Local Detection of Nitrogen-Vacancy Centers in a Nanodiamond

Oct 21, 2013 - Nitrogen-vacancy defect centers (NV) contained in nanodiamonds (NDs) are a promising candidate in quantum information processing and ...
0 downloads 0 Views 3MB Size
Download PDF
Letter pubs.acs.org/NanoLett

Local Detection of Nitrogen-Vacancy Centers in a Nanodiamond Monolayer Rémy Pawlak,*,† Thilo Glatzel,† Vincent Pichot,‡ Loïc Schmidlin,‡ Shigeki Kawai,† Sweetlana Fremy,† Denis Spitzer,‡ and Ernst Meyer† †

Department of Physics, University of Basel, Klingelbergstrasse 82, 4056 Basel, Switzerland Nanomatériaux pour les Systèmes Sous Sollicitations Extrêmes (NS3E), UMR 3208 ISL/CNRS/UdS, Institut franco-allemand de recherches de Saint-Louis (ISL), 5, rue du Général Cassagnou, 68300 Saint-Louis, France

‡

S Supporting Information *

ABSTRACT: Nitrogen-vacancy defect centers (NV) contained in nanodiamonds (NDs) are a promising candidate in quantum information processing and single photon sources due to the capability of controlling their assembly on various surfaces. However, their detection with traditional optical techniques becomes challenging when probing high NV densities at the nanometer scale. Here, we combine scanning probe techniques to characterize in a monolayer the structural and electronic properties of bucky-diamonds with sizes below 10 nm. We further observe by light-assisted Kelvin- and scanning tunneling spectroscopy a clear signature of negatively charged subsurface NV centers in NDs at the nanoscale where conventional techniques are limited. KEYWORDS: Nanodiamond, NV center, bucky-diamond, scanning tunneling microscopy, atomic force microscopy, Kelvin probe force microscopy

N

peaking at 560 nm.11 With the prospect of designing nanosized quantum devices based on NV centers, their confinement in nanodiamond hosts gives an edge to control their positioning on various surfaces. The spatial detection can be then reliably done using traditional optical methods,12 but such methods are limited when approaching the nanometer scale or increasing the density of defect centers. Addressing this issue might open new strategies in the built-up of ND based technologies in the field of electronics, photonics or quantum processing informations.4 Scanning tunneling microscopy (STM), atomic force microscopy (AFM), and related spectroscopic techniques are useful tools due to their high spatial resolutions. They allow the study of various properties (electronic, structural, and mechanical) of nanostructures with atomic precision.14−16 For instance, STM has been employed to investigate subsurface defects of semiconductors,13 and Kelvin probe force microscopy (KPFM) has demonstrated the determination of the charge states of single atoms, point defects, and molecules.18−20 Here, we demonstrate that the combination of current and force detection using a tuning fork based low-temperature STM/AFM is an excellent technique for the structural and

anodiamonds (NDs) are carbon particles with sizes from two to few tens of nanometers, which already find broad applications in polishing and as lubricant materials due to their excellent material properties.1 The combination of an inert core with a functionalizable surface also makes NDs an incomparable candidate for biomedical imaging or drug delivering systems.2 Very recently, a tremendous interest has raised up on the characterization and control of color centers within diamonds as promising systems for quantum information processing,3,4 single-spin magnetometry,5−7 and single photon sources.8 Color centers are lattice point defects of a crystalinducing new optical adsorption bands. They can be produced by incorporating foreign atoms such as nitrogen, silicon, or boron during the diamond growth process. The nitrogenvacancy center (NV) is one of the popular color centers in diamond and consists of a substitutional nitrogen atom associated with a vacancy in the carbon lattice. Encountered as neutral (NV0) or negatively charged (NV−1) in chemical vapor deposited (CVD) diamond, NVs charge states can be reliably detected by photoluminescent spectroscopy.9 Furthermore, the detection of long-lived spin states of NV−1 even at room temperature3 as well as the writing and read-out of these spin states with light3,8 have attracted a considerable attention on these centers. NV−1 have a maximum photon adsorption through the ground-state triplet (A3-E3) at λ = 532 nm (E = hc/ λ = 1.945 eV), so-called the zero-phonon line (ZPL).9,10 The ZPL is also accompanied with a broad phonon sideband © XXXX American Chemical Society

Received: June 19, 2013 Revised: September 19, 2013

A

dx.doi.org/10.1021/nl402243s | Nano Lett. XXXX, XXX, XXX−XXX

Nano Letters

Letter

combining numerical calculations and photoelectron spectroscopy24,25 have suggest that for small ND particles (2−3 nm) the carbon clusters stabilize in a “bucky-diamond” fashion. These structures consist of a “diamond core surrounded by a fullerene-like carbon network” and “bare, reconstructed surfaces” that are assumed to be hydrogen-free.24 Structural similarities have been also found for small Ge or Si nanoclusters where the reconstruction of the surface is primordial to promote the sp3 nature of the core.26,27 We think that our realspace observations confirm this bucky-diamond structural model. Indeed, topographic images show the coexistence of graphitic reconstructions and diamond structure whereas the core has a diamond character according to EELS measurements. The insulating nature was also confirmed by means of scanning tunneling spectroscopy (STS) showing a band gap of ∼6 eV. The value, slightly larger than for bulk diamond (5.5 eV), is in good agreement with previous numerical calculation for NDs predicting larger band gaps while decreasing the particle size.30 It is however not yet clear how the high N doping affects the band gap and future investigations using NDs with variable doping concentration might help to clarify this issue. To locally investigate the effect of radiation on the NV centers (Figure 2a), we conducted systematic spectroscopic measurements under five illumination conditions (dark, 400, 470, 530, and 590 nm). Bias dependencies of the tunneling current I(V) and the frequency shift Δf(V) were measured on different NDs with similar tip conditions (see Supporting Information Methods). Figure 2b show typical dI/dV spectra under these illumination conditions. The dark curve which corresponds to a measurement in dark shows a band gap of 6 eV. Under irradiation of the NDs with λ = 400 nm and λ = 470 nm while keeping the tip−sample separation constant, a systematic change of the dI/dV spectra (∼150 mV) is observed at the conduction band (inset of Figure 2b). Wavelengths larger than λ= 530 nm do not induce such changes (red and green curves Figure 2b). For comparison, STS spectra were also acquired on clean HOPG without showing relevant changes under identical illumination (not shown). The variation of the conduction band observed by STS under illumination implies that these nanostructures are n-type wide band gap semiconductors. The maximum photon energy used here (3 eV, λ = 400 nm) is insufficient to directly transfer electrons from the valence (VB) to the conduction band (CB) of ND. Hence, photoelectrons populating this band are irremediably coming from the nitrogen impurities (donor) contained in the ND core. Under sub-band gap illumination, a charge transfer from the filled states of these impurities lying in the ND gap to the CB occurs (Figure 2c).31 Filling of the CB then increases the local density of states (LDOS) that induces the shift in the STS spectra. At λ = 530 and 590 nm, this effect disappears since the charge transfer is forbidden. Although these states are present in the band gap, the large tip−sample separations utilized during the experiment and the weak contribution of the donor states to the STS spectra do not allow their observation here. This point will be discussed later. To confirm the photoionization of NV centers, we investigated the variation of the local contact potential difference (LCPD) between tip and sample under illumination by KPFM spectroscopic measurements. Figure 2d shows typical Δf(V) parabolas measured in dark (black curve) and λ = 400 nm (purple curve) for which the LCPD value is given by the parabola maxima. The shift in the LCPD between them is ∼190

electronic characterization of NDs at the atomic scale. Spectroscopic measurements under illumination further allow the detection of the subsurface NV centers in a reliable and systematic way. Figure 1a shows the derivative of a constant-current STM image of highly nitrogen-doped nanodiamonds deposited on

Figure 1. Surface structure of nanodiamonds. (a) Derivative STM image of the nanodiamond monolayer deposited on graphite by electrophoresis. The surface consists of graphitic-like defects (green area) and crystalline facets (red areas) respectively, as expected for the bucky-diamond structure, I = 60 pA, V = 900 mV. (b) STM topographic image and (c) its derivative image revealing crystalline facets of C(111) with atomic resolution, I = 10 pA, V = 900 mV.

HOPG by electrophoresis21 (see Supporting Information Methods). Prior to measurements, nanodiamonds were characterized by electron energy loss spectroscopy (EELS)22 confirming their diamond character as well as the nitrogen doping of ∼3 wt % within their core. According to the STM images, the NDs form a closed-packed monolayer and their diameters vary from 2 to 5 nm. This size distribution23 is also responsible of the large corrugation of the topographic image (Supporting Information Figure S1). At the atomic-scale, each nanostructure reveals at its surface crystalline facets (red areas) and graphitic-like structures (green area). The derivative image in Figure 1b shows the ND surface without carbon defects with an hexagonal periodicity of ∼0.2 nm, which can be attributed to relaxed surfaces of C(111) or graphene. Note also that the (100) facets were not been clearly observed during the experiments.28,29 Interestingly, previous studies of Raty et al. B

dx.doi.org/10.1021/nl402243s | Nano Lett. XXXX, XXX, XXX−XXX

Nano Letters

Letter

Figure 2. Charge transfer from the NV centers. (a) Schematic of a NV center in diamond. (b) dI/dV spectra taken above a single nanodiamond while illuminating the sample with five different wavelengths. Inset of (b) magnified dI/dV curves of both dark and illuminated conditions showing a shift of the conduction band of about 200 meV under 400 and 470 nm radiations. (c) Band diagram of NDs where the NV donor states allows under illumination (λ ≤ 470 nm) the excitation of electrons to the conduction band, (d) Δf(V) in dark condition (black curve) and under illumination (λ = 400 nm, purple curve). Both measurements are performed at the same relative tip−sample distance by opening the STM feedback at I = 10 pA, V = −3.1 V. Maxima of the Δf(V) curves (LCPD) are extracted with parabolic fits (superimposed curves in black). A shift of 190 mV ± 5 mV is clearly observed confirming the photoionization process under 400 nm irradiation. (e) LCPD values as a function of the tip−sample distance above a single diamond under different light conditions. In all graphs, colors correspond to the wavelengths (purple = 400 nm, blue = 470 nm, green = 530 nm, and red = 590 nm).

at the subsurface NV centers obtained under illumination. This LCPD change can thus be attributed to the creation of a local surface photovoltage (LSPV). At close tip−sample distance (z ≈ 0 nm), the difference of the LCPD values, Δ(LCPD), observed between λ = 400 and 470 nm is ∼50 mV ± 5 mV and arises from different photoionization processes of the NV centers. This LCPD difference can be interpreted as a variation of the charge transfer efficiency depending on the wavelengths. It might also reveal the variation of the NV−1 compared to NV0 ratio in the ND core which varies with respect to the irradiation.11 When illuminating at λ = 590 and 530 nm (red and green curves), a positive variation of the LCPD values is also observed whereas STS measurements do not show any charge transfers to the conduction band (Figure 2b). Here, the LCPD variation compared to dark is induced by the continuous photoexcitation of electrons through the A3-E3 electronic transition of the NV−1 centers. Although the photocharges are localized at the NV−1, they are detected as LCPD variation toward more positive values. The ΔLCPD between these wavelengths at z = 0 nm is ∼70 ± 5 mV. Close to the ZPL (λ = 530 nm),9,10 the maximum of electronic transition is obtained that induces the maximum LCPD shift compare to the situation in dark (∼190 ± 5 mV). For λ = 590 nm, which corresponds to the phonon sideband, the A3-E3 transition efficiency is decreased which results in a smaller LCPD shift compared to the ZPL (λ = 530 nm). Although no electrons are transferred to the ND conduction band with these wavelengths, the photoexcitation of electrons

mV. Such lateral displacement of the parabola maximum has been shown to be induced by the influence of localized charges in the vicinity of the surface,17,19 whereas the vertical shift of the Δf(V) is related to the stronger electrostatic force contribution arising from these charges. 32 Thus, these observations are consistent with the photoexcitation of the defect centers of NDs inducing the transfer of charges to the CB. Figure 2e shows the LCPD variations versus the tip−sample distance under the various illumination conditions (Supporting Information). In dark (dark curve) and at large tip−sample separation (z = 2 nm), the extracted LCPD value is ∼−300 mV. When entering the short-range regime of the electrostatic forces by decreasing the tip−sample distance, the LCPD decreases to −350 mV as a consequence of the appearance of charge images at the tip apex.32 Illuminating n-type NDs in which the band bending is upward should induce the compensation of the intrinsic surface band bending toward flat band condition. Consequently, a decrease of the work function of the NDs is expected under illumination.33 This created surface photovoltage (SPV) should then be detected as more negative CPD values during Kelvin measurements. This effect is however not observed during measurements due to the small intensity of light utilized.34 In contrast to this classical picture of SPV, the detection of localized charges in the vicinity of the surface changes the local work function and therefore induces LCPD variations.17,19 Experimentally, we observed such variation at close tip−sample distances (z ≤ 1.5 nm) due to the detection of negative charges C

dx.doi.org/10.1021/nl402243s | Nano Lett. XXXX, XXX, XXX−XXX

Nano Letters

Letter

Figure 3. Conductance spectroscopy of the NV−1 states under illumination. (a) and (b) I(V,z) spectra in dark and λ = 530 nm illumination, the z variation between curves is ∼15 pm. (c,d) corresponding I/V curves where all z distances are superimposed. (e) Comparison of dI/dV spectra at the closest tip−sample distance (z = 0 nm) obtained in dark (black curve) and 530 nm conditions (green curve). The electronic states of NV−1 centers appears under illumination with a HOMO−LUMO gap of about 1.8 eV coinciding with the A3-E3 transition. (f) Proposed band diagram of the nanodiamonds, the colored bars show the irradiation energies utilized during our experiment. STM feedback interrupted at I = 10 pA, V = −1.8 V, A = 8 mV, f 0 = 400 Hz.

at the NV−1 centers contained in the ND core can be reliably detected by the Kelvin spectroscopy method. To further confirm the presence of NV−1 centers in NDs, Figure 3 shows I(V) curves obtained in dark and for λ = 530 nm, respectively. Figure 3a,b shows the evolution of the conductance curves versus the tip−sample separation z, I(V,z), and Figure 3c,d, the corresponding I(V) plots where all z separations are superimposed. A comparison of these curves clearly reveals the appearance of electronic states when illuminating the sample (Supporting Information Figures S2 and S3). The dI/dV curves extracted at the closest tip−sample distance (Figure 3e) of both dark (black curve) and under illumination (green curve) confirm the presence of NV−1. The highest occupied molecular orbital (HOMO) of the NV−1 centers lies at ∼−1.2 eV below the Fermi level, and the lowest unoccupied molecular orbital (LUMO) at ∼0.6 eV. The HOMO−LUMO gap is ∼1.8 eV and coincides with the experimental ZPL of the NV−1 (1.945 eV)9,10 or theoretically predicted (1.75−1.85 eV).3 Although we cannot exclude the presence of other N centers, this observation points to the presence of NV−1 states in the ND core and allows a quantitative estimation of the HOMO−LUMO gap of the NV−1 centers. STM/KPFM-based measurements thus allow the detection of NV−1 centers and give a clear picture of the band diagram of single NDs (Figure 3g). We demonstrated the atomic-scale structural characterization of detonation NDs, showing a bucky-diamond structure, as well as the determination of their electronic properties (∼6.5 eV band gap). Interestingly, subsurface nitrogen-vacancy centers of the ND core can be also detected at the nanometer scale by combined current and force spectroscopic measurements under illumination. This particular asset might help for investigating nanosized quantum devices based on nanodiamonds up to

scales where usual optical techniques are limited. We also foresee that SPM-assisted studies might be helpful to elucidate the fundamental magnetic properties of color centers in semiconductors.

■

ASSOCIATED CONTENT

S Supporting Information *

The Material and Methods section, as well as the Supplementary Discussion and Figures S1−S2 are included. This material is available free of charge via the Internet at http:// pubs.acs.org.

■

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Notes

The authors declare no competing financial interests.

■

ACKNOWLEDGMENTS We acknowledge financial support from the Swiss National Science Foundation (NSF), the ESF EUROCORE program FANAS and the Swiss National Center of Competence in Research on “Nanoscale Science” (NCCR-NANO).

■

REFERENCES

(1) Mochalin, V. N.; Shenderova, O.; Ho, D.; Gogotsi, Y. Nat. Nanotechnol. 2011, 7, 11−23. (2) McGuinness, P. L.; Yan, Y.; Stacey, A. A.; Simpson, D. T.; Hall, L.; Maclaurin, D.; Prawer, S.; Mulvaney, P.; Wrachtrup, J.; Caruso, F. E.; Scholten, R. L.; Hollenberg, L. C. Nat. Nanotechnol. 2011, 6, 358− 363. D

dx.doi.org/10.1021/nl402243s | Nano Lett. XXXX, XXX, XXX−XXX

Nano Letters

Letter

(3) Weber, J. R.; Koehl, W. F.; Varley, J. B.; Janotti, A.; Buckley, B. B.; Van de Walle, C. G.; Awschalom, D. D. Proc. Natl. Acad. Sci. U.S.A. 2010, 107, 8515−8518. (4) Wrachtrup, J. Proc. Natl. Acad. Sci. U.S.A. 2010, 104, 9474−9480. (5) Balasubramanian, G.; Chan, I. Y.; Kolesov, R.; Al-Hmoud, M.; Tisler, J.; Shin, C.; Kim, C.; Wojcik, A.; Hemmer, P. R.; Krueger, A.; Hanke, T.; Leitenstorfer, A.; Bratschitsch, R.; Jelezko, F.; Wrachtrup, J. Nature 2008, 458, 648−651. (6) Maletinsky, P.; Hong, S.; Grinolds, M. S.; Hausmann, B.; Lukin, M. D.; Walsworth, R. L.; Loncar, M.; Yacoby, A. Nat. Nanotechnol. 2012, 7, 320−324. (7) Waldherr, G.; Beck, J.; Neumann, P.; Said, R. S.; Nitsche, M.; Markham, M. L.; Twitchen, D. J.; Twamley, J.; Jelezko, F.; Wrachtrup, J. Nat. Nanotechnol. 2012, 7, 105−108. (8) Aharonovich, I.; Castelletto, S.; Simpson, D. A.; Alastair Stacey, A.; McCallum, J.; Greentree, A. D.; Prawer, S. Nano Lett. 2009, 9, 3191−3195. (9) Davies, G. Nature 1976, 269, 498−500. (10) Mita, Y. Phys. Rev. B 1996, 53, 11360−11364. (11) Manson, N. B.; Harrison, J. P. Diamond Relat. Mater. 2005, 14, 1705−1710. (12) Bradac, C.; Gaebel, T.; Naidoo, N.; Sellars, M. J.; Twamley, J.; Brown, L. J.; Barnard, A. S.; Plakhotnik, T.; Zvyagin, A. V.; Rabeau, J. R. Nat. Nanotechnol. 2010, 5, 345−349. (13) Koenraad, P. M.; Flatte, M. E. Nat. Mater. 2011, 10, 91−100. (14) Gross, L.; Mohn, F.; Moll, N.; Liljeroth, P.; Meyer, G. Science 2009, 325, 1110−1114. (15) Pawlak, R.; Kawai, S.; Fremy, S.; Glatzel, T.; Meyer, E. ACS Nano 2011, 5, 6349−6354. (16) Pawlak, R.; Fremy, S.; Kawai, S.; Glatzel, T.; Fang, H.; Fendt, L.A.; Diederich, F.; Meyer, E. ACS Nano 2012, 7, 6318−6324. (17) Gross, L.; Mohn, F.; Liljeroth, P.; Repp, J.; Giessibl, F. J.; Meyer, G. Science 2009, 324, 1428−1431. (18) Barth, C.; Henry, C. R. Phys. Rev. Lett. 2007, 98, 136804. (19) Köning, T.; Simon, G. H.; Rust, H.-P.; Pacchioni, G.; Heyde, M.; Freund, H.-J. J. Am. Chem. Soc. 2009, 131, 17544−17545. (20) Walch, H.; Leoni, T.; Guillermet, O.; Langlais, V.; Scheuermann, A.; Bonvoisin, J.; Gauthier, S. Phys. Rev. B 2012, 86, 075423−1/7. (21) Schmidlin, L.; Pichot, V.; Josset, S.; Pawlak, R.; Glatzel, T.; Kawai, S.; Meyer, E.; Spitzer, D. Appl. Phys. Lett. 2012, 101, 253111− 253115. (22) Pichot, V.; Stephan, O.; Comet, O.; Fousson, E.; Mory, J.; March, K.; Spitzer, D. J. Phys. Chem. C 2010, 114, 10082−10087. (23) Pichot, V.; Risse, B.; Schnell, F.; Mory, J.; Spitzer, D. Sci. Reports 2013, 3, 1−6. (24) Raty, J.-Y.; Galli, G. Nat. Mater. 2003, 2, 792−795. (25) Raty, J.-Y.; Galli, G.; Bostedt, C.; van Buuren, T. W.; Terminello, L. J. Phys. Rev. Lett. 2003, 90, 037401−037405. (26) Williamson, A. J.; Bostedt, C.; van Buuren, T. W.; Willey, T. M.; Terminello, L. J.; Galli, G.; Pizzagalli, L. Nano Lett. 2004, 4, 1041− 1045. (27) Puzder, A.; Williamson, A. J.; Grossman, J. C.; Galli, G. Phys. Rev. Lett. 2002, 88, 097401−097405. (28) Bobrov, K.; Mayne, A. J.; Dujardin, G. Nature 2001, 413, 616− 619. (29) Nimmrich, M.; Kittelmann, M.; Rahe, P.; Mayne, A. J.; Dujardin, G.; von Schmidsfeld, A.; Reichling, M.; Harneit, W.; Kühnle, A. Phys. Rev. B 2010, 81, 201403−1/4. (30) Sasagawa, T.; Shen, Z.-X. J. Appl. Phys. 2008, 104, 073704−1/4. (31) Bonnell, D.; Rohrer, G. S.; French, R. H. J. Vac. Sci. Technol. B 1991, 9, 551−558. (32) Bocquet, F.; Nony, L.; Loppacher, C. Phys. Rev. B 2011, 83, 035411−1/13. (33) Rezek, B.; Nebel, C. E. Diamond Relat. Mater. 2005, 14, 466− 469. (34) Kronik, L.; Shapira, Y. Surf. Sci. Rep. 1999, 37, 1−206.

E

dx.doi.org/10.1021/nl402243s | Nano Lett. XXXX, XXX, XXX−XXX