Practical Applications of Quantum Sensing: A Simple Method to Enhance the Sensitivity of Nitrogen-Vacancy-Based Temperature Sensors

Nitrogen-vacancy centers in diamond allow measurement of environment properties such as temperature, magnetic and electric fields at the nanoscale level, of utmost relevance for several research fields, ranging from nanotechnologies to biosensing. The working principle is based on the measurement of the resonance frequency shift of a single nitrogen-vacancy center (or an ensemble of them), usually detected by monitoring the center photoluminescence emission intensity. Albeit several schemes have already been proposed, the search for the simplest and most effective one is of key relevance for real applications. Here we present a continuous-wave lock-in-based technique able to reach high sensitivity in temperature measurement at microscale or nanoscale volumes (4.8 ${\mathrm{mK}/\mathrm{Hz}}^{1/2}$ in $\mu {\mathrm{m}}^3$). Furthermore, the present method has the advantage of being insensitive to environmental magnetic noise that in general introduces a bias in the temperature measurement.

Figure 1

Optically detected magnetic resonance spectra: orientating the magnetic field $\mathbf{B}$ orthogonally to the N-V axis suppresses hyperfine contribution resulting in increased contrast and reduced linewidth of the resonant peak and hence in enhanced sensitivity.

Figure 2

Energy level diagram for axial and transverse magnetic field.

Figure 3

Schematic of the experimental setup. (a) Permanent magnet. (b) Microwave planar ring antenna. (c) CVD diamond sample. (d) Thermocouple. (e) Coil. (f) Green excitation laser. (g) Temperature-controlled chamber. (i) Red photoluminescence. (l) Photodetector.

Figure 4

Schematic of the lock-in amplifier (LIA) technique. The frequency of the microwave excitation $f_{\mathrm{MW}}$ is modulated by a wave of amplitude $f_{\mathrm{dev}}$ and frequency $f_{\mathrm{mod}}$. The resulting modulated photoluminescence signal, modulated at frequency $f_{\mathrm{mod}}$, is detected by a LIA. A temperature shift or an applied magnetic field induces a shift in the resonance frequency and hence a variation in the lock-in signal. In the inset is shown how the LIA spectrum derives from the ODMR spectrum. The red lines represent the modulated photoluminescence signal for different $f_{\mathrm{MW}}$. The LIA signal is greater at points of maximum derivative of the ODMR signal and it is null at the resonance of the ODMR signal.

Figure 5

Response of the N-V sensor (black line) in the proposed regime compared to the readout of a standard thermocouple (red line) to repeated thermal cycles. The period of the repetition is 1800 s and the temperature variation is 1 K. (a),(c) The TF regime, showing almost perfect agreement of the measurement values; (b),(d) SHfD. The lack of agreement between the two traces in the latter case is attributed to environmental field fluctuations. Panels (c),(d) are the enlargements of panels (a),(c), where the enlarged zone is defined by the blue dashed rectangle. The error bars are smaller than the experimental points in (a),(b), but they can be appreciated in (c),(d).

Figure 6

Comparison between the linear spectral densities of the readout of the N-V sensor in the TF regime (red line) and in the SHfD case (blue line). The dashed (continuous) lines correspond to the application (disabling) of an oscillating (25 Hz) magnetic field to the sample. It is apparent from the continuous lines that an improved sensitivity (lower than 5 ${\mathrm{mK}/\mathrm{Hz}}^{1/2}$) is achieved with our method. From the dashed lines it is demonstrated that with our technique the peak due to the 25-Hz magnetic field is one order of magnitude less prominent in our regime, proving the reduced sensitivity to magnetic noise.

Figure 7

LIA spectra for standard regime (a) and transverse field regime (b) with and without the application of an additional field ($\approx 1\mu \mathrm{T}$) along the (100) direction. The lock-in time constant is set to $\tau =100$ ms.