1 Introduction

Quartz crystal tuning forks (QCTFs) as a type of piezoelectric mechanical resonators are more attractive due to their small size, low cost, low power consumption, high precision and high quality factor (Q factor), long stability and fully digital frequency output signal [1]. Piezoelectric QCTFs are commonly used for the construction of real-time clock circuits in various electronic devices, such as electronic oscillators, wrist watches, portable music players and computer main-boards. Now, based on their resonant frequency variations under different perturbations, various high-precision QCTF sensors had been developed, including mass sensors for biochemical and microbiological applications [2, 3], stress sensors for measuring liquid density [4] and high-resolution atomic force microscopy [5,6,7], as well as pressure and temperature sensors in low-temperature physics [6,7,8,9,10,11,12,13,14].

In 2002, QCTF was first applied as an acoustic wave transducer by Kosterev et al. at Rice University [15], instead of a microphone in traditional photoacoustic spectroscopy [16], thus the so-called quartz-enhanced photoacoustic spectroscopy (QEPAS) was proposed, and then several updated QEPAS techniques have been successively reported [17]. To date, QEPAS has been widely used in the fields of atmospheric pollution monitoring, combustion diagnosis, industrial process control, combustion diagnostics, medical breath analysis and molecular spectroscopy [18,19,20,21,22,23,24,25,26] due to its various advantages, such as high quality factor, small size, low cost and immunity to environmental acoustic noise, which are critical factors for developing gas sensors. Recently, QCTF-based optical detector was proposed for gas absorption spectroscopy measurements and successfully demonstrated for standoff detection of volatile organic compounds (VOCs) in our group [27, 28], and compared with a commercial mercury cadmium telluride (MCT) infrared detector (PVMI-4TE-10.6, Vigo Systems). In case of QEPAS technique, the laser beam was directed through the QCTF gap between two tines, the acoustic wave generated by the photoacoustic effect was used to excite the QCTF vibration. In our case, the transmitted laser light was vertically focused onto the QCTF tines, the light wave emitted from the light source is used to excite the QCTF mechanical vibration. Analogous to the photoelectric effect of semiconductor photodetectors, we utilize the piezoelectric effect of the QCTF to gauge the light intensity. The incident modulated or pulsed light beam generates a periodic driving force on the tines of QCTF, when the modulated frequency or pulse repetition rate is matched with the QCTF resonant frequency, a significant signal enhancement can be achieved. Due to its flexibility to the environmental changes, the mechanical properties of QCTF will change accordingly with the external environmental conditions. Therefore, the influence of environmental parameters on QCTF intrinsic property should be carefully investigated as it was used as an actuator and sensor.

In this paper, we proposed a new type of photoelectric detector based on a standard QCTF with resonant frequency of ~ 32 kHz. To explore the capabilities of this technique, the impact of incident light beam excitation positions with respect to QCTF on signal amplitude, resonant frequency and Q factor was observed in details. Moreover, the influence of incident light intensity, operating pressure and temperature on QCTF intrinsic property was also investigated systematically. Finally, the QCTF-based photodetector was demonstrated for qualitative analysis of gasoline components by combing a broadband tunable external cavity quantum cascade laser (ECQCL). It is worth noting that the presented QCTF detector technology may pave the way towards developing a new type of photodetector without spectral bandwidth limit.

2 Experimental details

To investigate the characteristics of QCTF as an optical detector, a series of commercial tuning forks with standard geometry were employed. The dimensions of the QCTFs were length, width, and thickness equal to 3700, 600, and 300 µm, respectively, as depicted in Fig. 1a. A schematic of the experimental platform is designed as shown in Fig. 2. A near-infrared diode laser model with fixed emitting wavelength at 1550 nm was used as the excitation light source. To generate a periodic driving force, we performed this experiment by applying a sinusoidal modulation to the laser current with a constant modulation amplitude of 1 V, while scanning the modulation frequency around the QCTF resonance frequency (i.e., 32.768 kHz) to observe its resonant profile. The laser beam was collimated by a fiber collimator and subsequently focused on the tines of the QCTF (with a ~ 0.5-mm-diameter light spot). A visible red light was also coupled for facilitating the beam alignment. In view of the geometrical symmetry of the QCTF, only front and left-side excitation modes were investigated, which generally classified into four different modes, namely left side (black ellipse), left prong (yellow circle), middle gap (red circle) and right prong (blue circle), the coordinate system of the QCTF was defined as in Fig. 1b. Due to its piezoelectric effect, the QCTF will generate piezoelectric current signal, which was amplified and transformed into voltage signal by a home-made low-noise operational amplifier. Finally, the voltage signal of the QCTF detector was transferred to a digital acquisition board (NI USB-6259) and synchronously analyzed within a notebook-based Labview program. Unlike a custom method based on the lock-in amplifier technique, here the signal of the QCTF detector was directly extracted by fast Fourier transform (FFT) analysis algorithm [27]. To investigate the dependence of the QCTF on ambient pressure, the QCTF detector was placed in a sealed cell which has two CaF2 windows allowing the laser light to pass through. Moreover, a mechanical pump and pressure meter were used to control the cell pressure.

Fig. 1
figure 1

3D drawing of quartz tuning fork (a) and the laser excitation positions (b)

Full size image
Fig. 2
figure 2

Experimental setup used for investigating QCTF characteristic

Full size image

3 Experimental results and discussion

3.1 The influence of excitation location

As mentioned above, tuning forks can be excited by laser beam in their different superficial positions. Generally, the vertical surface can be easily excited than the sides, due to its larger surface area. The output laser beam was directed at a 0.5-mm-diameter light spot on the surface of the QCTF. Since the gap between the QCTF prongs is approximately 0.3 mm, the incident laser beam actually motivates both prongs in the case of “middle gap” excitation mode. In this excited mode, the dependence of the QCTF resonant response on the vertical coordinate was first investigated in detail. As in Fig. 3, the signal amplitude of the QCTF are plotted as a function of the distance, which is defined from the root of the QCTF and shown in the inset. From Fig. 3 it is clear that the optimal distance of between 2.4 and 2.7 mm can be determined. At the optimal distance of L = 2.6 mm, the dependence of the QCTF signal amplitude on incident laser power was also investigated. Figure 4 shows an example of the dependence of QCTF detector response on the incident laser power. A good linearity between the QCTF detector signal amplitude and the incident laser power was found in the range from 0 to 4.5 mW, which corresponds to the modulation amplitude between 50 and 300 mV. From this result, we can conclude that the minimal light power change of several µW levels can be detected by the QCTF detector; currently, the detectivity was mainly limited by the noise of the preamplifier circuit.

Fig. 3
figure 3

QCTF signal amplitude as a function of L at the excited mode of middle gap

Full size image
Fig. 4
figure 4

The relationship between the QCTF signal amplitude and the incident laser power (for details, see text)

Full size image

Moreover, we found that the optimal distance for other excited modes was also found within this range. Based on this finding, the optimal L = 2.6 mm was always selected for the following experiments. As an example, Fig. 5 demonstrates the observed and fitted resonant profiles at different excitation modes for a same QCTF. As expected, the optimal resonance frequency f0 is always constant, while the Q factor and signal amplitude shows significant difference. Finally, the related results for selected four different QCTFs are summarized in Table 1. For clarity the relationship between the experimentally measured Q factor and the QCTF signal amplitude at f0 was plotted in Fig. 6. It seems that the maximal signal amplitude at f0 is inversely proportional to the Q factor. Although the active area in side excitation mode is minimal, the signal amplitude is not the weakest for a certain QCTF. Therefore, the optimal excitation mode should be carefully selected for each QCTF in real application, and a trade-off should also be made between the maximal signal amplitude at f0 and the Q factor.

Fig. 5
figure 5

Resonant profiles of QCTF at different excitation positions for a middle gap, b left prongs, c right prongs, d side

Full size image
Table 1 Measured and calculated QCTF characteristic parameters
Full size table
Fig. 6
figure 6

Experimentally measured Q factor vs. the QCTF signal amplitude at f0

Full size image

3.2 The influence of pressure effect

In this experiment, the QCTF detector was installed in a home-made photoacoustic cell, in which pressure can be regulated, and the optimal excited mode determined from previous experiments was always adopted. For instance, Fig. 7 shows the dependence of one QCTF intrinsic property on its operating pressure at the excited mode of middle gap. As can be seen, the Q value and signal amplitude shows to be inversely proportional to its operating pressure, while the optimal resonant frequency f0 shows a good linear response to pressure. Linear regression leads to a regression coefficient of R2 = 0.997 for pressures between 0 and 1 atm. The maximum Q factor of over 20,000 in vacuum and a resonance frequency shift of approximate 7 Hz were founded. Therefore, the QCTF detector will provide better performance by fabricating it in a vacuum cavity.

Fig. 7
figure 7

Experimentally measured QCFT resonant profile (a), resonant frequency f0 (b), signal amplitude at f0 (c) and Q factor (d), as a function of its operating pressure, respectively

Full size image

3.3 The influence of temperature effect

As a thermo-sensitive component, the specific mechanical structure of the QCTF is particularly sensitive to the external temperature. The mechanical strain of the QCTF would occur with the temperature change, which in turn affects the shift of its resonant frequency. To explore the influence of temperature on the QCTF intrinsic property, the QCTF was installed in a commercial thermotank. Measurement at each temperature set point was repeated for tens of minutes until a stable signal had been recorded. As shown in Fig. 8, we found that the QCTF resonance frequency f0 presents a significant dependence on environmental temperature, frequency shift of 3 Hz was found between 10 and 60 °C for the selected tuning fork. However, the variation of signal amplitudes at f0 and Q factors shows no obvious regularity.

Fig. 8
figure 8

The experimentally measured QCFT resonant curve (a), signal amplitude at f0 (b), resonant frequency f0 (c) and Q factor (d) as a function of ambient temperature, respectively

Full size image

3.4 QCTF-based photodetector for gasoline analysis

Finally, the developed QCTF detector was further evaluated for VOC spectroscopy measurement. For this experiment, an experimental configuration similar to our previous work [26] was used. A broadband tunable ECQCL (Block Engineering, USA) with a tuning range of 1130–1437 cm−1 (or 6.96–8.85 µm) was used as the excitation source to target chemical VOCs. A gas sample cell with a total path length of 29.6 cm was used for recording absorption spectra of gasoline vapor. For signal processing, a self-established spectral analysis model integrated with one-dimensional cubic spline interpolation algorithm (CSIA) and multiple linear regression algorithm (MLRM) was developed for quantitative and qualitative analysis of gasoline components. A detailed description of VOC signal processing method has already been discussed in a previous paper [27]. For instance, Figs. 9 and 10 demonstrated the experimentally recoded absorption spectra of different type of gasoline (coded with 95# and 92# in China), and the corresponding fitted results, respectively. As expected, there are numerous VOC components found in the gasoline samples, mainly including alcohol, acetone, isopentane, toluene, benzene, n-pentane, 2-methyl-1-butene, isooctane, n-butane, and ethyl tert-butyl ether. Generally, our experimental results show a good agreement with other values measured by other different methods. In view of the complexity of gasoline components and other potential factors (such as usage of catalyst by supplier), details of quantitative analysis and composition identification are still ongoing.

Fig. 9
figure 9

Experimentally measured absorption spectrum of 95# gasoline and the fitted results

Full size image
Fig. 10
figure 10

Experimentally measured absorption spectrum of 92# gasoline and the fitted results

Full size image

4 Conclusion

In this study, a new type of photoelectric detector based on a standard QCTF with resonant frequency of ~ 32 kHz was developed for infrared spectroscopy measurements. Analogous to the photoelectric effect of traditional semiconductor detectors, we utilize the piezoelectric effect of the QCTF to gauge the light intensity. This piezoelectric detector requires a modulation of incident light to coincide with its resonant frequency. To explore the capabilities of this technique, the impact of incident light beam excitation positions with respect to QCTF on signal amplitude, resonant frequency and Q factor was observed in detail. In addition, the influence of incident light intensity, pressure and temperature on QCTF resonant property was also investigated systematically. Generally, we found that each quartz crystal tuning fork has different resonant property. Therefore, the influence of environmental parameters on QCTF intrinsic property should be carefully investigated and selected when it was used as an actuator and sensor, such as optimal excited positions, signal amplitude, Q factor, and linear dependence of light intensity, as well as frequency shift on temperature and pressure. For practical application, the QCTF-based photodetector was successfully demonstrated for qualitative analysis of gasoline components by combing a broadband tunable ECQCL. It is worth noting that the presented QCTF detector technology may pave the way towards developing a new type of photodetector without spectral bandwidth limit.