Highly sensitive cantilever-enhanced photoacoustic detection of hydrogen cyanide and methane in the mid-infrared region is demonstrated. A mid-infrared continuous-wave frequency tunable optical parametric oscillator was used as a light source in the experimental setup. Noise equivalent detection limits of 190 ppt (1 s) and 65 ppt (30 s) were achieved for HCN and CH4, respectively. The normalized noise equivalent absorption coefficient is 1.8 × 10−9 W cm−1 Hz−1/2.
© 2013 OSA
Trace gas monitoring by laser absorption spectroscopy is important for many scientific and industrial purposes. Detection limits of parts per million (ppm) to parts per trillion (ppt) can be achieved by various techniques, such as tunable diode laser absorption spectroscopy (TDLAS) , cavity enhanced absorption spectroscopy (CEAS) [2–5], and photoacoustic spectroscopy (PAS) [4,6]. Tunable diode lasers (TDLs) are workhorses for many of these techniques, because these are easy to use and inexpensive. However, these lasers are not available at all wavelength regions, which limits the number of detectable species and the achieved sensitivity. Especially, in the mid-infrared (MIR) region, the lack of tunable laser sources has been a challenge for a long time, although there has recently been great progress in the tunable quantum cascade laser (QCL) industry, which has increased the use of these setups in the MIR spectroscopy [7–9]. The MIR region contains strong fundamental ro-vibrational transitions (for example, CH, NH, and OH stretching modes) of various molecules. The strong absorption and the large number of detectable molecules make this region ideal for spectroscopy. Continuous-wave optical parametric oscillators (cw-OPOs) have been widely used as light sources for the MIR spectroscopy [10,11]. The OPOs possess wide tunability and watt-level output power, particularly wavelength regions below 4.0 µm, where QCL manufacturing has proven to be very challenging. Especially, the high output power is ideal for photoacoustic spectroscopy, because the intensity of the acoustic signal is directly proportional to the optical power. Thus, PAS combined with a tunable narrow-line mid-infrared OPO provides a versatile technique to measure trace gases with extremely high sensitivity and selectivity [12,13].
Photoacoustic spectroscopy is a straightforward technique, where the measured photoacoustic (PA) signal arises from a non-radiative relaxation of excited molecules. Periodically induced absorption creates an acoustic wave, which can be detected by a sensitive pressure sensor, e.g., a microphone. There are various techniques to increase the sensitivity of PAS. As an example, the acoustic signal can be enhanced by using a resonant cell [14,15], although the resonant condition limits the size of the cell and prevents small sample volumes. The PA measurement is strongly affected by ambient noise inside the cell. The 1/f-dependent noise can be reduced by using a wavelength-modulation technique [16,17], where the detection frequency is shifted to the higher harmonics of the modulation frequency. To increase the optical power, PA measurements have been performed, for example, inside the titanium sapphire ring laser cavity .
Essentially, in all of the PAS approaches discussed above, the sensitivity is limited by the microphone. The sensitivity of conventional electret or capacitive microphones is determined by the movement of the metal coated membrane of the microphone. The capacitive measurement itself disturbs the moving membrane by a damping effect, thus decreasing the sensitivity. In addition, the response of the membrane is non-linear with high pressure variations. In the last decade, novel sensors have been suggested to increase the sensitivity of the PAS measurement [19, 20]. Kosterev et al.  introduced a quartz tuning fork as a resonant element to obtain the acoustic signal. The lowest normalized noise equivalent absorption (NNEA) reported to date with quartz-enhanced photoacoustic spectroscopy (QEPAS) is 2.7 × 10−10 W cm−1 Hz−1/2, which was obtained by measuring SF6 in the MIR . The QEPAS enables the use of small sample volumes, but the effective absorption path length is limited to a few millimeters, owing to the high resonant frequency of a typical quartz tuning fork .
A different novel approach, introduced by Kauppinen et al. [20,22], is based on an extremely sensitive miniature silicon cantilever microphone. In cantilever-enhanced photoacoustic spectroscopy (CEPAS), the acoustic signal is detected by measuring the movement of the cantilever using a laser interferometer. The movement of the cantilever is about two orders of magnitude larger than that of a conventional membrane microphone, which increases significantly the sensitivity of the system. The response of the cantilever is also more linear, because the cantilever is connected to its frame just on one side, and thus only the bending motion is excited. Previously, cantilever-enhanced photoacoustic detection has been combined with black body radiation sources [22,23], tunable diode lasers [24–26], MIR light-emitting diodes (LEDs) , and QCLs . The best reported NNEA (3σ) with CEPAS spectroscopy is 3.4 × 10−10 W cm−1 Hz−1/2, which was achieved by measuring carbon dioxide at 1568.78 nm using a TDL with an optical power of 2.1 mW .
We report here a new implementation of a CEPAS system using a mid-infrared cw OPO as a light source. The sensitivity and stability of the setup are investigated by measuring hydrogen cyanide (HCN) and methane (CH4) rovibrational spectra between 3 and 3.3 µm.
2. Experimental setup
The experimental setup used in this work is schematically shown in Fig. 1. The design of the singly resonant continuous-wave OPO is similar to that presented in our previous papers [29,30]. The pump laser of the cw-OPO is a Yb fiber laser system operating at a wavelength of 1064 nm. The pump laser system has a maximum output power of 15 W. The pump beam is focused into an MgO-doped, periodically poled lithium niobate (MgO:PPLN) crystal. The nonlinear crystal is placed in a copper holder, the temperature of which can be stabilized anywhere between 20 °C and 200 °C with a precision of ± 10 mK. The two poling periods, 30.5 μm and 31 μm, used in this work, allow idler tuning between 3.0 and 3.4 μm (3330 to 2950 cm−1). Coarse tuning of the OPO wavelength within the tuning range of a single period can be done by changing the temperature of the crystal. Fine tuning and scanning of the OPO frequency are done by tuning the pump laser frequency with a piezoelectric actuator that controls its cavity length. With slow (~1 Hz) piezo scanning rates, the mode-hop-free tuning range of the pump laser is ~100 GHz.
The idler beam exiting the OPO cavity is collimated and separated from the residual pump and signal beams using a dichroic mirror. The single-mode output power of the idler is over 0.5 W. The linewidth of the beam is of the order of 1 MHz. A small fraction of the OPO power is directed to a wavemeter (EXFO, WA-1500). The rest of the power is focused through a commercial photoacoustic analyser (PA201).
Gasera Ltd., Turku, Finland, has manufactured the analyzer, which is equipped with a sensitive cantilever-enhanced PAS detector. The dimensions of the cantilever are 5 mm × 1.2 mm × 10 µm (length, width, and thickness). The PA cell is 95 mm long and 4 mm in diameter with a total volume of about 7 ml. The walls of the cell are gold coated. The analyzer is equipped with an automatic gas exchange and a temperature controller. The pressure of the cell can be reduced to 400 mbar. The movement of the cantilever is measured with a built-in laser interferometer. The PA201 provides an analog output signal proportional to the cantilever movement. Measurement software (Gasera Ltd.) is used for data acquisition via USB. The signal data rate is 16.8 Hz.
To obtain a photoacoustic signal, the OPO wavelength is modulated at the frequency f = 70 Hz, whereas the PA signal is detected at the frequency 2f. The wavelength modulation is done through the pump laser by applying a sinusoidal voltage to a piezoelectric actuator that controls its cavity length. In addition, to obtain a spectrum, a slow (f = 10 mHz) triangular voltage signal is applied to the piezoelectric actuator in order to scan the OPO back and forth over the wanted spectral region. The spectral region from 3331.30 to 3331.85 cm−1 was selected for HCN measurements. This region contains four rovibrational HCN (C-H stretching) absorption lines. The methane measurements were done in the spectral region from 3057.50 to 3057.90 cm−1. This region contains three strong rovibrational absorption lines of the ν3 vibrational band (C-H antisymmetric stretching). The spectral parameters of the measured lines are presented in Table 1. The measurements were recorded at a sample pressure of 400 mbar (~0.4 atm). The temperature of the cell was stabilized to 50 °C. The gas samples were mixed from a hydrogen cyanide standard of 10.2 ppm (AGA) and a methane standard of 5 ppm ± 2% (AGA) with nitrogen gas using two mass flow controllers (AREA, FC-785C). The total flow rate was ~1 liter per minute. The concentration of the hydrogen cyanide standard was verified in a separate cavity-ring down spectroscopy experiment .
3. Results and discussion
The performance of the OPO-CEPAS system was studied by measuring mid-infrared spectra of HCN and CH4. Although the main motivation of this work is merely to demonstrate the high potential of the OPO-CEPAS technique, HCN and CH4 are good test molecules owing to their importance in many applications of trace gas detection. Hydrogen cyanide is a highly toxic compound, which is used and released in many industrial processes. It is also one of the lethal fire gases, which is formed during the combustion of nitrogen-containing compounds. Despite its high toxicity, small amounts of HCN can be found in the human body due to endogenous production, bacteria, or ingestion of certain foods . Methane is an important trace gas in the atmosphere because of its effect on global warming . Methane has also a significant role in tropospheric chemistry [34,35].
Figure 2(a) shows a linear behavior of the PA signal of HCN as a function of the OPO power. In Fig. 2(b), the PA signal recorded at the center frequency of an HCN transition at 3331.584 cm−1 is presented as a function of the HCN concentration. The HCN concentration was varied between 8 ppb and 320 ppb. A detection limit (1σ) of 190 ppt was reached at an averaging time of one second with the presence of 35 ppb of HCN. The detection limit is extrapolated from the signal to noise ratio (SNR = 1), where the noise of the PA signal is measured at the center of the HCN transition with an OPO power of 0.5 W. This yields an NNEA of 1.8 × 10−9 W cm−1 Hz−1/2. For comparison, a detection limit (1σ) of 155 ppb (NNEA of 4.3 × 10−9 W cm−1 Hz−1/2) with a 1 s sensor time constant has been previously reported using QEPAS with a 50 mW telecom diode laser measuring an HCN line at 6539.11 cm−1 . It should be mentioned that the modulation depth was not optimized for our measurement conditions. With the 70 Hz modulation frequency used in the measurements, the achievable modulation amplitude is limited by the movement of the piezoelectric actuator of the pump laser system. The measured half width at the half maximum (HWHM) of the transition at 3331.58 cm−1 was 0.0430 cm−1 (in good agreement with the HITRAN value 0.0431 cm−1 ), while a maximum modulation amplitude of 0.0340 cm−1 was achieved at 70 Hz. The PA signal would be approximately 1.8 times stronger with the optimum modulation amplitude of ~2.2 × HWHM . Taking this into account, the smallest NNEA (1σ) would be 1.0 × 10−9 W cm−1 Hz−1/2. The fundamental detection limit of the photoacoustic analyzer is given by the lowest noise level achieved with the OPO beam blocked, which results in an NNEA (1σ) of 5.3 × 10−10 W cm−1 Hz−1/2.
Figure 3(a) shows a spectrum of 17 ppb HCN sample measured in a scanning mode with a single sweep of the OPO frequency from 3331.30 to 3331.85 cm−1. The OPO frequency was scanned with a speed of 0.012 cm−1/s (0.360 GHz/s). A derivative of a Lorentzian line shape was fitted to the spectrum. The fit residual, which is the difference between the measured points and the fitted Lorentz line shape, is shown below the spectrum. The standard deviation (1σ) of 0.30 mV in the residual gives a detection limit of ~720 ppt (60 ms integration time per a spectral point). Figure 3(b) shows a spectrum of a 2.1 ppm HCN sample, where four Lorentz line shapes were fitted. When the concentration (and hence the absolute signal level of the PA signal) is increased, the OPO power fluctuations become the dominant noise source in the measurement system. This can be seen as increased noise especially in the center of the absorption peak (see the residual in Fig. 3(b)). The behavior of the signal to noise ratio as a function of a PA signal level is illustrated in Fig. 4. The SNR is calculated as the ratio of the peak-to-peak value of the signal and the standard deviation of the fit residual. Within the region between 0 and 20 mV (0 and 50 ppb for the measured HCN line), the effect of OPO intensity noise is small or comparable with other noise sources, and the SNR increases linearly. At higher concentrations, the SNR saturates at around 200. Since the photoacoustic signal itself increases linearly versus the HCN concentration, this behavior is not critical for the performance of the measurement system.
Strong adsorption of HCN on the gold coated surfaces of the photoacoustic cell prevents long measurements. The HCN concentration decreases with time due to adsorption [33,38]. Because of this, we used methane to determine the optimum averaging time of the OPO-CEPAS system. The selected wavelength region for methane measurements, from 3057.50 to 3057.90 cm−1, covers three strong overlapping absorption lines, which complicates a detailed analysis of the spectrum. The measured 2f-signal is a combination of all these three peaks as is illustrated in Fig. 5, which shows the actual spectrum and the corresponding fit of a 400 ppb CH4 sample measured in a scanning mode with a single OPO frequency sweep. Figure 5 also contains simulated 2f spectra (dashed lines) of the three individual peaks. The same modulation amplitude (0.0340 cm−1) and frequency (70 Hz) were used as in the HCN measurements.
Stability measurements were done by measuring the PA signal at the center of the strongest methane absorption line at 3057.68 cm−1 with a CH4 concentration of 35 ppb. The OPO frequency was not actively locked to the center of the peak during the measurement. As a result, the OPO frequency deviation during the 5 min measurement was ~60 MHz. Black square markers in Fig. 6 show a typical log10-log10 Allan deviation [1,39] plot of the PAS system, while the OPO is running freely. A detection limit of 175 ppt was reached with an averaging time of ~15 s. The OPO power was monitored continuously by detecting the power which passed the photoacoustic cell. When the initial PA signal was divided with the normalized power, the detection limit was enhanced to 65 ppt for an averaging time of 30 s. This is illustrated in the Allan deviation plot with red markers. Due to the slow response of the power meter, the effect of the power compensation can be seen only with averaging times longer than a few seconds. Further improvement in the long-term stability of the PA signal would require active stabilization of the OPO frequency. In comparison with our results, a detection limit of ~200 ppt with an averaging time of 100 s has been previously reported for methane using a CEPAS setup with a black body radiation source . Our detection limit is roughly by a factor of three lower compared to a previously reported photoacoustic spectroscopy measurement done with a mid-infrared cw-OPO and a resonant cell .
We have reported highly sensitive detection of hydrogen cyanide and methane by a cantilever-enhanced photoacoustic spectrometer using a mid-infrared cw OPO as a light source. Noise equivalent detection limits of 190 ppt (with an averaging time of 1 s) and 65 ppt (with an averaging time of 30 s) were achieved for HCN and CH4, respectively. The best normalized noise equivalent absorption (1σ) was 1.8 × 10−9 W cm−1 Hz−1/2, which was limited by the power and the wavelength fluctuations of the OPO. The stability and the sensitivity of the system can be further improved by locking the OPO frequency and by stabilizing its power. The absorption path length can also be doubled by reflecting the OPO beam back after it has passed the PA cell. Access to lower sample pressure (<400 mbar) would increase the selectivity of the measurement and would make the system more useful, for example, for breath analysis, where sub-atmospheric pressures (~100 mbar) are often used . Detection of highly adsorptive species with the PA201 would require measurements in continuous flow, which is difficult with the present analyzer. The response time of the system with adsorptive molecules could be enhanced using less adsorptive coating materials than gold for the sample cell.
We have shown that the continuous-wave optical parametric oscillator (cw-OPO) based apparatus has potential for significant enhancement in the sensitivity of cantilever-enhanced photoacoustic spectroscopy (CEPAS) based trace gas detection, owing to the high output power of OPOs in the MIR wavelength range that is optimal for molecular spectroscopy. To our knowledge, this is the first time when a mid-infrared cw-OPO has been used as a light source for CEPAS.
The financial support of the University of Helsinki and the Academy of the Finland is gratefully acknowledged. Gasera Ltd. is gratefully acknowledged for providing the PA201 analyser. The University of Helsinki is also acknowledged for funding the laboratory equipment used in this work.
References and links
1. P. Werle, R. Mücke, and F. Slemr, “The limits of signal averaging in atmospheric trace-gas monitoring by tunable diode-laser absorption spectroscopy (TDLAS),” Appl. Phys. B 57, 131–139 (1993). [CrossRef]
2. D. Romanini, A. A. Kachanov, N. Sadeghi, and F. Stoeckel, “CW cavity ring down spectroscopy,” Chem. Phys. Lett. 264(3-4), 316–322 (1997). [CrossRef]
3. A. O’Keefe, J. J. Scherer, and J. B. Paul, “CW integrated cavity output spectroscopy,” Chem. Phys. Lett. 307(5-6), 343–349 (1999). [CrossRef]
4. M. W. Sigrist, R. Bartlome, D. Marinov, J. M. Rey, D. E. Vogler, and H. Wächter, “Trace gas monitoring with infrared laser-based detection schemes,” Appl. Phys. B 90(2), 289–300 (2008). [CrossRef]
5. I. Galli, S. Bartalini, S. Borri, P. Cancio, D. Mazzotti, P. De Natale, and G. Giusfredi, “Molecular Gas Sensing Below Parts Per Trillion: Radiocarbon-Dioxide Optical Detection,” Phys. Rev. Lett. 107(27), 270802 (2011). [CrossRef] [PubMed]
6. F. M. J. Harren, G. Cotti, J. Oomens, and S. te Lintel Hekkert, “Photoacoustic spectroscopy in trace gas monitoring,” in Ensyclopedia of Analytical Chemistry, ed. by R. A. Meyers (John Wiley, 2000), 2203–2226.
7. Y. A. Bakhirkin, A. A. Kosterev, C. Roller, R. F. Curl, and F. K. Tittel, “Mid-infrared quantum cascade laser based off-axis integrated cavity output spectroscopy for biogenic nitric oxide detection,” Appl. Opt. 43(11), 2257–2266 (2004). [CrossRef] [PubMed]
8. B. G. Lee, M. A. Belkin, R. Audet, J. MacArthur, L. Diehl, C. Pflugl, F. Capasso, D. C. Oakley, D. Chapman, A. Napoleone, D. Bour, S. Corzine, G. Hofler, and J. Faist, “Widely tunable single-mode quantum cascade laser source for mid-infrared spectroscopy,” Appl. Phys. Lett. 91(23), 231101 (2007). [CrossRef]
9. R. F. Curl, F. Capasso, C. Gmachl, A. A. Kosterev, B. McManus, R. Lewicki, M. Pusharsky, G. Wysocki, and F. K. Tittel, “Quantum cascade lasers in chemical physics,” Chem. Phys. Lett. 487(1-3), 1–18 (2010). [CrossRef]
10. I. Ricciardi, E. De Tommasi, P. Maddaloni, S. Mosca, A. Rocco, J.-J. Zondy, M. De Rosa, and P. De Natale, “A narrow-linewidth optical parametric oscillator for mid-infrared high-resolution spectroscopy,” Mol. Phys. 110(17), 2103–2109 (2012). [CrossRef]
11. J. Peltola, M. Vainio, V. Ulvila, M. Siltanen, M. Metsälä, and L. Halonen, “Off-axis re-entrant cavity ring-down spectroscopy with a mid-infrared continuous-wave parametric oscillator,” Appl. Phys. B 107(3), 839–847 (2012). [CrossRef]
12. A. K. Y. Ngai, S. T. Persijn, G. von Basum, and F. J. M. Harren, “Automatically tunable continuous-wave optical parametric oscillator for high-resolution spectroscopy and sensitive trace-gas detection,” Appl. Phys. B 85(2-3), 173–180 (2006). [CrossRef]
13. F. Kühnemann, K. Schneider, A. Hecker, A. A. E. Martis, W. Urban, S. Schiller, and J. Mlynek, “Photoacoustic trace-gas detection using a cw single-frequency parametric oscillator,” Appl. Phys. B 66(6), 741–745 (1998). [CrossRef]
14. M. Nägele and M. W. Sigrist, “Mobile laser spectrometer with novel resonant multipass photoacoustic cell for trace-gas detection,” Appl. Phys. B 70(6), 895–901 (2000). [CrossRef]
15. A. Miklos, P. Hess, and Z. Bozoki, “Application of acoustic resonators in photoacoustic trace gas analysis and metrology,” Rev. Sci. Instrum. 72(4), 1937–1955 (2001). [CrossRef]
16. M. E. Webber, M. Pushkarsky, and C. K. Patel, “Fiber-amplifier-enhanced photoacoustic spectroscopy with near-infrared tunable diode lasers,” Appl. Opt. 42(12), 2119–2126 (2003). [CrossRef] [PubMed]
18. D. Newnham, X. Zhan, O. Vaittinen, E. Kauppi, and L. Halonen, “High-resolution photoacoustic study of the 4ν1 band system of monofluoroacetylene using a titanium:sapphire ring laser,” Chem. Phys. Lett. 189(3), 205–210 (1992). [CrossRef]
21. V. Spagnolo, P. Patimisco, S. Borri, G. Scamarcio, B. E. Bernacki, and J. Kriesel, “Part-per-trillion level SF6 detection using a quartz enhanced photoacoustic spectroscopy-based sensor with single-mode fiber-coupled quantum cascade laser excitation,” Opt. Lett. 37(21), 4461–4463 (2012). [CrossRef] [PubMed]
22. J. Kauppinen, K. Wilcken, I. Kauppinen, and V. Koskinen, “High sensitivity in gas analysis with photoacoustic detection,” Microchem. J. 76(1-2), 151–159 (2004). [CrossRef]
23. J. Uotila, V. Koskinen, and J. Kauppinen, “Selective differential photoacoustic method for trace gas analysis,” Vib. Spectrosc. 38(1-2), 3–9 (2005). [CrossRef]
24. T. Laurila, H. Cattaneo, V. Koskinen, J. Kauppinen, and R. Hernberg, “Diode laser-based photoacoustic spectroscopy with interferometrically-enhanced cantilever detection,” Opt. Express 13(7), 2453–2458 (2005). [CrossRef] [PubMed]
25. V. Koskinen, J. Fonsen, K. Roth, and J. Kauppinen, “Cantilever enhanced photoacoustic detection of carbon dioxide using a tunable diode laser source,” Appl. Phys. B 86(3), 451–454 (2007). [CrossRef]
26. A. Parkes, K. Keen, and E. McNaghten, “Trace gas detection using a novel cantilever-based photoacoustic spectrometer with multiplexed optical fiber-coupled diode lasers and fiber-amplification,” Proc. SPIE 6770, 67701C, 67701C-7 (2007). [CrossRef]
27. T. Kuusela, J. Peura, B. A. Matveev, M. A. Remennyy, and N. M. Stus’, “Photoacoustic gas detection using a cantilever microphone and III–V mid-IR LEDs,” Vib. Spectrosc. 51(2), 289–293 (2009). [CrossRef]
28. J. Uotila, J. Lehtinen, T. Kuusela, S. Sinisalo, G. Maisons, F. Terzi, and I. Tittonen, “Drug precursor vapor phase sensing by cantilever enhanced photoacoustic spectroscopy and quantum cascade laser,” Proc. SPIE 8545, 85450I, 85450I-13 (2012). [CrossRef]
30. M. Vainio, J. Peltola, S. Persijn, F. Harren, and L. Halonen, “Thermal effects in singly resonant continuous-wave optical parametric oscillators,” Appl. Phys. B 94(3), 411–427 (2009). [CrossRef]
31. H. I. T. R. A. N. The, 2008 database, URL: http://www.hitran.com/.
32. G. Herzberg, Molecular spectra and molecular structure, Vol. 2: Infrared and Raman Spectra of Polyatomic Molecules (van Nostrand Reinhold Company, 1945).
33. F. M. Schmidt, M. Metsälä, O. Vaittinen, and L. Halonen, “Background levels and diurnal variations of hydrogen cyanide in breath and emitted from skin,” J Breath Res 5(4), 046004 (2011). [CrossRef] [PubMed]
34. P. Bergamaschi, M. Schupp, and G. W. Harris, “High-precision direct measurements of 13CH4/12CH4 and 12CH3D/12CH4 ratios in atmospheric methane sources by means of a long-path tunable diode laser absorption spectrometer,” Appl. Opt. 33(33), 7704–7716 (1994). [CrossRef] [PubMed]
35. D. Wuebbles and K. Hayhoe, “Atmospheric methane and global change,” Earth Sci. Rev. 57(3-4), 177–210 (2002). [CrossRef]
36. A. Kosterev, T. Mosely, and F. Tittel, “Impact of humidity on quartz-enhanced photoacoustic spectroscopy based detection of HCN,” Appl. Phys. B 85(2-3), 295–300 (2006). [CrossRef]
37. P. Kluczynski, J. Gustafsson, Å. M. Lindberg, and O. Axner, “Wavelength modulation absorption spectrometry — an extensive scrutiny of the generation of signals,” Spectrochim. Acta B 56(8), 1277–1354 (2001). [CrossRef]
38. D. Arslanov, “Optical parametric oscillator based real-time trace gas analysis for bio-medical applications,” Ph.D. Thesis, Chapter 7, Radbound University, Nijmegen (2012).
39. D. W. Allan, “Statistics of atomic frequency standards,” Proc. IEEE 54(2), 221–230 (1966). [CrossRef]