1. Introduction

Fluorescent bulk/nanodiamonds have attracted a special interest in many promising applications including quantum communication [1,2], quantum photonics [3,4], and quantum sensing in biological sciences [5–7]. For more demanding biological applications such as drug delivery, neurons imaging, and others, bright, photostable, round shape, and small (with size less than 10 nm) fluorescent nanodiamonds (FNDs) are needed [8]. Specifically, round and small FNDs (less than 10 nm) are of special interest for biological applications because they can easily enter most of the biological tissue membranes [8,9] and perform important biological tasks such drug delivery applications as they can be cleared through membranes of most biological cells, such as kidney membranes [7,8].

Unfortunately, the optical properties of commercially available FNDs decrease drastically as a function of NDs crystal size (below 10 nm) [7,9,10]. This unexpected degradation in small FNDs is due to random distribution of diamond color centers in small FNDs crystals which makes them more susceptible to interactions with surface defects which significantly reduce their sensitivity for quantum sensing, drug delivery, and bioimaging applications [8,10]. This is because of un cleaned surface and irregular shape of the small NDs crystals which could be a direct consequence of how they were fabricated. The most diamond world made by two main ways: (1) milling large high-pressure high temperature (HPHT) diamonds [11,12]; (2) detonation nanodiamonds [13,14]. In addition to the main diamond’s synthesis approaches, NDs were fabricated by non-detonation shock wave (laser ablation) [15], high power ultrasound waves [16], chemical vapor depositions (CVD) [17], and flowing plasma [18]. Due to the growth harsh conditions associated with those growth techniques, none of these techniques provides rounded and small FNDs with desired size below 10 nm that is close to quality of bulk diamonds.

Hydrothermal growth of NDs at mild pressure and temperature is strongly proposed as a good alternative. Over more than two decades, evidence for hydrothermal growth of diamond in the $\textrm {C}-\textrm {H}-\textrm {O}$ and $\textrm {C}-\textrm {H}-\textrm {O}$ halogen system at a high temperature (higher than 1000 $^{o}C$) was reported [19]. Diamonds produced from this experiment were large with irregular and ambiguous shapes. This attempt was followed by a hydrothermal growth of diamond particles over a hydrogenated cubic boron nitrite under alkaline hydrothermal conditions [20]. Hydrothermal formation of diamond from chlorinated organic compounds was achieved producing micron size diamond with unknown yield and optical properties [21,22]. Furthermore, formation of metastable nanosized diamond from a CHO fluid system [23] and in serpentinite-hosted hydrothermal systems [24] were performed. In fact, by investigating the final product of above-mentioned hydrothermal growth techniques, the desired size and shape of NDs are not accomplished.

Recently, it was reported multiple times that lowering hydrothermal growth conditions to around 150-200 $^{o}C$ and saturated water vapor pressure can produce small and rounded diamond nanocrystals with a size less than 10 nm [25–28]. It was theoretically predicted that sub-7 nm small nanodiamonds can be grown from hydrocarbons and graphene precursors at low pressure and low temperature (LPLT) [29,30]. This prediction has been experimentally realized by synthesis of small NDs in many experiments [18,28,31,32]. Recently, it was recently shown that sub-4 nm NDs can be grown from nitrated polycyclic aromatic hydrocarbons and graphene at a very low temperature of (423 K) [27]. Also, small diamond quantum dots called "n-diamond" were synthesized from a simple hydrocarbons material (glucose) at hydrothermal growth conditions [25,26]. However, it should be noted that most of these experiments [25–27,33] do not provide convincing evidence that they actually succeeded in growing diamonds. In fact, sometimes there is only one diagnostic showing evidence of diamond production, for instance Raman spectra or transmission electron microscope (TEM) images, but not both. It should be noted that there are a number of materials having a similar Raman spectrum to diamond, and others having similar lattice spacings [27].

The main goal of this study is to repeat at least one of the low temperature hydrothermal [26] growth experiments that provided the promising NDs and prove real growth of diamond. The key proof is the observation of a fluorescent color centers which are unique to diamond, like a magnetically sensitive nitrogen-vacancy (NV) or silicon-vacancy (SiV) with its well-known optical spectrum. This is strongest technique used few years back to identify the world’s smallest high-quality nanodiamonds in meteors [34]. The NV center in diamonds has a quantum property of electron spin with long coherence time, high photostability and easy optical pumping [35,36]. The spin-dependent properties of the NV center can be optically initiated and then manipulated using optical, and microwave-frequency techniques [37,38] enable sensitive detection of electric, magnetic fields as well as small fluctuations in temperature at a nanoscale system. The SiV center has also shown a great promise as a single-photon source for quantum applications as it has a narrow zero phonon line (ZPL) peaked at 738 nm with a high photon emission rate (short excited-state lifetime at ambient temperature) [17,39]. Furthermore, the SiV can be excited and detected within the biological transparency window which makes it preferable for quantum sensing in bio applications [40–42]. In addition, other near-infrared color centers in diamonds have recently shown a promising application, especially in quantum sensing applications [7,42].

In this work, we have synthesized small NDs from a simple hydrocarbon starting materials (glucose) at low temperature (220 $^{o}C$) and the saturated vapor pressure of water. The synthesized NDs where turned to fluorescent NDs after the required implantation and post-annealing to validate that the grown small nanocrystals are indeed cubic NDs. The grown small FNDs in this experiment have shown a stable fluorescent NV and SiV color centers with exceptional optical properties compared to bulk diamonds. The bottom-up synthesis of small FNDs is greatly expected to provide opportunities to create and tailor the properties of FNDs in the future for emerging imaging and quantum sensing technologies.

2. Experimental details

2.1 Nanodiamonds hydrothermal growth

One gram of glucose was dissolved in 10 mL deionized water, then 3 mL of concentrated hydrocloric acid solution was introduced into the solution. Subsequently, the mixed solution underwent ultrasonic treatment for 10 min. Then the mixed solution was transferred to a Teflon lined inside stainless steel autoclave and placed in an oven at 220 $^{o}C$ and saturated water vapor-pressure for 24 h. After the growth is complete, the resultant solution (orange color) was transferred into a dialysis bag and kept in water bath for two weeks. After that, the clean NDs solution was placed in a glass vial and stored for further use.

2.2 Custom-made confocal microscope for optical characterizations

To analyze the optical properties of the grown NDs, a confocal laser scanning microscope was designed and built. The confocal microscope was equipped with green and red lasers (532 nm and 660 nm), high and low magnification microscope objectives (100$\times$ and 50x, NA = 0.8 and 0.65, respectively, model number: LMPlanFL N, Olympus, Tokyo, Japan), and a spectrometer as well as photon counter. The grown FNDs samples were drop casted on quartz substrates and then placed on the confocal setup. FND samples were then scanned in the $\textrm {x}-\textrm {y}$ directions by a green (532 nm) for NV centers and red (660 nm) for SiV centers lasers using Thorlabs Galvano scanners (GVS 212, Newton, NJ, USA). The optical emission spectra were collected through the same microscope objective and analyzed with a custom-made spectrometer equipped with a starlight camera (Trius camera model SX-674, Starlight Xpress Ltd., Bottle Lane, UK) and a single photon counter (Hamamatsu photon counter model number H7155-21, Hamamatsu Photon-ics UK Limited, Welwyn Garden City, UK).

2.3 Preparation NDs samples for confocal imaging and optical characterizations

Quartz slides were first rinsed with acetone to remove any oils and dirt, the slides were then placed on a heating plate inside of a fume hood. Samples from the grown NDs were then dropped onto the slide using a transfer pipette and the temperature was raised to 200 $^{o}C$. Once the samples were completely dry, they were placed in a tube furnace set to 550 $^{o}C$ and allowed to oxidize in air for 10 minutes. After 10 minutes the samples were removed and cooled to room temperature before being placed on the confocal microscope for optical analysis. Each sample was then analyzed for Raman shift using green laser and for color centers using green and red lasers investigation after implantation and required post-annealing.

2.4 Ion implantation process

The grown NDs samples which showed promising diamond Raman line were prepared for ion implantation. So, several samples were irradiated at an irradiation commercial facility (CuttingEdge Ions, LLC, USA). The samples were divided to three groups, first samples were irradiated with only nitrogen at with implantation energy 5 keV with a dose of 1 × 10$^{14}$ion/$cm^{2}$. Second group of the samples were only irradiated with silicon implantation energy 5 keV with a dose of 1 × 10$^{14}$ion/$cm^{2}$. Finally, the third samples group were co-implanted with silicon plus nitrogen with implantation energy 5 keV with at the same doses mentioned above. Post annealing was required to mobilized the created vacancies in the implanted samples. For this, appropriate post-annealing was done in vacuum at 800 $^{o}C$ for 1h to create NV centers and 1100 $^{o}C$ for 1h to create SiV color centers.

2.5 Optically detected magneto resonance (ODMR) measurments

For the ODMR measurements, the FNDs samples were attached to a microwave board and placed on the confocal microscope. After focusing on the desired optical spot (NV centers) in the x-y scan, the microwave (MW) frequencies were swept over a specific range of the microwave frequencies ranging from 2700 MHz to 3000 MHz, and the fluorescent counts were plotted vs. MW frequency.

3. Experimental results and discussion

Experimentally, Small NDs were grown from a simple hydrocarbon precursor (glucose) at low temperature and low pressure following a hydrothermal growth method previously reported in [25,26]. Briefly, a carbon source material (glucose, 1g) was dissolved in deionized water (DI water, 20 ml) and 3 ml of concentrated hydrochloric acid (HCl, 38%) were added into the growth mixture. The growth mixture was then transferred into a hydrothermal growth stainless vessel equipped with a Teflon liner. The NDs growth experiment was carried out in a dry oven for 24 hours and the growth temperature was kept constant at 220 $^{o}C$ as illustrated in Fig. 1(a). After the growth is complete, the heat was turned off and a sample of the growth mix was carefully cleaned using a dialysis bag with ultra-small membrane ports (3.5 KDa) for almost two weeks and then cleaned in boiling acid to remove extra graphitic materials. For acid cleaning, the grown NDs sample was placed in a mixture of nitric and sulfuric acids (1:1,v:v) boiling at 120 $^{o}C$ in a reflux system for 4h. After that, the sample was washed three times with DI water using a high-speed centrifuge (50 krpm).

 

Fig. 1. (a) An illustration of the concept of nanodiamond growth from hydrocarbons (glucose) at hydrothermal growth conditions and made fluorescent after appropriate irradiation and post-annealing in vacuum. (b and c) Low and high magnification images of grown nanodiamonds ( after acid cleaning) with size ranging from 1 to 7 nm with cubic-diamond diffraction pattern (111) seen in most of the crystals. (d) Corresponding FFT image confirm that the particles are cubic diamond (111).

Download Full Size | PDF

Structural and size of the grown NDs were investigated using a transmittance electron microscope (TEM). For this, few drops of the cleaned NDs solution were deposited on a carbon TEM grid and placed inside the TEM microscope. As shown in Fig. 1(b), TEM images show a well dispersed and ultrasmall NDs crystals with small size less than 7 nm. The TEM diffraction pattern of these nanocrystals shows the cubic diamond lattice spacing of (111) as shown in Fig. 1(c and d) as reported in [18].

Raman measurements were performed for the grown NDs before and after acid cleaning and air oxidation at 550 $^{o}C$. Figure 2(a) shows TEM image of a small NDs buried in graphitic materials (amorphous carbons) before cleaning process. The corresponding Raman spectrum shows strong D, G, and 2D peaks from the graphitic materials (sp$^{2}$) as illustrated in Fig. 2(b). After appropriate acid cleaning and post annealing, Fig. 2(c) shows a TEM image of a clean NDs with a size less than 10 nm. The corresponding Raman spectrum (under green excitation) showed a clear NDs Raman signal peaked at 1325 $cm^{-1}$ as shown in Fig. 2(d). The observed small NDs Raman peak after appropriate cleaning and post annealing is in a good agreement with small diamond Raman line collected from 5 nm detonation [43]. Unlike HPHT NDs, we compare our NDs with detonation NDs as they both have a graphitic shell and require cleaning and post-annealing. In comparison to bulk diamond Raman peaked at 1332 $cm^{-1}$, the diamond Raman line of the grown small NDs is blue shifted to 1325 $cm^{-1}$ due to the confinement of optical phonons in NDs which known to cause a particle size-dependent shift to lower wavenumbers and broadening of the Raman scattering peak compared to bulk diamond [18,31]. It also worths mentioning that Teflon liner is made of polymer and can be a potential precursor for nanodiamonds, especially at high temperatures similar to our NDs growth conditions. To exclude such confusion, a control experiment was carried out without using glucose as a starting material and produced some organic materials which showed no D and G peaks and quickly vanished after oxidization at low temperatures less than 200 $^{o}C$. To explain the hydrothermal growth mechanism of the small NDs with size less than 7 nm, it was theoretically predicted and experimentally realized that hydrocarbons precursors can spontaneously be converted to small nanodiamonds (up to 7-8 nm) at atmospheric pressure and relatively low temperature [18,28–32]. Since H-terminated cubic diamond is the most stable form of carbon below 7 nm sizes [28], then either self-seeding or seeding by diamond-like molecules or even seeding by very small diamonds would preferentially produce diamonds up to this size. Then, if the growth temperature is kept below the surface reconstruction temperature of 900 $^{o}C$ [28,31], the subsequent growth is expected to continue to be cubic diamond. Note that, a hydrogen-rich growth mix is also desired since atoms like oxygen catalyze the graphitization of diamond surfaces at temperatures as low as 400 $^{o}C$ [28].

 

Fig. 2. (a) TEM image of the grown NDs buried inside graphitic materials which produce strong graphite and amorphous carbon signals. (b) An optical spectrum of grown nanodiamonds (uncleaned and shown in (a)) reveals a strong D and G peaks covering the observation of the grown nanodiamond Raman peak at 1325 $cm^{-1}$. (c) TEM image of the grown NDs after acid cleaning and air oxidation. (b, inset) is an illustration of the graphitic (amorphous carbons) covering the NDs supported by a TEM image. (d) Clean Raman spectrum of the cleaned NDs with a strong Raman peak at 1325 $cm^{-1}$ which corresponds to small NDs Raman peak.(d, inset) is an illustration of cleaned NDs supported by a TEM image.

Download Full Size | PDF

Now, to provide convincing evidence that the grown small nanocrystals are real diamond, more reliable diagnostic test is required. Raman spectra or transmission electron microscope (TEM) images are known to identify crystal’s structure, however there are number of materials having a similar Raman spectrum to diamond, and others having similar lattice spacings such as copper and some iron oxides phases [27]. Therefore, creation of fluorescent color centers which is unique to diamond is required. To do this, few drops of the cleaned NDs solution was dried on several pieces of quartz and annealed to 550 $^{o}C$ for 10 min in air to remove and remaining growth parent materials which could block NDs surface from irradiation.

To create NV centers in the grown small NDs, one sample was prepared to be implanted only with nitrogen to form fluorescent NV centers as illustrated in Fig. 3(a) after appropriate high temperature annealing in vacuum to mobilize the created vacancies. The grown small NDs with size less than 10 nm have shown exceptional stability and retained their dimensions and structure under high temperature annealing at high vacuum required to mobilize color centers after implantation. This is in a good agreement with small NDs (less than 10 nm) irradiated and post-annealed at the same conditions [13,34,44]. The implantation energy and dose were chosen based on SRIM simulation software to be 5 KeV at a dose of 1X10$^{14}$ion/$cm^{2}$ as illustrated in Fig. 3(b). The low implantation energy at 5 keV was carefully chosen as this energy value is shown in Fig. 3(b) to be around the middle of the implantation-stopping range. After the irradiation process was complete, a post annealing in vacuum at 800 $^{o}C$ for 1 hour was performed to mobilize the created vacancies, as illustrated in Fig. 3(a). Optical scan of the irradiated sample showed a clear and stable NV center emission with a clear zero-phonon lines (ZPLs) peaked at 575 nm and 637 nm, respectively as shown in Fig. 3(c). Bright NV spectrum illustrated in Fig. 3(c) suppressed weak Raman signal of the grown NDs. Even stronger proof of the presence of NV centers is provided by Optically Detected Magnetic Resonance (ODMR) spectrum, as illustrated in Fig. 3(d). Briefly, ODMR presents as a decrease in NV fluorescence when a microwave excitation is scanned over a ground state spin transition involving the m$_{s}$=0 and m$_{s}$=+/-1 levels in the triplet ground state [38,41,45]. Typically, the fluorescence change is a maximum of about 30% for single NVs and 10% for ensembles, where this value is reduced to about half when there is a line splitting. Fig.3 (d) shows a typical ODMR spectrum from the implanted sample while the observed 9.6% contrast is better than expected for NV ensembles in commercial small FNDs [38,41]. The stable NV center with exceptional photoluminescence and high contrast ODMR in such small size NV center can be attributed to the presence of a lot of OH groups termination on NDs surface since glucose is rich of OH group. It was shown that NV center in 5 nm detonation NDs terminated with OH retained exceptional optical properties observed in large diamonds [13].

 

Fig. 3. (a) an illustration of the ND irradiation and annealing process to create NV centers in the grown NDs. (b) Implantation energy and dose values estimation in small NDs with size less than 10 nm calculated using SRIM free software. (c) Clear NV fluorescence emission of a representative ND after irradiation and annealing. The NV spectrum shows NV0 and NV- zero-phonon lines peaked at 575 nm and 638 nm respectively. (d) Illustrates splitting-free ODMR spectrum centered at 2875 MHz with a relatively good contrast 9.7%. This ODMR spectrum shows also a narrow line width of 15MHz which is typical for NV ensembles in high-quality bulk diamond.

Download Full Size | PDF

Another strong evidence that the grown NDs are real diamond is to create another known diamond color center, for instance, SiV center. For this purpose, another sample with NDs solution was prepared for implantation with silicon atoms in a similar way as the nitrogen implanted sample as illustrated in Fig. 4(a). Appropriate implantation energy at 5 keV and dose of 1X10$^{14}$ion/$cm^{2}$ were chosen based on SRIM software simulation as illustrated in Fig. 4(b). Implanting the small NDs crystals with silicon atoms at energy of 5 keV was pulled from Fig. 4(b) which shows the implantation depth inside such small crystals. Post-annealing in vacuum at 1100 $^{o}C$ was performed to mobilize the created vacancies to form SiV center complex in the grown NDs. The implanted small NDs retained their diamond properties (high thermal stability) without graphitization as this procedure is standard for SiV color centers activation in similar small NDs (with size less than 10 nm) reported in [39,44]. Optical characterization of the implanted NDs revealed a strong SiV center emission with its distinguished ZPL peaked at 738 nm as shown in Fig. 4(c) [39]. We observed a broad SiV emission compared to SiV sharp emission observed in bulk diamond which can be attributed to strain present in small diamond nanocrystals. The laser excitation was filtered out by 680 nm long pass filter and NDs Raman peak was observed at 723 nm under 660 nm laser excitation as illustrated in Fig. 4(c). The 738 nm peak in the optical emission spectrum of the SiV center is the characteristic zero-phonon transition between its ground and excited states which corresponds to energy at 1.681 eV. Furthermore, the negatively-charged SiV center in diamond is a point defect formed by the replacement of two adjacent carbon atoms with a single silicon atom and the detected bright, narrow-band, optical transition between its ground and excited states [39,40]. Observation of SiV center in the grown small FNDs is expected as it was previously shown that silicon-vacancy (SiV) color centers are stable in diamonds down to 1.5 nm in meteor samples [34]. Also, creation of SiV center in NDs was performed in small NDs with average size of 20 nm [46] and in small NDs with size down to 10 nm [47]. In addition, silicon implantation into such small NDs was simulated at a certain energy and doses as illustrated above in Fig. 3(b) and in agreement with similar simulation performed for small NDs reported in [46]. Furthermore, co-implantation with N and Si atoms into the grown NDs showed a clear NV center emission with known NV and SiV centers ZPLs as illustrated in Fig. 4(d).

 

Fig. 4. (a) An illustration of the silicon irradiation and post- annealing process to create SiV centers in the grown NDs. (b) Silicon implantation energy and dose estimations in the grown small NDs calculated using SRIM free software. (c) Optical emission of the negatively-charged SiV center with a clear ZPL peaked at 738 nm excited with red laser (660 nm) after appropriate irradiation and post-annealing. (d) Optical emission spectrum of the grown NDs co-implanted with N and Si atoms. The optical spectrum shows a clear NV center emission with known NV and SiV centers ZPLs.

Download Full Size | PDF

4. Conclusion

We have successfully grown ultra-small fluorescent nanodiamonds from a simple hydrocarbon precursor at hydrothermal growth conditions. The grown NDs were rounded and ultra-small with size less than 7 nm and confirmed using electron micro-scope diffraction, optical characterizations, presence of known diamond color centers. The grown small NDs were made fluorescent after suitable irradiation and post-annealing. The result was Nitrogen-Vacancy color centers showing a high contrast and splitting-free ODMR spectrum which is an indication of high-quality diamond. SiV centers were also created in the grown NDs showing a strong and sharp optical spectrum. Creation of such color centers in NDs grown at hydrothermal conditions is a key signature that prove that the grown crystals are real diamond not any other form of carbon. This innovative diamond growth technique holds promise for virtually any industrial application of diamond that can benefit from a simple and low-cost growth. The resulting diamond are also of sufficiently good quality for demanding applications like biology.