may be combined in any combination, except combinations where at least some of such features and / or steps are mutually exclusive. Each feature disclosed in the specification, including the claims, abstract, and drawings, can be replaced by alternative features serving the same, equivalent, or similar purpose, unless expressly stated otherwise.

[0059] Additionally, this written description makes reference to particular features. It is to be understood that the disclosure in this specification includes all possible combinations of those particular features. For example, where a particular feature is disclosed in the context of a particular aspect, that feature can also be used, to the extent possible, in the context of other aspects.

[0060] Also, when reference is made in this application to a method having two or more defined steps or operations, the defined steps or operations can be carried out in any order or simultaneously, unless the context excludes those possibilities.

[0061] Although specific aspects of this disclosure have been illustrated and described for purposes of illustration, it will be understood that various modifications may be made without departing from the spirit and scope of the invention. Accordingly, the invention should not be limited except as by the appended claims.

Claims

WHAT IS CLAIMED IS:

1. A microfluidic device, comprising:an aqueous phase of analyte and nanodiamonds in an aqueous inlet;a first oil inlet;a first junction fluidically connected to the aqueous inlet and the first oil inlet, the first junction configured to cause droplets of the analyte and quantum sensors to become suspended in oil from the first oil inlet;a second junction fluidically connected to the first junction and a second oil inlet, the second junction configured to control spacing between the droplets;a microwave source adjacent the second junction operable to apply microwaves to the droplets at a frequency greater than a frequency of a flow of the droplets; anda microscope arranged to illumination the droplets and sense signals from the droplets while the microwaves are applied to perform noise-suppressed quantum sensing.

2. The microfluidic device as claimed in claim 1, the device further comprising a circuitous path fluidically connected to the second junction;3. The microfluidic device as claimed in claim 2, wherein the device further comprises a collection chamber connected to the circuitous path.

4. The microfluidic device as claimed in claim 1, the device further comprising a collection chamber connected directly to the second junction.

5. The microfluidic device as claimed in claim 1, wherein the microwave source comprises a modulated source.

6. A composition of matter, comprising:a microdroplet of fluid containing an analyte and quantum sensors.

7. The composition of matter as claimed in claim 6, wherein the quantum sensor comprises nanodiamonds.

8. The composition of matter as claimed in claim 6, wherein the microdroplet comprises one of water droplets in oil, or oil droplets in water.

9. A method of detecting an analyte, comprising:dispersing an analyte into droplets containing quantum sensors therein; controlling a flow of the droplets into a region of a microscope objective, the flow having a determined frequency of droplets;applying a signal to the region;andtaking a measurement of the analyte.

10. The method as claimed in claim 9, wherein the signal comprises a microwave signal.

11. The method as claimed in claim 9, wherein the signal comprises a magnetic field.

12. The method as claimed in claim 9, further comprising a laser applied to the analyte.

13. The method as claimed in claim 12, wherein one of either the laser or the signal is modulated.

14. The method as claimed in claim 9, wherein applying a signal to the region comprises applying a signal having a frequency of droplets and locking the frequency of the droplets and the frequency of the signal to suppress background noise.

15. The method as claimed in claim 9, further comprising:adding magnetic beads to the droplets;applying the signal to the region comprises applying a magnetic field; andtaking measurement of the analyte comprises detecting the presence of the analyte.

Images