Open Access
Add to BibSonomyAbstract
A Q-switched Nd:YAG laser providing nanosecond pulse durations and millijoule pulse energies is suitable for typical biomedical PA applications. However, such lasers are both bulky and expensive. An alternative method is to use a diode laser, which can achieve a higher pulse repetition frequency. Although the energy from a diode laser is generally too low for effective PA generation, this can be remedied by using high-speed coded laser pulses, with the signal intensity of the received signal being enhanced by pulse compression. In this study we tested a version of this method that employs coded excitation. A 20-MHz PA transducer was used for backward-mode PA detection. A frequency-coded PA signal was generated by tuning the interval between two adjacent laser pulses. The experimental results showed that this methodology improved the signal-to-noise ratio of the decoded PA signal by up to 19.3 dB, although high range side lobes were also present. These side lobes could be reduced by optimizing the compression filter. In contrast to the Golay codes proposed in the literature, the proposed coded excitation requires only a single stimulus.
©2011 Optical Society of America
1. Introduction
It has been proposed that the photoacoustic (PA) effect could be used to distinguish soft tissues with different optical absorption properties [1], since thermal expansion induced by electromagnetic energy will generate acoustic waves from the region of interest. Generally, a Q-switched Nd:YAG laser is effective at generating a PA signal since it can provide an energy of millijoules within a pulse duration of tens of nanoseconds. However, such lasers are bulky and expensive. As an alternative, diode lasers with continuous-wave excitation and pulsed excitation are available for PA applications [2]. The size of the diode lasers is more compact and the price is cheaper than a conventional Nd:YAG laser. In addition, they achieve high pulse repetition frequencies (PRFs). The main drawback of a diode laser is that its pulse energy is too low for the effective generation of a PA signal. Therefore, strategies for increasing the efficacy of PA signal generation using a diode laser need to be developed.
One method used to boost the instantaneous output of a diode-laser-based PA system is simply to combine several diodes [3,4] and merge the corresponding laser fibers into a bundle. Such a method is potentially useful in imaging superficial vascular structures [4]. The stimulus laser energy can also be increased using coded excitation, which has been widely applied in conventional ultrasound imaging [5]. The high PRF of a diode laser makes it able to produce coded laser pulses, with the resultant signal intensity being enhanced by pulse compression. A Golay code has previously been proposed for generating a PA signal, since in principle it provides perfect side-lobe cancellation whilst enhancing the signal intensity [6,7]. Moreover, it is possible to use orthogonal Golay codes to excite a sample at two wavelengths simultaneously in order to speed up PA imaging [7].
However, a PA system that combines several laser diodes is complex due to the need to have additional laser diodes and drives. Moreover, further boosting the transmitted laser energy by lengthening the laser pulse duration reduces the axial resolution [3,4]. While the use of coded excitation is more feasible in PA imaging, two problems are anticipated. First, due to the insufficient PRF (i.e., maximum of 500 kHz) of lasers used in previous studies, the relatively long pulse repetition interval of the codes reduced the detection range [7]. The time-consuming processing is analogous to a common averaging method. Second, since a diode laser can only transmit a uniphase code, to achieve Golay-coded excitation with code pair [1,1] and [1, –1], for example, two excitations (i.e., [1 0] and [0 1]) are required to effectively obtain the result from the desired excitation (i.e., [1, –1]), where the [1 0] response is subtracted from that for [0 1]. Therefore, three excitations are required to achieve the Golay-coded excitation in this case. In other words, the use of multiple excitation reduces the imaging frame rate and makes the procedure susceptible to motion artifacts.
This study employed coded excitation using optical pulse modulation to effectively generate PA signals using a high-speed diode laser. This technique involves generating a frequency-coded PA signal using consecutive laser pulses, and hence the instantaneous frequency of the PA signal is determined by the interval between two consecutive laser pulses. We chose linear frequency modulation (FM) pulses to demonstrate the feasibility of this proposed method, with the generated PA signal being decoded (i.e., pulse compression) by matched filtering.
The proposed method provides several advantages. First, the complete system is compact since it uses only a single laser diode. Second, only a single coded excitation sequence is required to generate the PA signal. Therefore, the PA imaging is faster than for codes that require multiple excitations. The highest pulse frequency achieved in our experiment is 7-MHz, which corresponds to a pulse period with 71.42-ns. Lastly, a PA transducer integrating an optical fiber was proposed to enable backward-mode PA detection, which is potentially useful for PA imaging.
2. Methods
2.1 Selecting the coding parameters
Generating a frequency-coded PA signal using a diode laser requires the determination of parameters for coding the laser pulses so that the desired spectral characteristics can be achieved. One of the important coding parameters is the pulse duration. Generally, the pulse duration must be sufficiently short to meet the thermal and stress confinement criteria to effectively generate a PA signal [1]. Thus, there is a trade-off between the confinement criteria and the pulse energy, which is proportional to the pulse duration.
Another important coding parameter is the pulse period, which determines the frequency of the laser excitation. This study used laser pulses with a 50% duty cycle; that is, the pulse period was determined after selecting the pulse frequency. Three linear FM codes were designed to code the diode laser in this study. The FM codes were formed by combining single cycle codes with decreasing frequency. The instantaneous frequency ranges of the 16-cycle FM codes 1, 2 and 3 are designed to be 5.0–2.5 MHz, 6.0–3.0 MHz, and 7.0–3.5 MHz.
The instantaneous code frequency of the FM codes as a function of cycle number is shown in Fig. 1 . Taking FM code 1 as an example, the frequencies of the 16-cycle codes are {5.00 4.83 4.66 4.50 4.33 4.16 4.00 3.83 3.66 3.50 3.33 3.16 3.00 2.83 2.66 2.50 MHz}. Similarly, the frequencies of the 1-, 2-, 4-, 8-, and 16-cycle codes are {5.00 MHz}, {5.00 4.83 MHz}, {5.00 4.83 4.66 4.50 MHz}, and {5.00 4.83 4.66 4.50 4.33 4.16 4.00 3.83 MHz}, respectively.
The spectra of FM codes 1, 2, and 3 with different code cycles are shown in Fig. 2a –2c, respectively. Note that shorter code cycles produce broader bandwidths due to the shorter signal duration. Also, harmonics are evident since the FM codes consist of unipolar square waves.
Fig. 2 (a–c): Amplitudes of frequency component in dB of FM codes 1, 2, and 3, respectively, for different code cycles.
2.2 Experimental setup
The experimental setup for PA signal generation with the proposed coded excitation is shown in Fig. 3 . The FM codes generated by an arbitrary-waveform generator (AWG) (model AWG-420, Tektronix, Tokyo, Japan) were supplied to a high-speed diode laser (model PhoxX-830-140, Omicron, Germany) to generate the corresponding laser pulses. The wavelength of this diode laser is 830 nm and its peak output power is 140 mW for continuous-wave excitation. The rise and fall times of the laser pulse are both less than 1 ns. The dimensions (length × width × height) of the laser head and laser controller are 100 × 40 × 40 mm and 120 × 62 × 40 mm, respectively. A high-frequency 20-MHz single-element PA transducer made with lithium niobate (LiNbO3) was used. The focal depth is 12 mm and the aperture size is 6 mm. The –6 dB fractional bandwidth is 45%. With a 0.65-mm hole in its surface for insertion of an optical fiber, this transducer can achieve effective backward-mode PA detection. Note that using a high-frequency PA transducer will achieve a high spatial resolution in PA imaging, and it is also possible to obtain fine structural information by obtaining an ultrasound image in combination with the PA image. However, the PA signal generated in our experiments had a lower frequency, and so signal averaging was necessary to improve the signal-to-noise ratio (SNR). Finally, the received signals were amplified by a pulser/receiver (model 5077PR, Panametrics, Waltham, MA) and then acquired by a PCI-based A/D card (CompuScope 14200, Gage, Lachine, Canada). The data were stored in a PC for offline processing.
A polyvinylchloride film was used in experiments to demonstrate PA signal generation with different FM codes and different code cycles (i.e., 1, 2, 4, 8 and 16). The SNRs of the decoded PA signals with 16-cycle FM codes 1, 2, and 3 were evaluated. The improvement in the SNR was defined as the SNR of the coded PA signal minus the SNR of the single pulsed PA signal.
3. Results
The one-cycle FM code 1 is shown in Fig. 4a as a solid line. Because the code frequency is 5 MHz, the resultant pulse duration is 100 ns (i.e., half a cycle at 5 MHz). The laser pulse as measured by a photo detector (model ET-2020, Electro-Optics Technology, Traverse City, MI), which is shown as the dotted line in Fig. 4a, was in good agreement with the code generated by the AWG. The generated PA signal is shown in Fig. 4b. Throughout this study, 3000-times averaging was applied to the detected PA signal. The spectra of the code, laser pulse, and the generated PA signal are shown in Fig. 4c. As shown in Fig. 2a, the one-cycle FM code 1 has a broadband spectrum. The frequency response of the generated PA signal by using one-cycle pulse excitation is approximate 3 MHz, which is determined by the pulse frequency, the dimension of the polyvinylchloride film [8,9] as well as optical absorption of the object [10].
Fig. 4 (a): Code signal generated by the AWG and the corresponding diode laser output signal. (b): PA signal generated using the code signal in (a). (c): Spectra of the code signal, laser output signal, and PA signal.
The 2-, 4-, 8-, and 16-cycle FM codes (for code 1) are shown in Fig. 5a –5d, respectively. The instantaneous frequency of the 16-cycle signal in Fig. 5d is clearly decreasing (i.e., FM). The resultant PA signals are shown in Fig. 5e–5h. The corresponding spectra of both the transmitted laser pulse and the generated PA signal are shown in Fig. 5i–5l. As shown in Fig. 2a, the bandwidth of the FM code decreases from a one-cycle code to an eight-cycle code, and then broadens again for longer code cycles; consistent results for the resultant PA signal spectra are shown in Fig. 5i–5l.
Fig. 5 (a–d): Laser excitation codes for the 2-, 4-, 8-, and 16-cycle FM code 1. (e–h): PA signals generated using the code signals in (a)–(d), respectively. (i–l): Spectra of the laser pulses in (a)–(d) and the PA signals in (e)–(h).
The time–frequency representations of the laser excitation codes and the generated PA signals for 2, 4, 8 and 16 pulse cycles are shown in Fig. 6a –6d and Figs. 6e–6h, respectively. Again the results show the good agreement between the spectra of the laser pulses and the generated PA signals. In addition, when 16 cycles are present (Fig. 6d and 6h), both of the signals change linearly with frequency (i.e., from 5.0 to 2.5 MHz).
Fig. 6 (a–d): Time–frequency representations of the laser excitation codes for 2-, 4-, 8-, and 16-cycle FM code 1. (e–h): Corresponding time–frequency representations of the generated PA signals.
The PA signals generated using 16-cycle FM codes 1, 2, and 3 are shown in Fig. 7a , 7e, and 7j, respectively. The calculated linear frequency-modulated signals are shown in Fig. 7b, 7f, and 7j (i.e., 5.0–2.5 MHz for FM code 1, 6.0–3.0 MHz for FM code 2, and 7.0–3.5 MHz for FM code 3). The time-reversed calculated signals served as matched filters for decoding (i.e., pulse compression). The decoded PA signals are shown in Fig. 7c, 7g, and 7k, respectively.
Fig. 7 (a–d): Generated PA signal, calculated linear frequency-modulated signal, decoded PA signal, and spectra of the PA signal and the calculated linear frequency-modulated signal for FM code 1, respectively. (e–h): Results for FM code 2. (i–l): Results for FM code 3.
The spectra of the PA signals and the calculated linear frequency-modulated signal are shown in Fig. 7d, 7h, and 7l. As designed, the −6 dB bandwidths of FM codes 1, 2, and 3 were 5.0–2.5, 6.0–3.0, and 7.0–3.5 MHz, respectively. In addition, the two spectra exhibited good agreement at low frequencies but not at higher frequencies. This effect is clearest in Fig. 7l, and is also evident from the amplitude of the PA signal in Fig. 7i being noticeably lower for the higher frequencies (i.e., at the beginning of the signal). This is possibly caused by the finite response time of the diode laser. In addition, the size of the polyvinylchloride film also affects frequency content.
The SNRs of the decoded PA signals of different FM codes 1, 2, and 3 were also evaluated. The SNR was defined as the square of the peak-to-peak amplitude ratio of the PA signal to a reference noise level. The noise level was calculated by finding the maximum peak-to-peak value of the signal within a region of interest, as illustrated schematically in Fig. 8a .
Fig. 8 (a): Calculation of SNR of the PA signal. (b): SNRs of the pulsed decoded PA signals of FM codes 1, 2, and 3, and the SNR improvements (decoded PA minus pulsed PA).
The SNRs of the pulsed PA signals and the decoded PA signals using 16-cycle FM codes excitation are shown in Fig. 8b. The results for pulsed PA were obtained by using one-cycle FM codes with pulse frequencies of 5, 6, and 7 MHz (dashed line in the figure). The highest SNR of the pulsed PA signal was 10.3 dB, and the SNR decreased to 7.8 dB when the pulse frequency increased. In contrast, the SNRs were much higher for the decoded PA signals than for the pulsed PA signals (thin solid line in the figure): they were 25.5, 27.1, and 26.4 dB for FM codes 1, 2, and 3, respectively. The SNR differences between the pulsed PA signal and the coded PA signal—which are the SNR improvements of the coded excitation method—are shown as the thick solid line in the figure: they were 15.2, 19.3, and 18.5 dB, respectively. Note that the SNR gain from the time-bandwidth product is only around 10.5 dB. The additional improvement came from pulse compression when matched filtering is applied and noise is suppressed during filtering operations.The axial resolutions of the pulsed and decoded PA signals were also obtained by calculating the –6 and –12 dB widths of their signals after envelope detection. The envelopes of the PA signals of FM codes 1, 2, and 3 are shown in Fig. 9a –9c, respectively (the thin and thick lines in the figure indicate the pulse and code inputs, respectively). The results show that the decoded PA signals improve the SNR at the expense of some loss in axial resolution; however, this loss could be reduced by using a different type of pulse compression filter. The –6 and –12 dB widths are summarized in Table 1 .
Table 1. Axial beam widths (in millimeters) of decoded and pulsed PA signals
4. Conclusions
As shown in Fig. 7, the finite response time of the diode laser and the dimension of the polyvinylchloride film could determine the frequency response of PA signal at higher frequencies. Our experimental results indicated that the amplitude of the generated PA signal was more affected when the frequency was higher than around 5 MHz. This also affected the performance of pulse compression, particularly when this employed the driving waveform of the diode laser. Alternative pulse compression methods may improve the pulse compression performance. In addition, prebiased driving waveforms could also be designed by taking the frequency response of the diode laser into account.
The experimental results demonstrated that a frequency-coded PA signal can be generated using a high-speed diode laser with linear FM codes. The trade-off between SNR improvement and axial resolution degradation was also considered. The maximum SNR improvement of the decoded PA signal after compression was 19.3 dB in our study. Under such circumstance, the −6 dB and −12 dB axial widths increased to 2.69 and 2.88 times compared to pulsed PA signal. It also has to be pointed out that there are other factors dictating frequency of the photoacoustic signal produced, such as the size of the optical absorber. On the other hand, the SNR would be changed by using a different duty cycle because as the duty-cycle changes, the pulse energy also changes. With the same laser, we have found that there is a maximum PA response with the pulse duration at 80-ns.
The proposed method utilizes the advantages of lower cost and compactness of a diode laser relative to a conventional Nd:YAG laser.
Acknowledgments
Financial support from the National Science Council, the National Health Research Institutes, NTU Center for Genomic Medicine, and the NTU Nano Center for Science and Technology is gratefully acknowledged.
References and links
1. M. Xu and L. V. Wang, “Photoacoustic imaging in biomedicine,” Rev. Sci. Instrum. 77(4), 041101–041122 (2006). [CrossRef]
2. K. Maslov and L. V. Wang, “Photoacoustic imaging of biological tissue with intensity-modulated continuous-wave laser,” J. Biomed. Opt. 13, 024006 (2008). [CrossRef] [PubMed]
3. T. J. Allen and P. C. Beard, “Pulsed near-infrared laser diode excitation system for biomedical photoacoustic imaging,” Opt. Lett. 31(23), 3462–3464 (2006). [CrossRef] [PubMed]
4. T. J. Allen, and P. C. Beard, “Dual wavelength laser diode excitation source for 2D photoacoustic imaging,” Proc. SPIE 6437, 1U1–1U9 (2007).
5. R. Y. Chiao and X. Hao, “Coded excitation for diagnostic ultrasound: a system developer’s perspective,” IEEE Trans. Ultrason. Ferroelectr. Freq. Control 52(2), 160–170 (2005). [CrossRef] [PubMed]
6. M. P. Mienkina, A. Eder, G. Schmitz, C. S. Friedrich, N. C. Gerhardt, and M. R. Hofmann, “Simulation study of photoacoustic coded excitation using Golay codes,” Proc. IEEE Ultrason. Symp. 1242–1245 (2008).
7. M. P. Mienkina, C. S. Friedrich, N. C. Gerhardt, M. F. Beckmann, M. F. Schiffner, M. R. Hofmann, and G. Schmitz, “Multispectral photoacoustic coded excitation imaging using unipolar orthogonal Golay codes,” Opt. Express 18(9), 9076–9087 (2010). [CrossRef] [PubMed]
8. G. J. Diebold, M. I. Khan, and S. M. Park, “Photoacoustic “signatures” of particulate matter: optical production of acoustic monopole radiation,” Science 250(4977), 101–104 (1990). [CrossRef] [PubMed]
9. G. J. Diebold, T. Sun, and M. I. Khan, “Photoacoustic monopole radiation in one, two, and three dimensions,” Phys. Rev. Lett. 67(24), 3384–3387 (1991). [CrossRef] [PubMed]
10. P.-C. Li, C.-W. Wei, and Y.-L. Sheu, “Subband photoacoustic imaging for contrast improvement,” Opt. Express 16(25), 20215–20226 (2008). [CrossRef] [PubMed]
