1. Introduction

The nitrogen-vacancy (NV) color center is a point defect in diamond consist of a substitutional nitrogen atom and an adjacent vacancy in the diamond lattice [1,2]. It has three charge states, two of which are fluorescent, namely neutral and negatively charge states ($NV^{o}$ and $NV^{-}$) with zero-phonon lines (ZPLs) peaked at 575 nm and 637 nm respectively [1–3]. Both charge states have exceptional long-term photostability [1–3]. The negatively charge $NV^{-}$ center is arguably the most interesting color center in diamond due to its strong fluorescence-dependence on magnetic fields. Its great potential in many promising applications in physics and biology [4] has been demonstrated. These include quantum information [3,5–12], magnetometry [13–15], super-resolution imaging [16–18] and quantum-enhanced bio-sensing [19–21].

Photo-induced ionization and recombination of the NV charge states is intensity and wavelength dependent [22]. Upon optical illumination, the NV centers undergo stochastic charge-state transitions between the negatively charge state ($NV^{-}$) and the neutral charge state ($NV^{o}$) [22,23] which depend strongly on illumination wavelength. They also depends strongly on the concentration electron donor impurities in diamond lattice in the sense that the switching away from the negative state is suppressed for high donor concentrations. In super-resolution applications, high-contrast switching between $NV^{-}$ and $NV^{o}$ is desirable to reduce the laser intensity requirements [16]. However, for magnetic sensing in biological systems, stability of the $NV^{-}$ in the presence of red illumination is highly desirable due to its lower photo-toxicity than the usual green illumination. Doping bulk diamond crystals with a high content of electron donors such as nitrogen and phosphorous was proposed to stabilize the negative charge state [24,25]. This approach shows promise for bulk diamonds, but since most fluorescent nanodiamonds (FND) are produced by crushing bulk diamonds and charge traps are produced by mechanical damage, the charge stability tends to degrade as FND size decreases. This makes it difficult to reduce the size to less than about 10 nm, and the random nature of such charge traps causes each FND to have a different charge stability.

Recently, it was demonstrated that slow growth of FNDs at low temperature can results in a high-purity material with optical and magnetic properties comparable to pure bulk diamond [26–28]. Here, we demonstrate these ultra-pure FNDs experience strong charge-state switching for red (or blue) laser excitation, and that this switching can be significantly suppressed by increasing growth temperature, so as to incorporate more nitrogen impurities.

2. Experimental details

2.1 Diamond growth mixture

In addition to the diamond-like molecule (1-Adamantylamine, (AURUM Pharmatech LLC >98%)) the main component of the growth mix consisted of an easily cracked source of carbon, namely heptamethylnonane, $((CH_{3})_{3}CCH_{2}CH(CH_{3})CH_{2}C(CH_{3})_{2}CH_{2}C(CH_{3})_{3}$ , Sigma Aldrich > 98%). Other reactants included: tetracosane $((CH_{3}(CH_{2})_{22}CH_{3})$, Sigma Aldrich > 98%) to make the growth mix solid at room temperature, a hydrazine derivative to suppress graphite formation, and a chlorinated hydrocarbon to achieve mutual solubility of the seed molecule and the other reactants. The approximate proportions of the reactants were: seed molecule (20$\mu$l) + Heptamethylnonane(200$\mu$l) + Tetracosane(2ml) + Tetramethylhydrazine(10$\mu$l). The chlorinated hydrocarbon was determined by a combination of trial-and-error and availability and consisted of a mix of chlorobenzene and chloroform. As mentioned above the role of the hydrazine derivative (tetramethylhydrazine) in the growth mixture is to prevent graphite and diamond-like carbon formation. However, it also provides a large concentration of nitrogen impurities in the growth mix.

2.2 Diamond growth setup

To optically monitor pressure during nanodiamond growth inside the diamond anvils cell (DAC), we designed a custom-built optical microscope setup. This setup is equipped with a choice of low and high magnification objectives (Mitutoyo 100x and 10x), a CCD camera for live imaging, and a custom-built spectrometer. The spectrometer consists of visible-near infrared (NIR) grating, and a pinhole (100 $\mu m$) imaged through a 4f imaging system onto a CCD camera (Trius camera model SX-674). Ruby fluorescence, Raman emission, and photoluminescence of growth mix were collected and analyzed using the spectrometer numerous times before, during, and after growth. To heat the DAC, it was inserted into a custom-built oven consisting of heater wire would around a boron nitride core, insulated with alumina and encased in steel with windows in the top and bottom plates.

2.3 Sample extraction

After diamond growth is complete, the sample was extracted from the diamond anvil cell by a needle with a 2$\mu$m tip (American Probe Technologies, Inc.). To insure a successful transfer of nanodiamonds from growth chamber, we dip the tip of the extraction needle into 200$\mu$l of an isopropanol/ethanol solution, repeating several times until the growth chamber is emptied. For TEM and optical characterization a few drops of the nanodiamond solution was placed on a lacy carbon TEM grid and quartz slides respectively.

3. Experimental results and discussion

To study the charge stability of the negatively charged NV center in slow and fast grown FNDs, we designed two diamond growth experiments following our previous work [26–28]. As seen in Fig. 1(a), a diamond-like molecule (1-adamntylamine) was mixed with more reactive hydrocarbons (Heptamethylnonane and tetracosane) at a ratio optimized to give a solid at room temperature. Tetramethylhydrazine was also introduced into the growth mixture as a graphite-suppression agent during diamond growth process. Tetramethylhydrazine will decompose at the growth temperature and give liquid nitrogen (at high pressure) which protect the growth mix from the chamber walls [29].

 

Fig. 1. (a) An illustration showing how nitrogen impurity concentration is expected to depend on growth temperature. The growth mix consisting of 1-Adamantylamine, Heptamethylnonane, Tetracosane and Tetramethyhydrazine was exposed to a high pressure (10 GPa) and then either low or high temperature in diamond anvil cell (DAC) for 24h. (b) Proposed physical mechanism for or exclusion or inclusion of nitrogen impurities in a diamond crystal at different growth speeds. (c) A schematic drawing of a confocal microscope used to observe the NVs. It is equipped with a 100x microscope objective, CW (blue, green, and red) lasers, a microwave wire, a photon counter, and a spectrometer.

Download Full Size | PDF

The growth mixture for each experiment was then placed into a gasket chamber and exposed to a high pressure (10 GPa) and variable heat using two identical diamond anvil cells (DACs). One growth experiment was carried out at relatively low temperature (400 $^{o}C$) for 24 hours and the other experiment was carried out at higher temperature (650 $^{o}C$) for the same growth period. We believe that diamond, like many crystals, tend to exclude impurities at slow growth rates but include more impurities at faster growth rates, as illustrated in Fig. 1(b). Since the growth rate depends on temperature, at low temperature the diamond is expected to exclude nitrogen as it grows. This works so well that even though the Tetramethylhydrazine in the mixture supplies large amounts of nitrogen, low-nitrogen FNDs are produced at 400 $^{o}C$ growth temperature. In contrast, high temperature diamond growth is highly expected to include much more nitrogen resulting in low-purity NDs containing many substitutional-nitrogen (P1 center) electron donors.

After diamond growth was completed, the temperature and pressure were then returned to ambient and the samples were extracted using a sharp-tip needle and dispersed in isopropanol/ethanol solution (see methods). The presence of nanodiamonds in both experiments was confirmed using a transmission electron microscope (TEM). For this, few drops of the extracted nanodiamond solutions were dropped on two different lacy carbon TEM grids and then placed on a TEM heating stage. After heating to remove unconsumed growth material we see many NDs crystals. TEM diffraction of NDs showed a crystal lattice spacing 2.06 $^{o}A$ which matches the diamond (111) lattice planes spacing reported in [30].

Optically active NV color centers are not expected to be created in diamonds grown at high pressure, so the crystals were post-irradiated with helium implantation energy 50 KeV with a dose of $2X10^{13} ions/cm^{2}$, followed by high temperature annealing in vacuum. For this purpose, a few drops of the nanodiamond solution from both experiments were placed on two different silicon chips for ion irradiation. The chips were then heated to 550 $^{o}C$for 8 minutes to drive off the volatile components that could shield the FNDs from the ions. The ion irradiation was conducted in a commercial ion irradiation facility at an energy of 50 keV at room temperature. After irradiation, post annealing at 750 $^{o}C$ for 2 hours was necessary to mobilize the vacancies and create NVs.

Optical properties of NV centers in the NDs grown at low and high temperature were investigated on a custom multi-color scanning confocal microscope as shown in Fig. 1(c). This confocal microscope is equipped with 100x (N.A. 0.90) objective, 473, 532, and 637 nm continuous wave (CW) lasers, photon counter, spectrometer, and a microwave system to study spin transitions of NV centers.

In the slow-grown (low-temperature) FND sample, we found many spots uniformly distributed on the silicon substrate. Upon green (532 nm) excitation at a power of 200 $\mu W$, the collected optical fluorescence from each spot shows a clear spectrum of the NV center emission with $NV^{0}$ and $NV^{-}$ zero-phonon lines peaked at 575 nm and 638 nm respectively as shown in Fig. 2(a). Upon blue (473 nm) excitation at the same power as green, we observed a clear depletion of the negatively charged NV center, along with sharpening of the zero-phonon lines for both, as seen in Fig. 2(b). Furthermore, Fig. 2(c) shows upon red excitation at the same, or higher, power, the negatively charged NV center emission was strongly suppressed which indicates that it was completely ionized. Note that any fluorescence from the $NV^{o}$ charge state is blocked by the low-pass filter that removes the strong laser light.

 

Fig. 2. The NV optical spectra in FNDs grown slowly, at low temperature (400 $^{o}$C), is shown in (a-c) and grown quickly, at high temperature (650 $^{o}$C) in (d-f). (a) Optical spectrum shows the characteristic NV center spectrum with $NV^{o}$ and $NV^{-}$ zero-phonon lines (ZPLs) centered at 575 nm and 637 nm, respectivily upon green (532 nm) laser. (b) Partially ionized NV center emission under blue (473 nm) laser. (c) Compeletely ionized NV- optical spectrum under red (637 nm) laser. Any NVo spectrum is blocked by a 650 nm long-pass filter which filters out the 637 nm laser. (d-f) show the corresponding optical spectra in FNDs grown rapidly at 650 $^{o}$C. The data suggests that the negtively charged NV center is stablized by electron donors (P1 center) expected to have higher concentration if the fast-grown FNDs grown at high temperature.

Download Full Size | PDF

In the fast-grown (high-temperature) FND sample, we again found many spots uniformly distributed on the silicon substrate. Upon green (532 nm) excitation at power of 200 $\mu$W, the collected optical fluorescence spectrum from each spot shows a clear spectrum of the NV center emission with $NV^{o}$ and $NV^{-}$ zero-phonon lines peaked at 575 nm and 638 nm respectively as illustrated in Fig. 2(d). Here it is seen that the ratio of $NV^{-}$ to $NV^{o}$ is similar to the slow-grown case. However, upon blue (473 nm) excitation at the same power as green, we observed very little change in the NV spectrum as shown in Fig. 2(e). This is in contrast to the slow-grown case. An even stronger difference is seen when the fast-grown FNDs were excited with red (637 nm) light at 200 $\mu W$, as seen in Fig. 2(f). Here a strong optical spectrum of NV- color center is seen, compared to essentially none for the slow-grown material.

It is theorized that the enhanced stability of the $NV^{-}$ charge state under blue and red excitations, for the fast-grown FNDs, is the result of more substitutional-nitrogen (P1) electron donors being present in the diamond. It is well-known that nitrogen donors in P1 centers have an activation energy (EA = 1.70 eV) which shallower than the acceptor level of the NV center, and therefore the P1 can easily donate electrons to $NV^{o}$, thereby changing its state to $NV^{-}$ [24,25].

As an independent check on the purity of FNDs grown at low and high temperature, Optically Detected Magnetic Resonance (ODMR) in the NV centers is used. Briefly, in ODMR the optical excitation initially pump the $NV^{-}$ into the $m_{s}=0$ spin sublevel of the triplet ground state. When a resonant microwave field induces a magnetic transition between the $m_{s}=0$ spin sublevel and the $m_{s}=\pm 1$ levels, a significant decrease of NV fluorescence results [14,31].

All of the NV color centers in slow-grown FNDs grown at low temperature show high-contrast ODMR spectra. This is shown in Fig. 3(a-b) where a high-contrast (11%) ODMR spectrum with a narrow width of 15 MHz is seen. As described in [32] , the narrow line without zero-field splitting is an indication of low P1 content in the diamond. In contrast, ODMR for fast-grown NVs at high growth temperature (see Fig. 3(c-d)) has a much lower (4.7%) fluorescence contrast and clear zero-field splitting, which is in a good agreement with commercial FNDs [33–36]. Again the larger zero-field splitting in fast-grown FNDs is attributed to local electrical fields caused by electron donors (presumably P1 centers), rather than strain as previously assumed in [32] and in a good agreement with our previous work [28].

 

Fig. 3. (a) Optical spectrum of FNDs grown slowly, at low temperature, shows the characteristic NV center spectrum with $NV^{o}$ and $NV^{-}$ zero-phonon lines (ZPLs) centered at 575 nm and 637 nm, respectively, upon green (532 nm) laser illumination. (b) The corresponding optically detected magnetic resonance (ODMR) spectrum of the FNDs described in part (a), again under a green (532 nm) excitation laser. The high-contrast (11%) and zero-field splitting free ODMR spectrum indicate high-purity FNDs which resembles good quality NVs in a bulk crystal. (c) The NV center emission of FNDs grown rapidly, at high-temperature. (d) ODMR spectrum of FNDs grown at high temperature shows lower (4.7%) contrast and clear zero-field splitting which is in a good agreement with commercial FNDs. This clear splitting in the ODMR spectrum could be attributed to local electrical fields caused by electron donors (presumably P1 centers) present in much higher concentration in fast grown (high temperature) FNDs.

Download Full Size | PDF

4. Conclusion

In this work, we demonstrated strong control over the charge-state stability of the NV color center in fluorescent nanodiamonds (FNDs) by varying growth rate (via temperature). This was done by comparing the charge stability for different excitation wavelengths for FNDs grown at both high temperature (fast growth) and low temperature (slow growth) at high pressure. Optically detected magnetic resonance (ODMR) measurements confirmed that this charge stability came from the incorporation of many more substitutional-nitrogen (P1) donors in the case of fast growth. This work opens the door to engineering of the charge stability of NVs in FNDs so as to optimize the NV for numerous applications especially those involving both super-resolution and magnetic sensing.