Open Access
Add to BibSonomy
Abstract
On-line measurement is a trend of development toward laser-based applications. We present a fiber-integrated force sensor device for laser power measurement with both CW mode and pulse mode based on laser radiometric heat and radiation force sensing simultaneously. The sensor device is fabricated using a standard microfabrication process. Laser intensity is determined through the displacement of a movable mirror measured by an integrated Fabry-Perot interferometer. Compared with the performance of the device in the ambient condition, a non-linearity error of 0.02% and measurement uncertainty of 2.06% is observed in the quasi-vacuum condition for CW laser illumination. This device can measure a CW laser power with a 46.4 μW/Hz1/2 noise floor and a minimum detection limit of 0.125 mW. For a pulsed laser, a non-linearity error of 0.37% and measurement uncertainty of 2.08% is achieved with a noise floor of 1.3 μJ/Hz1/2 and a minimum detection limit of 3 μJ.
© 2023 Optica Publishing Group under the terms of the Optica Open Access Publishing Agreement
1. Introduction
Generally, laser power and energy are measured by photodiode, pyroelectric, and thermal detector. These methods require the conversion of absorbed photons into measurable current, voltage, or heating for measurement. None of them can achieve on-line measurement for the laser in use, which results in timeliness error in the measurement. A beam splitter can be implemented to monitor the power by picking up a small portion of laser beam for on-line measurement, but the measurement accuracy becomes very sensitive to the power dependent stability of the splitting ratio. The laser radiation force as an on-line measurement approach for laser power and pulse energy measuring arouses the interest of researchers. The laser can be used for further purposes after the measurement. This concept is first proposed by Lebedev [1] and has been widely studied in the measurement of laser power. The predominant approaches include the torsion balance method [2–4], plate capacitor method [5–8], cantilever method [9,10], etc. The principle of these methods lies in the detection of the displacement caused by the laser radiation force. For the torsion balance method, the laser power is determined by measuring the torsional angle caused by the laser radiation force on the torque-balanced system. However, the disturbance of environmental airflow and vibration significantly reduces the accuracy of the measurement. Besides, the entire measurement system is bulky and complicated. For the plate capacitor method, the miniaturization of the device can be achieved using the microfabrication process [11,12]. However, the parasitic capacitance and the pull-in effect limit the measurement resolution and range [7]. The cantilever method is limited to low laser power applications due to the stiffness decreasing with the increase in size [11,13].
The reported measurement of laser power based on the laser radiation force is from 0.1 mW to 50 kW with a measurement uncertainty above 1% [3,11–19]. Among them, the Ref. [11] and [15] have demonstrated the feasibility of Fabry-Perot optical interferometer on the laser measurement based on the radiation force. The common structure is a fixed mirror combined with a movable mirror for sensing the laser radiation force. Numerous studies have focused on extending the measurement range, decreasing the measurement uncertainty, and expanding the applications. Using a high reflective mirror is the most effective way to explore the upper limit of the laser power. The highest laser power reported for on-line measurement is 50 kW, which takes advantage of a high-reflectivity mirror [12]. Extensive studies have been focused on the approaches of uncertainty suppressing, whereby the accuracy of laser traceable measurement can be enhanced. In practice, the measurement uncertainty is dominated by external disturbances including airflow, vibration, and thermal drifts. Isolating the sensor device from the ambient environment by placing it in an air shield is the most common scheme to reduce measurement uncertainty [13]. In terms of applications, it has been demonstrated that the on-line monitoring of the input laser power (∼kW level) using the laser radiation force in the welding process is beneficial to improve the welding quality [20]. However, the laser on-line measurement for a low power or energy range has not been widely used. Application of on-line measurement can further accelerate the research of low power or energy laser in tissue regeneration [21], laser therapy [22] and nanomaterials processing [23–25]. Furthermore, the F-P cavity configuration is widely used in sound signal detection [26–29], nanoscale probe [15], gas composition analysis [30,31]. As a robust and scalable platform, MEMS (Micro-Electro-Mechanical Systems) technique potentially promotes the miniaturization and uniformity of these sensor devices.
In this work, we present a fiber-integrated sensor device for laser power and energy measurement based on the standard microfabrication process. A home-built fiber coupler is integrated into the microfabricated sensor device to form a Fabry-Perot interferometer for accurate displacement detection induced by laser illumination. This configuration enables laser power measurement for both CW mode and pulse mode. We systematically investigate the contribution of the laser radiometric heat and laser radiation force for CW and pulsed laser power measurements. To evaluate the laser power and energy with the sensor device, the mechanical vibration models are employed to describe the response with CW and pulsed laser illumination. Moreover, comparative experiments are implemented to confirm that placing the sensor device in a quasi-vacuum condition is beneficial to decrease the measurement uncertainty.
2. Sensor design and fabrication
The sensor device consists of the sensor chip and the fiber coupling module [32], as shown in Fig. 1(a). The optical assembly of the sensor chip and the fiber coupling module offers light access to measure the displacement of a suspended mirror with an external fiber Bragg grating interrogator (FBGI). After entering the Fabry-Perot cavity from optical fiber, the wideband light ranging from 1510 nm to 1590 nm interferes between the surface of the gradient index (GRIN) lens and the back-reflecting surface of the mirror and thus produces resonant peaks in the reflection spectra. The fiber (Thorlabs, UHNA1) in used is single mode fiber with a mode field diameter of 4 μm, a length of 0.5 m and a spectral bandwidth of 1100 - 1600 nm. As a laser beam illuminates the upper surface of the mirror vertically, the photon momentum is transferred to the mirror and the structure thermal expansion is induced by the laser absorption, whereby the spring flexure produces a displacement $\mathrm{\Delta}h$. This displacement leads to a change in the cavity length and the resonance wavelength shifts accordingly in the reflection spectra collected by FBGI. By tracking the resonant wavelength shift, the laser power or pulse energy can be determined. As shown in Fig. 1(b), the circular mirror is supported by a three-arms Archimedes spiral spring flexure [11]. The gaps between the spring arms ensure the pressure balance inside and outside of the Fabry-Perot cavity.
Fig. 1. (a) Schematic diagram of the fiber-integrated sensor device. (b) Microscope image of the front side view of the gold-coated Archimedes spiral springs flexure. The middle circular area supported by the spring arms is used for laser reflection. (c) Schematic of the fabrication process of the sensor device. (d) Photograph of the sensor device coupled with the fiber coupling module. (e) The simulated and measured spectrum obtained by FBGI.
The fabrication process of the sensor chip is presented in Fig. 1(c). The process starts with growing a 1-µm-thick SiO2 layer by thermal oxidation on a 4-inch double-side polished silicon wafer with a thickness of 300 µm. A 20-nm-thick Ti adhesive layer and a 400-nm-thick Au layer are sputtered on the surface of the silicon wafer successively. Afterwards, a composite of metallic layers is patterned with photolithography and ion beam etching is performed to form the circular mirror. In the following step, a cylindrical cavity with a 2.4-mm diameter and 200-µm depth is engraved with deep reactive ion etching (DRIE). Then a second-round DRIE is performed to engrave the Archimedes spiral spring arms at the bottom of the cavity. Due to the 200 µm spacing during the second lithography, the size of the etching structure is 10 µm larger than the previous design size. A thinner mirror makes a lighter proof mass, whereby the mirror is more sensitive to the incoming pressure induced by the laser radiation. Finally, a 400-nm-thick Au layer is deposited on the surface after the trench. The overall size of the measuring chip is 10 × 10 × 0.3 mm3. Note that, both sides of the mirror are coated with gold and a 92% reflectivity is obtained.
In the interferometry-based measurement, sensitivity and accuracy are determined by Q factor. The F-P cavity is constructed with two high-reflective surfaces, corresponding to the backside of the mirror and the front side of the fiber coupling module. The GRIN lens (Pitch: 0.25; NA:0.46) coated with alternating layers of SiO2 and Ta2O5 is mounted to the backside of the chip with epoxy, as shown in Fig. 1(d). Light interferes between the two surfaces and the wavelength for resonant mode in the cavity generates peaks in the reflection spectrum captured by the fiber grating interrogator in real time. After analyzing the full width at half maximum of the peak in the spectrum, the $Q$ factor can be obtained by
where the ${\lambda _0}$ is the resonant wavelength and $FWHM$ is the full width at half maximum of the resonant peak in the reflection spectrum. The simulated Q factor of the sensor is 5345 and the experimental result is 2421, as shown in Fig. 1(e). The loss of the Q factor is attributed to the imperfection of the surface and misalignment error. A promising solution is to replace the plano–plano Fabry-Perot cavity with the plano–concave cavity [33–35]. Further, the cavity length can be derived by [36] $FSR$ represents the free spectral range. $n$ represents the refractive index of the air in the F-P cavity. The calculated cavity length of the sensor device is 149.78 µm. The optical path of resonant light in the F-P cavity must be an integral multiple of its wavelength, therefore, the cavity length is also described as Combining Eq. (2) and Eq. (3), we find the cavity-to-wavelength conversion ratio m is 1:97. In the measurement, the resonance peak wavelength is extracted to monitor the movement of the suspended mirror in real time with a sampling frequency of 5000 Hz.The spring stiffness is an important parameter to determine the sensitivity and dynamic range of the sensor device and must be measured accurately for further calculation. By modeling the geometry and inserting the parameters of the materials in use, the spring stiffness is calculated to be 3189 N/m. Then, we measure the peak wavelength shifts by rotating the sensor device 180° (clockwise) from the upward position to the downward position. Due to gravity, 23.75 nm displacement is observed. The measurement is repeated 10 times and the spring stiffness of 3172 N/m is obtained, with an uncertainty of 2.1%. The measured stiffness is in good agreement with the simulated result.
With a relatively low power or pulse duration involved, the contribution of radiometric heat and radiation force are analyzed by comparing the displacement using the same laser power or energy in Comsol Multiphysics. Since the reflectivity of the gold layer is 92%, ∼8% of laser power is absorbed, which causes thermal expansion of the spring, as shown in Fig. 2(a). According to the volume thermal expansion formula $\mathrm{\Delta}V = \beta {V_0}\mathrm{\Delta}T$, the spring will expand $\mathrm{\Delta}V$ due to the temperature variation $\mathrm{\Delta}T$ caused by the laser illumination. The $\beta $ and ${V_0}$ represent the coefficient of volume expansion of silicon and initial volume of the mirror respectively. The $\mathrm{\Delta}T$ is determined by the formula $\mathrm{\Delta}T = {Q_{Laser}}/cm$, where c and m represent the heat specific capacity of silicon and the mass of the sensor device. The ${Q_{Laser}}$ represents the input laser energy. The theoretical displacement of mirror is plotted, as well as the simulated displacement with and without a spring in the Fig. 2(b). The trend of theoretical displacement is consistent with the simulated value, as shown in Fig. 2(b). However, the theoretical value is 1.4 times higher than simulated value as the thermal convection is involved in simulation. For CW mode, the simulated response time of displacement with a laser power of 500 mW is plotted in Fig. 2(c). It takes 101 s for the spring to reach 95% of the equilibrium state caused by the laser heating effect. The displacement of spring caused by the thermal effect that corresponds to the laser power ranging from 5 mW to 1000 mW are obtained in Fig. 2(d). By Eq. (4), the displacement caused by laser radiation force can be derived, as shown in Fig. 2(d),
where P represents the input laser power, ${E_{laser}}\; $ and $dt$ represent the input pulse energy and pulse duration, F represents the laser radiation force, c represents the speed of the light, R and $\alpha $ represent the reflectivity and absorption coefficient of the mirror, and $\theta $ represents the angle of laser incidence. As shown in Fig. 2(d), the displacement caused by the laser radiation force is 6 orders smaller than that caused by the laser radiometric heat effect. With a CW laser involved, the radiometric heat effect dominates the response of the sensor device. For pulse mode, the simulated time traces of mirror temperature and displacement caused by the thermal effect with a 7 ns duration are plotted in Fig. 2(e). The temperature and displacement of mirror increase within 15 ns. Owing to the thermoelastic effect, the temperature and displacement of mirror are observed oscillating decay and tend to reach an equilibrium state after 4.8 µs. The oscillation is possibly due to the photoacoustic effect. The displacement due to the heating effect is less than 1 nm, much smaller than that caused by radiation effect. In addition, frequency of oscillation is ∼MHz, thus the oscillation cannot be sampled by FBGI. The response induced by of laser radiation force is shown in Fig. 2(f). It is found that the displacement caused by the laser radiometric heat is 2149 times smaller than that caused by the laser radiation force effect. Therefore, the radiation force effect is dominant for ultra-short laser pulse illumination.Fig. 2. (a) Simulation of sensor thermal expansion with and without spring structure. (b) Plot of theoretical and simulated displacement caused by heat effect. The black star, green dot and purple triangle indicate theoretical displacement, simulated displacement without and with spring structure respectively. (c) Simulated time traces of the displacement of mirror caused by laser radiometric heat with a laser power of 500 mW in a vacuum condition. (d) Plot of the simulated displacement caused by heat effect and calculated displacement caused by force effect as a function of laser power in a vacuum condition. (e) Simulated time traces of the displacement of mirror with laser energy at 1000 µJ and 108 µJ in a vacuum condition. The solid line represents the temperature change, and the dash line represents the displacement change. (f) Plot of the simulated displacement caused by heat effect and calculated displacement caused by force effect as a function of laser energy in a vacuum condition.
3. Results and discussion
3.1 CW laser measurement
To evaluate the performance of the sensor device with CW laser illumination, a simple test system is constructed, as shown in Fig. 3(a). A 532-nm laser (Coherent, Verdi-V6) passes through an acoustic optical modulator (AOM, CETC) and then focuses on the center of the mirror. The diameter of the focal spot on the sensor is ∼500 µm in diameter with a Gaussian beam profile obtained by the beam analyzer (Thorlabs, BP209IR1/M), as shown in Fig. 3(b). AOM is driven by a waveform generator (Keysight, 33600A). AOM modulation is turned off for CW measurement and on for modulation measurement.
Fig. 3. (a) Schematic of the power test system for CW laser. The CW laser is modulated by AOM driven by a waveform generator. The chamber provides a quasi-vacuum condition. AOM: acoustic optical modulator; WG: waveform generator; PC: personal computer; FBGI: fiber Bragg grating interrogator. (b) Observed beam profile of focused laser beam at the sensor position.
By simulation and calculation, the displacement induced by the laser radiation force effect is 6 orders smaller than that caused by the laser radiometric heat effect. Therefore, the response of the sensor device is dominated by the laser radiometric heat effect. Considering the disturbance from the ambient environment (e.g., airflow and heat source), keeping the sensor device in a quasi-vacuum condition is an effective way to avoid ambient noise. To evaluate the influence of airtight packaging, the performance of the sensor device is investigated at the ambient condition (AC) and quasi-vacuum condition (QVC) for comparison, respectively. The pressure for the quasi-vacuum condition is about 10−1 Pa. First, the CW laser is normal incident on the sensor device at AC for 600 s. The time trace of the mirror displacement with different laser power is plotted in Fig. 4(a). The displacement exponentially declines and tends to reach an equilibrium state. To protect the Au reflection layer on the mirror, the maximum power we used is constrained to 1000 mW. In contrast, the sensor device is placed in a vacuum chamber for airflow and heat source rejection and the response curve of the mirror is plotted in Fig. 4(b). Due to the heat exchange through airflow, the temperature of the thermal equilibrium state caused by laser radiometric heat is lowered and thus the rise time at AC is 20 s to 40 s shorter than that at QVC, as shown in Fig. 4(c). To evaluate the sensitivity and linearity of the sensor, the wavelength shifts from the initial peak wavelength to the averaged peak wavelength reaching the equilibrium state (the grey area in Fig. 4(a) and Fig. 4(b)) are plotted as a factor of input laser power. Then, a linear fitting is employed to the function. As shown in Fig. 4(d) and 4(e), the sensitivity of 40 pm/mW with a non-linearity error of 1.24% at AC and sensitivity of 84 pm/mW with a non-linearity error of 0.02% at QVC is achieved. The non-linearity error is quantified by averaged the fit residuals [11]. In Fig. 4(d, e), the green and purple dots are represented the measured value, while the black dots are represented the simulated value at the same laser power. The trend of measured wavelength shifts is consistent with the simulated result. The difference between the simulated and measured values comes from the parameter setting and size error by fabrication in the simulation. Further, the measurement error is analyzed by calculating the standard deviation of the peak wavelength after reaching the equilibrium state. We find that the averaged measurement error at AC is 27 times higher than that at QVC due to the impact of the nonuniform external airflow. The detection limit is determined by the SNR. After turning off the laser, the standard deviation of the peak wavelength is 3.5 pm at QVC and 4.7 pm at AC. Thus, the estimated detection limits of 0.125 mW and 0.352 mW are achieved at QVC and AC, respectively. Compared with the measurement errors and the detection limits obtained for two cases, it is evident that the stability is superior at QVC. The errors in the x axis in Fig. 4(d, e) indicate power fluctuation measured by the power meter.
Fig. 4. (a) Measurements of the peak wavelength with laser power ranging from 5 to 1000 mW at AC. The grey area represents that the displacement reaches an equilibrium state. The inset shows the decay of the peak wavelength from 0 s to 15 s. (b) Measurements of the peak wavelength with laser power ranging from 0.5 to 500 mW at QVC. The grey area represents that the displacement reaches an equilibrium state. The inset shows the decay of the peak wavelength from 0 s to 15 s. (c) Plots of the rise time as a function of laser power at QVC and AC, respectively. (d) The Plot of the measured (in purple) and simulated (in black) peak wavelength shift as a function of laser power at AC. (e) The Plot of the measured (in purple) and simulated (in black) peak wavelength shift as a function of laser power at QVC. The errors in the x axis in (d) and (e) indicate power fluctuation measured by the power meter. (f) The plot of the noise density as a function of laser power at QVC. The noise density spectra are plotted in the inset with laser power of 0 mW, 50 mW, and 100 mW, respectively.
Fourier transform of representative segments reaching equilibrium state ranging from 300 s to 600 s in Fig. 4(b) is used to analyze the noise spectra for laser power from 0 to 500 mW at QVC. The noise density floor value is averaged from 1 Hz to 100 Hz in the spectrum. As shown in Fig. 4(f), the noise density increases with the rising of the laser power. The noise floor of 46.4 µW/Hz1/2 at QVC is obtained. The 1/f noise possibly originates from the ambient disturbance.
3.2 Modulated CW laser measurement
To evaluate the performance of the sensor device for the intensity-modulated laser at QVC, a laser modulated at 2 Hz with a 50% duty cycle is injected, as shown in Fig. 3(a). The response curves of the sensor are plotted in Fig. 5(a). The response time is above 100 s. When the peak wavelength reaching the equilibrium state, the time trace of the peak wavelength from 400 s to 401.5 s is plotted in the Fig. 5(a) and the periodic rectangular pulse modulation waveform are shown in Fig. 5(b). Compared with the modulation speed, the temperature rise is a relatively slow response process. Therefore, the response of peak wavelength is an approximate triangular wave in one period, as shown in Fig. 5(b).
Fig. 5. (a) Measurements of the change of the resonance peak wavelength at a laser power of 100 mW and 500 mW with a modulation frequency of 2 Hz at QVC. (b) The time traces of the peak wavelength from 400 s to 401.5 s are plotted in black for 100 mW and red for 500 mW and the driving periodic rectangular pulse modulation waveform is labeled accordingly. (c) The Fourier transform spectrums of the peak wavelength at a laser power of 100 mW and 500 mW from 300 s to 600 s. (d) The plot of the vibration amplitude at 10 Hz as a function of the modulated laser power ranging from 1 mW to 1000 mW at QVC.
Taking a Fourier transform of the peak wavelength from the time segment from 300 s to 600 s, peaks at the modulation frequency and its harmonics are observed, as shown in Fig. 5(c). Due to the slow drift with 500 mW illumination, the low frequency noise of for 500 mW is higher than that of 100 mW. The harmonic frequencies of the modulation frequency in Fig. 5(c) originate from the distribution of the sensor response feature in the frequency domain. It can be described by a typical forced vibration model with a periodic excitation applied [37,38]. The excitation in this system can be described as a periodic function,
From Eq. (7), the vibration amplitude of the spring reflects the intensity of the laser power. We use the amplitude at the modulation frequency to describe the vibration amplitude of the spring. The amplitude is obtained by the ratio of the height of the peak to a flattened noise floor as shown in Fig. 5(c). Taking frequency domain analysis, the amplitude at the frequency of modulation is increased linearly with laser power from 1 mW to 1000 mW at a modulation frequency of 10 Hz, as shown in Fig. 5(d). The sensitivity of the laser power measurement is 20.53 dB/mW with a non-linearity error of 0.76% at QVC.
3.3 Pulsed laser measurement
To evaluate the performance of the sensor device with pulsed laser illumination, the test system is modified based on the Fig. 3(a). The AOM is replaced by a beam splitter. A 532-nm pulsed laser (Spectra-Physics) with a duration of 7 ns is split into two arms with equal energy through the beam splitter. One arm of the laser is focused on the sensor through a lens (f = 150 mm) and the other arm is collected by an energy meter (Ophir, PE50-DIF-v2) for laser energy calibration.
In simulation, it is found that the displacement induced by the laser radiometric heat effect is 2149 times weaker than that caused by the laser radiation force. Therefore, the response of the sensor device is governed by the radiation force and can be described by the free vibration model [39]. Considering the initial displacement and velocity are zero, the response of the sensor device is described as
where ${x_p}(t )$ represents the oscillation of the peak wavelength with a single pulse applied, $h$ represents the amplitude of the vibration, $\zeta = 0.02$ represents the damping coefficient at QVC, and ${p_d} = {p_n}{({1 - {\zeta^2}} )^{1/2}}$ represents the vibration frequency. The damping rate of the spring in the quasi-vacuum condition is calculated to be 0.2 by the formula ${\zeta _{spring}} = \; \zeta /2\sqrt {mk} $. The vibration frequency is calculated to be 12673 Hz. The vibration frequency is greater than three times the sampling frequency of FBGI, therefore, the vibration frequency cannot be full sampled by FBGI. By fitting the upper envelope of the response curve with the exponential term of Eq. (8), h index can be used to quantify the vibration amplitude of the spring caused by the momentum transfer applied by the laser pulse. The h index as a function of the laser energy is plotted in Fig. 6(a). The sensitivity of the sensor is 1.38 pm/µJ with a non-linearity error of 0.37% at QVC and 0.75 pm/µJ with a non-linearity error of 0.69% at AC. The measurement error is analyzed by calculating the 95% confidence interval of the h index. It is found that the measurement errors of the two cases are at the same level due to the limited sampling rate of the FBGI. Turning off the pulsed laser, the noise density floor at AC is three times higher than that at QVC as shown in Fig. 6(b). According to the SNR, detection limits of 3 µJ and 10 µJ are obtained at QVC and AC, respectively. It attributes to the isolation of the disturbance in the vacuum chamber. In addition, as shown in Fig. 6(c) and 6(d), the amplitude at QVC is doubled compared with the result at AC given the same pulse energy due to the smaller air friction. The trend of measured displacement in Fig. 6(c) is consistent with the calculated values in Fig. 2(f).Fig. 6. (a) Plots of the h index as a function of pulse energy at QVC and AC, respectively. Linear fitting curves are plotted in black and red for QVC and AC. (b) Noise density spectra of the sensor device without laser injection at QVC and AC, respectively. (c) Measurements of the changes in peak wavelength changes with 1046 µJ, 729 µJ, 404 µJ, 125 µJ, and 63 µJ pulse energy injected at QVC. Exponential fitting curves are plotted in red. (d) Measurements of the changes in peak wavelength changes with 1000 µJ, 835 µJ, 499 µJ, 108 µJ, and 48 µJ pulse energy injected at AC.
When the pulsed laser is operated at a repetition rate of 10 Hz, the response curves of the sensor at QVC and AC are shown in Fig. 7(a) and 7(b), respectively. The amplitude of each pulse is significantly different due to the instability of the laser source. In addition, the peak wavelength drifts slowly over time due to the accumulation of radiation heat, which could be mitigated by an optical film with a higher reflectivity [40].
Fig. 7. (a) Measurement of the change in peak wavelength of the sensor device with laser energies of 500 µJ (10 Hz) and 100 µJ (10 Hz) at QVC. A segment of wavelength shift from 1.0 s to 1.2 s is plotted below for details. (b) Measurement of the change in peak wavelength of the sensor device with laser energies of 500 µJ (10 Hz) and 100 µJ (10 Hz) at AC. A segment of wavelength shift from 1.0 s to 1.2 s is plotted below.
3.4 Measurement uncertainty
The measurement uncertainty is coming from a combination of the electronics noise, the laser instability and external vibration and airflow. Plots of uncertainty as a function of power for CW laser and energy for pulsed laser are shown in Fig. 8, respectively.
Fig. 8. Plot of the measurement uncertainty as a function of CW laser power at QVC (a) and AC (b). Plot of the measurement uncertainty as a function of pulsed laser energy at QVC (c) and AC (d). The red dashed line represents the uncertainty reaching steady state.
Plots of measurement certainty as a function of power for CW laser and energy for pulsed laser are shown in Fig. 8, respectively. In Fig. 8, it is found that the uncertainty drops with increasing laser power at QVC and converges to a steady state of 2.06%. However, the measurement uncertainty at AC is kept at a level of ∼19.85% in the range of 10 to 1000 mW due primarily to the thermal noise and convection of external airflow for the unshielded environment. For pulsed laser, the measurement uncertainty decreases faster for the sensor at QVC and reaches 2.08% at ∼65 µJ, whereas the measurement uncertainty at AC drops down to 3.26% after the energy approaches 600 µJ then noise begins to dominate.
Finally, the performance of the sensor is summarized and compared in Table 1. Both CW laser and pulsed laser power measurement can be achieve using the sensor device. However, compared with Refs. [11,12], the measurement range and measurement uncertainty of the sensor device need to be further extended and improved.
Table 1. Performance of the sensor device and comparison
4. Conclusion
We develop a miniaturized and reliable sensor for laser power and energy measurement using a standard microfabrication process. A Fabry-Perot interferometer is integrated into the sensor device by the optical assembly. Archimedes spiral force-sensing spring with a stiffness of 3172 N/m is designed and fabricated. A 532-nm CW laser and a pulsed laser are used to test the performance of the microfabricated sensor device. This device enables on-line laser power measurement for both CW mode at milliwatt scale and pulse mode at millijoule scale. Compared with the performance of the sensor device at AC, the device at QVC can measure a CW power with 84 pm/mW sensitivity and 0.125 mW detection limit. For pulsed mode, a sensitivity of 1.38 pm/µJ with a detection limit of 3 µJ is achieved at QVC. Comparative experiment results demonstrate the performance of the proposed device can be further optimized with hermetic sealing. Based on the capability of the microfabrication process, further research will focus on the following aspects. First, develop the maskless etching process to replace the plano–plano Fabry-Perot cavity with the plano–concave cavity. This helps to increase the alignment tolerance in the optical assembly. Second, common mode rejection could be employed to reduce the noise floor caused by the intrinsic vibration of the floating structure. Finally, vacuum package the sensor device by the anodic bonding process at wafer level helps the sensor better adapted to laser-based applications.
Funding
CAS Strategic Pilot Project (XDC07030200); National Key Research and Development Program of China (2021YFB3202500); R&D Program of Scientific Instruments and Equipment, Chinese Academy of Sciences (YJKYYQ20190026).
Acknowledgments
The authors thank Prof. Xiaohong Ge from the Shanghai Institute of Microsystem and Information Technology, Chinese Academy of Sciences, for sample preparation and useful discussion.
Disclosures
The authors declare no conflict of interest.
Data availability
The data presented in this paper is available from the authors upon request.
References
1. V. M. Berezanskaya, I. Y. Doskoch, and M. A. Man’ko, “Light-pressure experiments by P. N. Lebedev and modern problems of optomechanics and quantum optics,” J. Russ Laser Res. 37(5), 425–433 (2016). [CrossRef]
2. Y. Yuan, “A new pulse laser energy meter,” Rev. Sci. Instrum. 61(6), 1743–1746 (1990). [CrossRef]
3. P. Pinot and Z. Silvestri, “Optical power meter using radiation pressure measurement,” Measurement 131, 109–119 (2019). [CrossRef]
4. F. Mueller, S. Heugel, and L. J. Wang, “Femto-Newton light force measurement at the thermal noise limit,” Opt. Lett. 33(6), 539–541 (2008). [CrossRef]
5. I. Ryger, A. Artusio-Glimpse, P. Williams, G. Shaw, M. Simons, C. Holloway, and J. Lehman, “MEMS non-absorbing electromagnetic power sensor employing the effect of radiation pressure,” Proceedings 2(13), 767 (2018). [CrossRef]
6. I. Ryger, R. Balogh, S. Chamraz, A. Artusio-Glimpse, M. Stephens, P. A. Williams, and J. Lehman, “Feedback control of a nonlinear electrostatic force transducer,” Sensors 20(24), 7337 (2020). [CrossRef]
7. I. Ryger, A. B. Artusio-Glimpse, P. Williams, N. Tomlin, M. Stephens, K. Rogers, M. Spidell, and J. Lehman, “Micromachined force scale for optical power measurement by radiation pressure sensing,” IEEE Sensors J. 18(19), 7941–7948 (2018). [CrossRef]
8. G. C. Spalding, K. Dholakia, A. Walowitz, D. Rahn, J. Lehman, K. Rogers, P. Williams, I. Ryger, and A. B. Artusio Glimpse, “Mechanical characterization of planar springs for compact radiation pressure power meters,” Optical Trapping and Optical Micromanipulation XV (SPIE2018), 10723 (2018).
9. D. Ma, J. L. Garrett, and J. N. Munday, “Quantitative measurement of radiation pressure on a microcantilever in ambient environment,” Appl. Phys. Lett. 106(9), 091107 (2015). [CrossRef]
10. P. R. Wilkinson, G. A. Shaw, and J. R. Pratt, “Determination of a cantilever's mechanical impedance using photon momentum,” Appl. Phys. Lett. 102(18), 184103 (2013). [CrossRef]
11. A. B. Artusio-Glimpse, I. Ryger, N. A. Azarova, P. A. Williams, J. A. Hadler, and J. H. Lehman, “Miniature force sensor for absolute laser power measurements via radiation pressure at hundreds of watts,” Opt. Express 28(9), 13310–13322 (2020). [CrossRef]
12. P. A. Williams, A. B. Artusio-Glimpse, J. A. Hadler, D. W. King, K. A. Rogers, I. Ryger, T. Vo, and J. H. Lehman, “Radiation-pressure-enabled traceable laser sources at CW powers up to 50 kW,” IEEE Trans. Instrum. Meas. 68(6), 1833–1839 (2019). [CrossRef]
13. P. Williams, J. Hadler, F. Maring, R. Lee, K. Rogers, B. Simonds, M. Spidell, M. Stephens, A. Feldman, and J. Lehman, “Portable, high-accuracy, non-absorbing laser power measurement at kilowatt levels by means of radiation pressure,” Opt. Express 25(4), 4382–4392 (2017). [CrossRef]
14. J. H. Lehman, K. Rogers, D. Rahn, and P. Williams, “Inline laser power measurement by photon momentum,” Appl. Opt. 58(5), 1239–1241 (2019). [CrossRef]
15. Z. Li, S. Gao, U. Brand, K. Hiller, S. Hahn, and H. Wolff, “Traceable laser power measurement using a micro-machined force sensor with sub-piconewton resolution,” in Proceedings of IEEE 20th International Conference on Solid-State Sensors, Actuators and Microsystems and Eurosensors XXXIII (TRANSDUCERS and EUROSENSORS XXXIII) (IEEE, 2019), pp. 1603–1606.
16. G. A. Shaw, J. Stirling, J. Kramar, P. Williams, M. Spidell, and R. Mirin, “Comparison of electrostatic and photon pressure force references at the nanonewton level,” Metrologia 56(2), 025002 (2019). [CrossRef]
17. S. Vasilyan, T. Frohlich, and E. Manske, “Total momentum transfer produced by the photons of a multi-pass laser beam as an evident avenue for optical and mass metrology,” Opt. Express 25(17), 20798–20816 (2017). [CrossRef]
18. R. Wagner, F. Guzman, A. Chijioke, G. K. Gulati, M. Keller, and G. Shaw, “Direct measurement of radiation pressure and circulating power inside a passive optical cavity,” Opt. Express 26(18), 23492–23506 (2018). [CrossRef]
19. A. B. Artusio-Glimpse, K. A. Rogers, P. A. Williams, and J. H. Lehman, “High amplification laser-pressure optic enables ultra-low uncertainty measurements of optical laser power at kilowatt levels,” Metrologia 58(5), 055010 (2021). [CrossRef]
20. P. Williams, J. Sowards, and B. Simonds, “Measuring laser beam welding power using the force of light,” Weld. J. 95(3), 30–34 (2016).
21. A. Zennifer, A. Subramanian, and S. Sethuraman, “Design considerations of bioinks for laser bioprinting technique towards tissue regenerative applications,” Bioprinting 27, e00205 (2022). [CrossRef]
22. S. N. Meller, M. C. de Oliveira, G. F. Krueger, K. Z. Younan, H. T. Gassen, P. A. G. Hernández, S. P. Moure, and S. A. Q. Miguens, “Micro-CT analysis of the mandibular bone microarchitecture of rats after radiotherapy and low-power laser therapy,” Lasers Med. Sci. 37(6), 2645–2653 (2022). [CrossRef]
23. V. Amendola and M. Meneghetti, “What controls the composition and the structure of nanomaterials generated by laser ablation in liquid solution?” Phys. Chem. Chem. Phys. 15(9), 3027–3046 (2013). [CrossRef]
24. J. Zhang, M. Chaker, and D. Ma, “Pulsed laser ablation based synthesis of colloidal metal nanoparticles for catalytic applications,” J. Colloid Interface Sci. 489, 138–149 (2017). [CrossRef]
25. K. Yin, L. Wang, Q. Deng, Q. Huang, J. Jiang, G. Li, and J. He, “Femtosecond laser thermal accumulation-triggered micro-/nanostructures with patternable and controllable wettability towards liquid manipulating,” Nano-Micro Lett. 14(1), 97 (2022). [CrossRef]
26. B. Fischer, “Optical microphone hears ultrasound,” Nat. Photonics 10(6), 356–358 (2016). [CrossRef]
27. S. Lorenzo and O. Solgaard, “Optical fiber-facet multiplexed monolithic silicon pressure sensors,” IEEE Sensors J. 20(18), 10598–10606 (2020). [CrossRef]
28. S. Lorenzo, A. Kroo, Y.-P. Wong, and O. Solgaard, “Thermally-compensated optical fiber silicon sensor platform,” IEEE Sensors J. 21(21), 24121–24128 (2021). [CrossRef]
29. S. Lorenzo, Y.-P. Wong, and O. Solgaard, “Optical fiber photonic crystal hydrophone for cellular acoustic sensing,” IEEE Access 9, 42305–42313 (2021). [CrossRef]
30. H. Thienpont, J. Mohr, H. Zappe, H. Nakajima, N. P. Ayerden, G. de Graaf, P. Enoksson, and R. F. Wolffenbuttel, “A highly miniaturized NDIR methane sensor,” Proc. SPIE 9888, 98880D (2016). [CrossRef]
31. B. Zhang, K. Chen, Y. Chen, B. Yang, M. Guo, H. Deng, F. Ma, F. Zhu, Z. Gong, W. Peng, and Q. Yu, “High-sensitivity photoacoustic gas detector by employing multi-pass cell and fiber-optic microphone,” Opt. Express 28(5), 6618–6630 (2020). [CrossRef]
32. F. Xie, Y. Hu, L. Li, C. Wang, Q. Liu, N. Wang, L. Wang, S. Wang, J. Cheng, H. Chen, and Z. Wu, “A microfabricated fiber-integrated diamond magnetometer with ensemble nitrogen-vacancy centers,” Appl. Phys. Lett. 120(19), 191104 (2022). [CrossRef]
33. K. P. Larsen, J. T. Ravnkilde, and O. Hansen, “Investigations of the isotropic etch of an ICP source for silicon microlens mold fabrication,” J. Micromech. Microeng. 15(4), 873–882 (2005). [CrossRef]
34. G. Wachter, S. Kuhn, S. Minniberger, C. Salter, P. Asenbaum, J. Millen, M. Schneider, J. Schalko, U. Schmid, A. Felgner, D. Huser, M. Arndt, and M. Trupke, “Silicon microcavity arrays with open access and a finesse of half a million,” Light: Sci. Appl. 8(1), 37 (2019). [CrossRef]
35. J. Fait, S. Putz, G. Wachter, J. Schalko, U. Schmid, M. Arndt, and M. Trupke, “High finesse microcavities in the optical telecom O-band,” Appl. Phys. Lett. 119(22), 221112 (2021). [CrossRef]
36. E. Kim, S.-J. Jeon, S.-H. Kim, D.-H. Kim, J.-S. Lee, and Y.-W. Choi, “Wavelength shift measurement with resonance characteristic prediction algorithm using intensities at discrete wavelengths,” IEEE Trans. Instrum. Meas. 69(8), 5666–5672 (2020). [CrossRef]
37. J. Ye, Z. Q. Lu, C. Chen, and M. Q. Wang, “Power analysis of a single degree of freedom (DOF) vibration energy harvesting system considering controlled linear electric machines,” in Proceedings of IEEE Transportation Electrification Conference and Expo (IEEE, 2017), pp. 158–163.
38. S. Ye, J. Zhang, B. Xu, L. Hou, J. Xiang, and H. Tang, “A theoretical dynamic model to study the vibration response characteristics of an axial piston pump,” Mech. Syst. Signal PR 150, 107237 (2021). [CrossRef]
39. Y. M. Chen, Q. X. Liu, and J. K. Liu, “Time-dependent decay rate and frequency for free vibration of fractional oscillator,” J. Appl. Mech. 86(2), 024501 (2019). [CrossRef]
40. P. A. Williams, K. A. Rogers, M. Divoky, G. Porrovecchio, J. Tesar, M. Smid, and J. H. Lehman, “Extreme laser pulse-energy measurements by means of photon momentum,” Opt. Express 30(5), 7383–7393 (2022). [CrossRef]
