Review of Low-Cost Photoacoustic Sensing and Imaging Based on Laser Diode and Light-Emitting Diode

sensors

Review Review of Low-Cost Photoacoustic Sensing and Imaging Based on Laser Diode and Light-Emitting Diode

Hongtao Zhong 1, Tingyang Duan 1, Hengrong Lan 1, Meng Zhou 1,2 and Fei Gao 1,* 1 Hybrid Imaging System Laboratory, School of Information Science and Technology, ShanghaiTech University, Shanghai 201210, China; [email protected] (H.Z.); [email protected] (T.D.); [email protected] (H.L.); [email protected] (M.Z.) 2 Shanghai Advanced Research Institute, Chinese Academy of Sciences, Shanghai 201210, China * Correspondence: [email protected]; Tel.: +86-021-2068-5379

Received: 12 May 2018; Accepted: 25 June 2018; Published: 13 July 2018

Abstract: Photoacoustic tomography (PAT), a promising medical imaging method that combines optical and ultrasound techniques, has been developing for decades mostly in preclinical application. A recent trend is to utilize the economical laser source to develop a low-cost sensing and imaging system, which aims at an affordable solution in clinical application. These low-cost laser sources have different modulation modes such as pulsed modulation, continuous modulation and coded modulation to generate different proﬁles of PA signals in photoacoustic (PA) imaging. In this paper, we review the recent development of the photoacoustic sensing and imaging based on the economical laser sources such as laser diode (LD) and light-emitting diode (LED) in different kinds of modulation types, and discuss several representative methods to improve the performance of such imaging systems based on low-cost laser sources. Finally, some perspectives regarding the future development of portable PAT systems are discussed, followed by the conclusion.

Keywords: photoacoustics imaging; sensing; low-cost laser source; laser diode

1. Introduction PAT is a kind of hybrid imaging modality, which employs the high imaging contrast of optical imaging and deep penetration of ultrasound imaging [1–5]. In PAT, the tissue is irradiated by the sufficient short-pulsed laser beam, and absorbs the energy partially, and then converts the absorbed energy into heat. The transient temperature elevation causes the expansion and contraction of the tissue and then produces the ultrasound wave, which will be further received by ultrasound transducers for image reconstruction [6–12]. PAT can be used in vivo for tumor detection, blood oxygenation mapping, animal or human organs imaging, etc. [13]. Conventionally, the light source usually comprises of a Q-switched Nd:YAG laser with nanosecond pulse duration and hundreds of mJ pulse energy, which is commonly used in different kinds of PAT imaging implementations, such as optical-resolution photoacoustic microscopy (OR-PAM), acoustic-resolution photoacoustic microscopy (AR-PAM) and photoacoustic computed tomography (PACT) [14–19]. From organelles to small animals or human organ imaging, and covering functional, metabolic and histologic imaging, the flash-lamp-pumped laser source can provide the sufficient pulse energy (several hundred mJ) and enough signal-to-noise ratio (SNR) of PA signals. The above-mentioned works are summarized in References [13,20–22]. However, this kind of laser source is widely used in the laboratory but is not economical and compact in clinical application since its high price and bulky size. Moreover, the repetition rate of most high-power laser sources is relatively low (~tens of Hz), which may limit the imaging speed when data averaging is required. Additionally,

Sensors 2018, 18, 2264; doi:10.3390/s18072264 www.mdpi.com/journal/sensors Sensors 2018, 18, x FOR PEER REVIEW 2 of 24

Sensorssources2018 is, relatively18, 2264 low (~tens of Hz), which may limit the imaging speed when data averaging2 of 24is required. Additionally, another type of laser source, diode pumped laser such as that mentioned in Reference [23], can operate in the repetition rate of 50 Hz with the 80 mJ per pulse. another type of laser source, diode pumped laser such as that mentioned in Reference [23], can operate In recent years, the low-cost and compact photoacoustic imaging system has been developing in the repetition rate of 50 Hz with the 80 mJ per pulse. fast by utilizing laser diodes or light emitting diodes. Although the output peak energy of these In recent years, the low-cost and compact photoacoustic imaging system has been developing semiconductor devices are low (nano-joule to micro-joule per pulse), it has the following advantages: fast by utilizing laser diodes or light emitting diodes. Although the output peak energy of these (1)semiconductor The small sizes devices can are reduce low (nano-joulethe volume toof micro-joulethe imaging per system pulse), significantly; it has the following advantages: (2) Commercially available, which can be easily obtained from the industrial market; (1)(3) HighThe small repetition sizes rate, can reduce up to tens the volumeof kilohertz, of the which imaging can systemcompensate signiﬁcantly; for the low output energy to (2) improveCommercially the SNR available, by averaging, which canwhich be is easily promising obtained to achieve from the real-time industrial imaging; market; (3)(4) DoesHigh not repetition require rate, the active up to tenscooling of kilohertz, system si whichnce the can pulse compensate duty cycle for is theless low than output 1%; energy to (5) Theimprove stability the SNRwell meets by averaging, the photoacoustic which is promising imaging tosystem achieve as the real-time laser diode imaging; can maintain the (4) constantDoes not output require energy the active when cooling working system over since 3000 theh [24]. pulse duty cycle is less than 1%; (5) The stability well meets the photoacoustic imaging system as the laser diode can maintain the In addition to the above advantages, the ability of photoacoustic imaging by using these constant output energy when working over 3000 h [24]. semiconductor devices has also been verified in many experiments such as those in References [24– 28]. In addition to the above advantages, the ability of photoacoustic imaging by using these semiconductorBecause such devices semiconductor has also been laser veriﬁed diodes in can many oper experimentsate in high frequency such as those with in high-speed References current [24–28]. driver,Because there suchare several semiconductor modulation laser diodesways in can sign operateal excitation, in high frequencysuch as pulsed with high-speed modulation current and continuousdriver, there modulation. are several Pulsed modulation modulation ways incan signal be further excitation, divided such into as short-pulsed pulsed modulation modulation and basedcontinuous on LDs, modulation. short-pulse Pulsed modulation modulation based on can LE beDs, further long-pulse divided modulation into short-pulsed and coded modulation excitation. Continuousbased on LDs, modulation short-pulse can modulation be further based divided on LEDs,into chirp long-pulse modulation, modulation sinusoidal and codedmodulation excitation. and hybridContinuous modulation. modulation Here, can this be furtherreview dividedpaper will into chirpdiscuss modulation, the recent sinusoidaldevelopment modulation of low-cost and photoacoustichybrid modulation. sensing Here, and imaging this review system paper catego willrized discuss by thedifferent recent modulation development ways. of low-cost In the photoacousticmeantime, some sensing signal and enhancemen imaging systemt methods categorized to improve by different the SNR modulation of PA imaging ways. system In the meantime, will also besome discussed. signal enhancement methods to improve the SNR of PA imaging system will also be discussed.

2. Pulse Excitation Using laser laser diode diode to to generate generate short short pulse pulse is isa very a very common common method method to excite to excite the thePA PAsignals. signals. To Togenerate generate the thePA signals PA signals efficiently, efficiently, the thermal the thermal and stress and stressconfinements confinements should should be satisfied. be satisfied. These Theserequirements requirements mean that mean the that pulse the width pulse of width the lase ofr the should laser be should short beenough; short enough;and in order and to in attain order enoughto attain SNR, enough the SNR,laser theenergy laser per energy pulse per should pulse be should large beenough. large enough.Meanwhile, Meanwhile, utilizing utilizing LD aims LD at decreasingaims at decreasing the volume the and volume cost of and the cost imaging of the system imaging potentially system potentially for point-of-care for point-of-care testing. Thus, testing. such Thus,a short such and ahigh-energy short and high-energy pulse is a challenge pulse is aduring challenge the duringdevelopment the development of a laser system. of a laser system. From the perspective of the laser driver circuit, the the typical typical driver driver circuit circuit shows shows as as in in Figure Figure 11.. The operation principle of this circuit is based on the charge and discharge of the capacitor for pulse generation. The The pulse pulse current current will will flow flow through through from from LD to the switch device as illustrated by the arrow in in Figure Figure 11.. In In order order to to attain attain the the minimi minimizedzed pulse pulse width width and and maximize maximize the the peak peak current, current, the parasiticthe parasitic inductance inductance (Lr, (L Lcr, Lcand and Ld) L shouldd) should be beas assmall small as as possible possible [29]. [29]. Additionally, Additionally, these these kind kind of circuits are not able to modulate the pulse width sinc sincee the time constant of the circuit is fixed. fixed. To To our best knowledge, the ultra-compact pu pulselse laser driver is reported in Reference [[24],24], which has the size of 40 ×5050 mm mm2,2 up, up to to 10 10 kHz kHz of of repetition repetition and and the the energy energy of of 1.7 1.7 mJ mJ per per pulse pulse within within 40 40 ns ns width. width.

FigureFigure 1. 1. StructureStructure of of the the typical typical laser laser diode diode driver driver.. Reproduced withwith permissionpermission fromfrom Ref.Ref. [[29].29].

Sensors 2018, 18, 2264 3 of 24

Sensors 2018, 18, x FOR PEER REVIEW 3 of 24 2.1. Short-Pulse Modulation Based on Laser Diode 2.1. Short-Pulse Modulation Based on Laser Diode There are many literatures on PA imaging based on laser diode by short-pulse modulation. Here, we willThere introduce are many some literatures of typical on and PA recentimaging studies. based on laser diode by short-pulse modulation. Here, we willDynamic introducein vivo somephotoacoustic of typical and imaging recent studies. is a signiﬁcant research area, which can monitor the bloodDynamic velocity, in oxygen vivo photoacoustic saturation and imaging circulating is a significant tumor to research diagnose area, the which processing can monitor of disease. the Smallblood animalsvelocity, playoxygen a signiﬁcant saturation roleand incirculating preclinical tumor studies to diagnose and effectively the processing reveal of the disease. experiment Small results.animals Inplay Reference a significant [30], role the in dynamic preclinicalin vivo studiesimaging and effectively in a mouse reveal brain the was experiment achieved results. by using In theReference pulsed [30], laser the diode dynamic as an in illumination vivo imaging source. in a mouse This system brain was utilizes achieved an 803 by nm using pulsed the pulsed laser diode laser anddiode a fast-scanningas an illumination single-element source. This ultrasound system utiliz transducer.es an 803 nm The pulsed repetition laser rate diode of and the pulseda fast-scanning laser is aboutsingle-element 7 kHz and ultrasound its cross-sectional transducer. imaging The repetition time is about rate of 5 sthe with pulsed a signal-to-noise laser is about ratio 7 kHz of and about its 48cross-sectional dB. The spatial imaging resolution time is is about about 185 5 µsm with for 5a MHzsignal-to-noise transducers. ratio The of demonstration about 48 dB. ofThe dynamic spatial imagingresolution was is performedabout 185 byum monitoring for 5 MHz the transducers. uptake and The clearance demonstration process of of indocyanine dynamic imaging green (ICG) was inperformed the cortical by monitoring region of a the mouse uptake brain, and whichclearance is shown process in of Figure indocyanine2. It is green not very (ICG) clear in the to cortical show theregion deduction of a mouse of ICG brain, from which the is imaging shown in results Figure in 2. Figure It is not2d–f, very but clear from to Figureshow the2g, deduction the amplitude of ICG offrom the the PA signalimaging is decreasingresults in Figure apparently 2d–f, frombut from point Figure (d) to point2g, the (e). amplitude Additionally, of the as PA mentioned signal is indecreasing Reference apparently [30], the lateral from resolutionpoint (d) to of point this system(e). Additionally, could be further as mentioned improved in byReference using multiple [30], the ultrasoundlateral resolution transducers. of this system could be further improved by using multiple ultrasound transducers.

FigureFigure 2.2.In Invivo vivoPLD-PAT PLD-PAT brain brain imaging imaging and and pharmacokinetics pharmacokinetics ofof ICGICG intravenouslyintravenously injectedinjected inin toto ratrat brain:brain: PhotographPhotograph ofof ratrat brainbrain beforebefore ((aa)) andand afterafter ((bb)) removingremoving thethe scalp.scalp.In In vivovivobrain brain imagesimages atat differentdifferent scanscan timestimes ((cc)) 00 s, (d) 2 min, ( (ee)) 6 6 min, min, and and ( (ff)) 13 13 min. min. (g (g) )Graph Graph shows shows the the quantification quantiﬁcation of ofICG ICG signal signal in in the the superior superior SS SS during during 25 25 min min follo followingwing injection. The red arrow indicatesindicates thethe ICGICG injectioninjection point. point. ReproducedReproduced withwith permission permission from from Ref. Ref. [ 30[30].].

TheThe aboveabove imaging system system verify verify the the possibility possibility of of dynamic dynamic imaging imaging in invivo. vivo Moreover,. Moreover, the thefeasibility feasibility of diagnosing of diagnosing some some diseases diseases such suchas synovitis as synovitis is also is demonstrated also demonstrated [31]. The [31 ].image The system image systemis shown is in shown Figure in 3a, Figure which3a, is which a handheld is a handheld probe integrated probe integrated with a laser with diode a laser source, diode laser source, driver, laserultrasound driver, transducers ultrasound and transducers a prism [32]. and aThe prism laser [32 source]. The in laser this sourceresearch in is this at 808 research nm wavelength. is at 808 nm It wavelength.can provide It1.3 can mJ provide per pulse 1.3 mJwith per 90 pulse ns duration with 90 an nsd duration its repetition and its rate repetition can be rateup to can 10 be kHz. up toIn 10Reference kHz. In Reference[31], it shows [31], multimodality it shows multimodality photoacoustic photoacoustic and ultrasound and ultrasound imaging imaging of synovitis of synovitis in finger in fingerjoints.joints. The clinically The clinically evident evident synovitis synovitis is detect is detecteded by by comparing comparing the the inflamed inflamed and non-inflamednon-inflamed proximalproximal interphalangealinterphalangeal joints,joints, whichwhich isis shownshown inin FigureFigure3 b.3b. The The clinical clinical study study with with a a low-cost low-cost and and compactcompact photoacoustic photoacoustic and and ultrasound ultrasound probe probe is firstlyis firs demonstrated.tly demonstrated. This This handheld handheld probe probe is also is used also used in the flow imaging system, which can achieve quantitative flow velocities from 12 to 75 mm/s [33]. Such a compact solution paves the way for clinical investigation of a photoacoustic imaging system based on economical pulse laser diode.

Sensors 2018, 18, 2264 4 of 24 in the ﬂow imaging system, which can achieve quantitative ﬂow velocities from 12 to 75 mm/s [33]. Such a compact solution paves the way for clinical investigation of a photoacoustic imaging system Sensorsbased 2018 on economical, 18, x FOR PEER pulse REVIEW laser diode. 4 of 24

Figure 3. TheThe imaging imaging setup setup (a ()a and) and PA/US PA/US and and US/PD US/PD images images of an of inflamed an inflamed (upper (upper row) row)and non- and inflamednon-inflamed contra-lateral contra-lateral joint (bottom joint (bottom row) of row) an RA of patient an RA patient(b). PA/US (b). images PA/US in images (I) show in a (I)difference show a indifference color between in color inflamed between and inflamed non-inflamed, and non-inflamed, corresponding corresponding to an increase to in an amplitude increase inlevels. amplitude When discardinglevels. When low discarding PA amplitudes low PA in (II), amplitudes only features in (II), in only the inflamed features injoint the are inflamed visible. jointCorresponding are visible. US-PDCorresponding images are US-PD shown images in (III). are shownThe blue in line (III). in The the blue PA/US line images in the PA/US indicates images the ROI indicates used thefor quantificationROI used for quantification of PA features ofin PAthe featuressynovial in spac thee. synovial The 0 dB space. level is The the 0maximum dB level isPA the amplitude maximum from PA theamplitude inflamed from joint. the d inflamed = dermis; joint. dv = ddorsal = dermis; vein;dv pp = = dorsalproximal vein; phalanx; pp = proximal pip = proximal phalanx; interphalangeal pip = proximal joint;interphalangeal mp = middle joint; phalanx; mp = s middle= synovium; phalanx; t = extens s = synovium;or tendon. tReproduced = extensor tendon.with permission Reproduced from with Ref. permission[31]. from Ref. [31].

The low output energy of the pulse diode causes many times the averaging signal required to The low output energy of the pulse diode causes many times the averaging signal required to improve the SNR of PA signals, which will limit the clinical application of the pulsed laser diode improve the SNR of PA signals, which will limit the clinical application of the pulsed laser diode since the imaging speed is not enough and the imaging penetration will be significantly limited. Thus, since the imaging speed is not enough and the imaging penetration will be signiﬁcantly limited. improving the energy is of paramount importance. The proposed PAM system reported in Reference Thus, improving the energy is of paramount importance. The proposed PAM system reported [34] utilizes the laser source that can output 325 W peak power pulse within 50ns. It is not necessary in Reference [34] utilizes the laser source that can output 325 W peak power pulse within 50 ns. for averaging during data acquisition in this system. Therefore, the imaging speed can be improved It is not necessary for averaging during data acquisition in this system. Therefore, the imaging significantly, which reaches ~370 A-lines per second. Moreover, there is no need for mechanical speed can be improved signiﬁcantly, which reaches ~370 A-lines per second. Moreover, there is no scanning of the sample since two galvanometer scanning mirrors serve as the two-dimensional need for mechanical scanning of the sample since two galvanometer scanning mirrors serve as the scanning of the beam through an aspheric focusing lens. The mouse ear ex vivo imaging shown in Figure 4 verified the feasibility of this system. Another method to improve output energy is to combine several laser diodes [25].

Sensors 2018, 18, 2264 5 of 24 two-dimensional scanning of the beam through an aspheric focusing lens. The mouse ear ex vivo imaging shown in Figure4 veriﬁed the feasibility of this system. Another method to improve output energySensors 2018 is to, 18 combine, x FOR PEER several REVIEW laser diodes [25]. 5 of 24

Figure 4. ((aa,,cc)) PAM PAM image image of of the the vasculature vasculature onon aa mousemouse earear exex vivo.vivo. The color bars represent normalized PA amplitude. (b,d) Photograph of the mouse ear in ( a,,c), respectively. Reproduced with permission from Ref. [[34].34].

In addition to the above-mentioned reports, there are also many other interesting studies on In addition to the above-mentioned reports, there are also many other interesting studies on photoacoustic imaging based on pulsed laser diode. Imaging the custom-made phantoms such as photoacoustic imaging based on pulsed laser diode. Imaging the custom-made phantoms such as blood blood vessel phantoms or other tissue-mimicking phantoms are developed by many imaging systems vessel phantoms or other tissue-mimicking phantoms are developed by many imaging systems [35–42]. [35–42]. For ex vivo imaging, several imaging systems are tested on chicken breast, mouse ear and For ex vivo imaging, several imaging systems are tested on chicken breast, mouse ear and porcine porcine ovary etc. [43–50]. For in vivo imaging, blood vessels [26,51,52], hemoglobin monitor [53] and ovary etc. [43–50]. For in vivo imaging, blood vessels [26,51,52], hemoglobin monitor [53] and animal animal brain [54] imaging systems are also studied based on pulsed laser diodes. For photoacoustic brain [54] imaging systems are also studied based on pulsed laser diodes. For photoacoustic sensing sensing based on such economical laser sources, some systems are developed for gas detection based on such economical laser sources, some systems are developed for gas detection [55,56]. [55,56]. Last but not least, it should be noted that laser diode has two types. One is the pulsed laser diode Last but not least, it should be noted that laser diode has two types. One is the pulsed laser diode and another one is the continuous laser diode. Most photoacoustic imaging systems use the pulsed and another one is the continuous laser diode. Most photoacoustic imaging systems use the pulsed laser diode. Recently, the continuous laser diode is verified for its feasibility of working at pulsed laser diode. Recently, the continuous laser diode is verified for its feasibility of working at pulsed mode by overdriving while avoiding the catastrophic optical damage and thermal damage [57]. It can mode by overdriving while avoiding the catastrophic optical damage and thermal damage [57]. It be a complementary solution of the pulsed laser diode since the continuous laser diode has a broader can be a complementary solution of the pulsed laser diode since the continuous laser diode has a available wavelength. broader available wavelength. 2.2. Short-Pulse Modulation Based on LEDs 2.2. Short-Pulse Modulation Based on LEDs By overdriving the high power LEDs beyond their continuous current damage threshold, LEDsBy can overdriving be another the alternative high power pulsed LEDs lightbeyond emission their continuous source, which current are damage more economical threshold, LEDs than LDscan be and another less sensitive alternative to static pulsed electricity light emission [58]. Theoretically, source, which both are LEDs more and economical LDs are semiconductor than LDs and devices,less sensitive which to means static theirelectricity driver [58]. circuits Theoreti are similarcally, both [58– 60LEDs]. The and fabrication LDs are semiconductor process and technology devices, ofwhich LEDs means allow their multi-wavelength driver circuits inare a similar single LED[58–60]. [61]. The The fabrication multi-wavelength process and LEDs technology can be controlled of LEDs toallow an output-specificmulti-wavelength wavelength in a single by LED the [61]. electronic The multi-wavelength signals, which is LEDs an attractive can be controlled characteristic to an in multi-wavelengthoutput-specific wavelength PA imaging. by The the multi-wavelengthelectronic signals, characteristic which is an ofattractive LED can characteristic significantly reduce in multi- the sizewavelength of the illumination PA imaging. system The multi-wavelength since it is not necessary characteristic to combine of LED several can LDs significantly in different reduce wavelengths. the size However,of the illumination the single-wavelength system since lightit is not intensity necessary from to multi-wavelength combine several LEDsLDs in will different decrease wavelengths. because the totalHowever, chip areathe single-wavelength is constant, which islight critical intensity in PA from imaging. multi-wavelength Moreover, the LEDs efficiency will decrease of delivering because the the total chip area is constant, which is critical in PA imaging. Moreover, the efficiency of delivering the light power of LEDs from optical fiber is very low (10~20%), which means the higher power LEDs are needed to provide enough light density on the imaging object [62]. This will increase the volume of LEDs illumination system. The different principle of light emission between LDs and LEDs leads to the broader spectrum in the emission wavelength from LEDs shown in Figure 5 [61]. According to the temperature sensitive characteristics of semiconductor devices, the emission wavelength may

Sensors 2018, 18, 2264 6 of 24 light power of LEDs from optical ﬁber is very low (10~20%), which means the higher power LEDs are needed to provide enough light density on the imaging object [62]. This will increase the volume of LEDs illumination system. The different principle of light emission between LDs and LEDs leads to Sensorsthe broader 2018, 18 spectrum, x FOR PEER in REVIEW the emission wavelength from LEDs shown in Figure5[ 61]. According to6 of the 24 temperature sensitive characteristics of semiconductor devices, the emission wavelength may shift shift considerably as the operation temperature increases [58]. This feature may influence the considerably as the operation temperature increases [58]. This feature may inﬂuence the accuracy of accuracy of PA imaging results. It should be noticed that the low prices of LEDs enable it to still be PA imaging results. It should be noticed that the low prices of LEDs enable it to still be attractive in attractive in PA imaging. PA imaging.

Figure 5. IndividualIndividual spectra spectra of of 23 23 LEDs. LEDs. Reproduc Reproduceded with with permission permission from Ref. [63]. [63].

In the field field of PA imagingimaging and sensing based on short-pulsed modula modulation,tion, gas gas monitoring, monitoring, phantom imaging and in vivo imaging were demonstrated. Gas monitoringmonitoring systems systems were were developed developed in the in early the stageearly ofstage LEDs. of Lay-Ekuakille LEDs. Lay-Ekuakille et al. developed et al. developed a multi-gas sensing system for estimating the concentration of anaesthetic gas and nitrous a multi-gas sensing system for estimating the concentration of anaesthetic gas and nitrous oxide (N2O) oxidein hospital (N2O) to in protect hospital the to facultiesprotect the and faculties patients and [64 patients]. Based [64]. on 4–4.5 Basedµm on LEDs, 4–4.5 thisμm LEDs, system this detects system an detectsattention an threshold attention levelthreshold of gas level in a roughof gas in range. a roug Otherh range. gas detectionOther gas such detectio as ozonen such by as blue ozone LEDs by blue with LEDs460 nm with and 460 water nm vapor and water by IR vapor LEDs withby IR 1450 LEDs nm wi wereth 1450 reported nm were in Reference reported [ 65in, 66Reference], respectively. [65,66], respectively.To image a tissue mimicking phantom, a 623 nm LED was overdriven at 50 A which is 20 times largerTo than image its nominala tissue mimicking current in Reference phantom, [ 27a 623]. Figure nm LED6a shows was overdriven the setup of at the 50 PAT A which imaging is 20 system times largercomprising than ofits three nominal tubes current filled with in Reference human blood. [27]. TheFigure energy 6a shows per pulse the is setup 9 µJ with of the 500 PAT Hz repetition imaging systemrate. The comprising duty cycle of is wellthree below tubes 1%filled to preventwith human the LED blood. damage. The energy Figure per6b is pulse the reconstructed is 9 μJ with 500 image Hz repetitionof detected rate. time The resolved duty PA cycle signals is well as a functionbelow 1% of scanto prevent angle. Itthe shows LED that damage. the penetration Figure 6b depth is the of reconstructedthis system can image be 1.5 of cm. detected Figure6 timec shows resolved the reconstructed PA signals as PA a imagingfunction of of three scan tubes. angle. The It shows diameters that theof the penetration tubes in this depth image of this is within system 0.1 can mm, be 1.5 which cm.is Figure close to6c theshows real the values. reconstructed Moreover, PA this imaging work has of threeutilized tubes. a compact The diameters four wavelength of the tubes LED, in this shown image in Figureis within7, to 0.1 implement mm, which two is close excitation to the strategies real values. to Moreover,increase SNR. this One work is tohas use utilized all four a wavelengthscompact four simultaneously wavelength LED, to design shown a in Wiener Figure filter, 7, to andimplement another twoone isexcitation to use the strategies extended to Golay increase coded SNR. excitation. One is to The use latterall four excitation wavelengths method simultaneously shows the potential to design for afaster Wiener data filter, acquisition. and another The multi-wavelengthone is to use the exte imagingnded Golay based coded on several excitation. single-wavelength The latter excitation LEDs is methoddescribed shows in Reference the potential [67], which for faster illustrates data theacquisition. color imaging The multi-wavelength system. imaging based on severalXianjing single-wavelength Dai et al. demonstrated LEDs is described the in vivo in imagingReference of [67], vasculature which illustrates based on LEDsthe color [60]. imaging A 1.2 W system.405 nm LED with 200 ns pulses and repetition rate of 40 kHz was used to excite the photoacoustic signalsXianjing in the Dai PAM et al. imaging demonstrated system (shownthe in vivo in Figureimaging8a). of Figurevasculature8b shows based an on ultra-compact LEDs [60]. A LED1.2 W 405compared nm LED with with a coin.200 ns Figure pulses8c,e and are repetition the maximum rate of amplitude 40 kHz was projection used to image excite and the3D photoacoustic volumetric signalsimage underin the thePAM dotted imaging red squaresystem in (shown 8f, respectively. in Figure The8a). dottedFigure red8b squareshows an is theultra-compact scanning area LED of compared4 mm × 4 mm.with a Figure coin. 8Figured is the 8c,e cross-sectional are the maximu imagem amplitude along the dottedprojection line image in 8c. and The 3D signal volumetric of each imagescanning under point the averaged dotted red 4000 square times andin 8f, the respectively imaging speed. The isdotted 10 A-lines red square per second. is the Thisscanning study area shows of 4the mm potential × 4 mm. of Figure a portable 8d is handheld the cross-sectional photoacoustic image imaging along devicethe dotted based line on in LEDs. 8c. The signal of each scanning point averaged 4000 times and the imaging speed is 10 A-lines per second. This study shows the potential of a portable handheld photoacoustic imaging device based on LEDs.

Sensors 2018, 18, 2264 7 of 24 Sensors 2018, 18, x FOR PEER REVIEW 7 of 24 SensorsSensors 2018, 18, x FOR PEERPEER REVIEWREVIEW 7 of7 of24 24

Figure 6.6. ((a)) PhotoacousticPhotoacoustic imaging imagingimaging setup. setup. setup. ( (b( b)) )Time-resolved Time-resolved Time-resolved photoacoustic photoacoustic photoacoustic signals signals signals of of of three three three 1.4 1.4 1.4 mm mmmm −1 −1 tubes filled withwith humanhuman bloodblood (35%(35% haematocrit)haematocrit) and and immersed immersed in in 1% 1% Intralipid Intralipid (μ (sμ’ s=’ =10 1mm mm),−1 ),(c− )(1 c) tubes filled ﬁlled with with human human blood blood (35% (35% haematocrit) haematocrit) and and immersed immersed in in 1% 1% Intralipid Intralipid (μ (sµ’ s= 1= mm 1 mm), (c),) (reconstructedreconstructedc) reconstructed photoacousticphotoacoustic photoacoustic image.image. image. PP == P9 9 =μ μJ, 9J,J, µN NNJ, = = N= 5000. 5000.5000. = 5000. Reproduced ReproducedReproduced Reproduced with withwith with permission permissionpermission permission from fromfrom from Ref. Ref.Ref. Ref. [27]. [27].[27]. [ 27 ].

Figure 7. Schematic and photograph of the four wavelength LED (460, 530, 590, and 620 nm). Figure 7.7. SchematicSchematicSchematic andand and photographphotograph photograph ofof of thethe the fourfour four wavelength wavelength LED LED (460, (460, 530, 530, 590, 590, and and 620 620 nm). nm). Reproduced with permission from Ref. [27]. Reproduced with permission fromfrom Ref.Ref. [[27].27].

Figure 8. In vivo photoacoustic imaging of mouse ear vasculature. (a) LED-PAM system. L1, L2. Lens; Figure 8. In vivo photoacoustic imaging of mouse ear vasculature. (a) LED-PAM system. L1, L2. Lens; FigureSH, sample 8. InIn vivo vivoholder; photoacousticphotoacoustic UST, ultrasound imaging imaging transducer. of of mouse mouse (ear earb) Photographvasculature. vasculature. of( (aa ))LED. LED-PAM LED-PAM (c) Maximum system. L1, L1,amplitude L2. L2. Lens; Lens; SH, sample holder; UST, ultrasound transducer. (b) Photograph of LED. (c) Maximum amplitude SH,projection sample (MAP) holder; image. UST, ( ultrasoundd) Cross-sectional transducer. imaging. ( b)) Photograph(e) 3D volumetric of LED. image. (c) Maximum (f) Photograph amplitude of projection (MAP) image. (d) Cross-sectional imaging. (e) 3D volumetric image. (f) Photograph of projectionmouse ear. (MAP)All the image.scale bars ( d )indicate Cross-sectional 0.5 mm in imaging. length. Reproduced ( (ee)) 3D volumetric with permission image. ( ( fffrom)) Photograph Ref. [60]. of mouse ear. All the scale bars indicate 0.5 mm in length. Reproduced with permission from Ref. [60]. mouse ear. All the scale bars indicate 0.5 mm in length. Reproduced with permission from Ref.Ref. [[60].60]. More recently, a portable light emitting diode—based photoacoustic imaging (PLED-PAI) More recently, a portable light emitting diode—based photoacoustic imaging (PLED-PAI) systemMore with recently,recently, a handheld a portablea portable probe light waslight emitting developed emitting diode—based [68].diode—based Figure photoacoustic 9a photoacousticis the handheld imaging imaging probe (PLED-PAI) in (PLED-PAI) this systemdual system with a handheld probe was developed [68]. Figure 9a is the handheld probe in this dual withsystemmodality a handheldwith (ultrasound/photoacousti a handheld probe was probe developed wasc) developedimaging [68]. Figure system. [68].9 aFigure isIt theintegrates handheld9a is the with handheld probe 2 LED in thisprobe arrays dual in and modalitythis 128dual modality (ultrasound/photoacoustic) imaging system. It integrates with 2 LED arrays and 128 (ultrasound/photoacoustic)modalityelements of(ultrasound/photoacousti an ultrasound transducer. imagingc) system. Eachimaging LED It integratessystem. array has It withfourintegrates 2rows LED of arrayswith 36 single 2 and LED embedded 128 arrays elements and LEDs. of 128 an elementsThe emitted of anpulse ultrasound width allows transducer. changing Each from LED 50 nsarray to 150 has ns four with rows a 5 ns of step 36 single size and embedded the repetition LEDs. The emitted pulse width allows changing from 50 ns to 150 ns with a 5 ns step size and the repetition The emitted pulse width allows changing from 50 ns to 150 ns with a 5 ns step size and the repetition

Sensors 2018, 18, 2264 8 of 24 ultrasound transducer. Each LED array has four rows of 36 single embedded LEDs. The emitted pulse widthSensors allows 2018, 18 changing, x FOR PEER from REVIEW 50 ns to 150 ns with a 5 ns step size and the repetition rate can be8 tunedof 24 between 1 kHz, 2 kHz, 3 kHz and 4 kHz. This system has the frame rates of 30 Hz, 15 Hz, 10 Hz, 6 Hz, rate can be tuned between 1 kHz, 2 kHz, 3 kHz and 4 kHz. This system has the frame rates of 30 Hz, 3 Hz, 1.5 Hz, 0.6 Hz, 0.3 Hz and 0.15 Hz, which depends on the choice of repetition rates of LEDs. 15 Hz, 10 Hz, 6 Hz, 3 Hz, 1.5 Hz, 0.6 Hz, 0.3 Hz and 0.15 Hz, which depends on the choice of repetition The axial resolution is 268 µm and the lateral resolution is between 550 µm and 590 µm under the 70 ns rates of LEDs. The axial resolution is 268 μm and the lateral resolution is between 550 μm and 590 pulse width. The penetration depth can be 3.2 cm when detecting a pencil lead inside chicken breast μm under the 70 ns pulse width. The penetration depth can be 3.2 cm when detecting a pencil lead withinside a frame chicken rate breast of 15 Hz.with The a frame dual rate modality of 15 imagingHz. The dual of a freshmodality enucleated imaging rabbit of a fresh eye illustrates enucleated the clinicalrabbit utilityeye illustrates of PLED-PAI, the clinical which utility is shown of PLED-PAI, in Figure8 b–e.which The is yellowshown dottedin Figure box 8b–e. shows The clearly yellow the retinaldotted vessels box shows in 2 cmclearly depth. the Figureretinal8 vesselsf is the in imaging 2 cm de resultpth. Figure of a human 8f is the wrist, imaging which result shows of a human the skin andwrist, vessels which by shows yellow the arrows. skin and vessels by yellow arrows.

FigureFigure 9. 9.(a ()a PLED-PAI) PLED-PAI probe probe with with imagingimaging plane and illu illuminationmination source source are are shown shown schematically. schematically. LEDLED array array design design is is also also shownshown inin thethe inset—thereinset—there were alternating rows rows of of LEDs LEDs with with different different wavelengths. (b–e) Evaluation of PLED-PAI on rabbit eye. (b) B-mode ultrasound image when fresh wavelengths. (b–e) Evaluation of PLED-PAI on rabbit eye. (b) B-mode ultrasound image when fresh enucleated rabbit eye was embedded in 1% agar. (c) B-mode photoacoustic/ultrasound image of rabbit enucleated rabbit eye was embedded in 1% agar. (c) B-mode photoacoustic/ultrasound image of rabbit eye using 690 nm (d) B-mode Photoacoustic/ultrasound image of rabbit eye using 850 nm. (e) B-mode eye using 690 nm (d) B-mode Photoacoustic/ultrasound image of rabbit eye using 850 nm. (e) B-mode photoacoustic/ultrasound image when both 690 and 850 nm are used at the same time. Retinal vessels photoacoustic/ultrasound image when both 690 and 850 nm are used at the same time. Retinal vessels are imaged in a depth of 2 cm. (f) Photoacoustic image of skin and vasculature. Skin and blood vessel are imaged in a depth of 2 cm. (f) Photoacoustic image of skin and vasculature. Skin and blood vessel are shown using yellow arrows. Reproduced with permission from Ref. [68]. are shown using yellow arrows. Reproduced with permission from Ref. [68]. Furthermore, this portable PA imaging system has been exploited for accurate visualization of clinical metal needles [69] and human placental vasculature imaging [70]. The visualization of clinical metal needles is critical in guiding minimally invasive procedures, and PA imaging has shown

Sensors 2018, 18, 2264 9 of 24

Furthermore, this portable PA imaging system has been exploited for accurate visualization of clinical metal needles [69] and human placental vasculature imaging [70]. The visualization of clinical metal needles is critical in guiding minimally invasive procedures, and PA imaging has shownSensors promising 2018, 18, x potential FOR PEER REVIEW in this field. Wenfeng Xia et al. have developed a combined handheld9 of 24 PA and ultrasound imaging system for real-time visualization of clinical metal needles. In this system, promising potential in this field. Wenfeng Xia et al. have developed a combined handheld PA and the needle is inserted into chicken breast tissue ex vivo and is visible with PA imaging at depths of up ultrasound imaging system for real-time visualization of clinical metal needles. In this system, the to 2 cm.needle The is SNR inserted of both into USchicken images breast and tissue PA imagesex vivo and decreases is visible as with the needlePA imaging insertion at depths angle of increased.up to Figure2 10cm.a–c The briefly SNR of show both the US imagingimages and results PA images of this de systemcreases and as the a supplementary needle insertionvideo angle isincreased. provided in ReferenceFigure [69 10a–c]. From briefly Figure show 10 thea–c, imaging it can results be seen of thatthis system the visibility and a supplementary of the needle video and vesselis provided was low in ultrasoundin Reference images [69]. From but it Figure is high 10a–c, in PA it can images be seen and that dual the modality visibility of images. the needle The and human vessel placental was vasculaturelow in ultrasound imaging based images on but such it is a high handheld in PA images probe hasand thedual potential modality toimages. be the The guidance human placental in minimally invasivevasculature fetal interventions, imaging based such on assuch twin-two-twin a handheld pr transfusionobe has the syndromepotential to (TTTS), be the toguidance optimize in the treatmentminimally outcomes. invasive The fetal preliminary interventions, study such in Reference as twin-two-twin [70] demonstrated transfusion the syndrome feasibility (TTTS), of real-time to PA/USoptimize imaging the of treatment the human outcomes. placenta. The Figure preliminar 10d is they study photograph in Reference of human [70] placentademonstrated whose the facial feasibility of real-time PA/US imaging of the human placenta. Figure 10d is the photograph of human vasculatures are visible in PA images (Figure 10e,f), which means the anastomosing vessels may be placenta whose facial vasculatures are visible in PA images (Figure 10e,f), which means the identifiedanastomosing in TTTS. vessels However, may thebe identified clinical application in TTTS. Ho ofwever, guiding the clinical TTTS basedapplication on PA of imagingguiding TTTS requires us to developbased on aPA miniature imaging requires probe that us to is develop small enough a miniature to be probe delivered that is through small enough the working to be delivered channel of a fetoscope.through These the working studies channel and such of a fetoscope. handheld These imaging studies systems and such make handheld a big imaging step towards systems the make clinical applicationsa big step based towards on LEDs.the clinical applications based on LEDs.

FigureFigure 10. ( a10.–c )(a Photoacoustic–c) Photoacoustic and and ultrasound ultrasound (US)(US) imagingimaging of of a aneedle needle inserted inserted towards towards a vessel a vessel mimicking phantom. (a) US image; (b) PA image at 850nm; (c) PA + US image overlay. (d–f) Human mimicking phantom. (a) US image; (b) PA image at 850nm; (c) PA + US image overlay. (d–f) Human placental vasculature imaging. Photograph (d) and PA images (e,f) of a portion of the human placental vasculature imaging. Photograph (d) and PA images (e,f) of a portion of the human placenta. placenta. Reproduced with permission from Refs. [69,70]. Reproduced with permission from Refs. [69,70]. 2.3. Long-Pulse Modulation and Quasi-CW Modulation 2.3. Long-Pulse Modulation and Quasi-CW Modulation Long-pulse modulation and quasi-CW modulation can induce a photoacoustic nonlinear effect. Long-pulseThe former is modulation a novel method and to quasi-CW induce dual modulation PA signals canby using induce the a long photoacoustic pulse without nonlinear satisfying effect. The formerthe stress is aconfinement novel method [71]. toGenerally, induce duala PA PAsignal signals consists by of using a positive the long waveform pulse withoutand a negative satisfying the stresswaveform confinement with the [satisfaction71]. Generally, of stress a PA confin signalement consists and thermal of a positive confinement waveform as shown and in aFigure negative waveform11a. Due with to the the satisfaction pulse width of longer stress confinementthan the stress and of the thermal confinement confinement time, asa PA shown signal in will Figure be 11a. separated into two parts, the positive one and the negative one as shown in Figure 11b. The positive Due to the pulse width longer than the stress of the confinement time, a PA signal will be separated signal and negative signal are induced by the sharp rising and falling edges of the square laser pulse, respectively. Meanwhile, the absolute amplitude of the second signal is larger than that of the first

Sensors 2018, 18, 2264 10 of 24 into two parts, the positive one and the negative one as shown in Figure 11b. The positive signal and negativeSensors 2018 signal, 18, x FOR are inducedPEER REVIEW by the sharp rising and falling edges of the square laser pulse, respectively.10 of 24 Meanwhile, the absolute amplitude of the second signal is larger than that of the ﬁrst signal since thesignal heat since accumulation the heat accumulation is greater than is greater heat diffusion. than heat Withdiffusion. the increase With the of increase pulse width, of pulse if width, the heat if accumulationthe heat accumulation is smaller is thansmaller the than heat the diffusion, heat diffusion, the second the signalsecond will signal drop will back drop to theback normal to the levelnormal shown level inshown Figure in 11 Figurec. On the11c. contrary,On the contrary if the heat, if accumulationthe heat accumulation equals to equals the heat to diffusion, the heat thediffusion, second the signal second will signal be saturated. will be saturated. The theoretical The theoretical analysis is analysis derived is in derived Reference in [Reference71]. This long[71]. laserThis long pulse-induced laser pulse-induced dual PA (LDPA) dual PA nonlinear (LDPA) effectnonlinear can be effect observed can be with observed low-power with low-power laser diode laser and hasdiode much and broaderhas much application broader application compared withcompared those with that onlythose happen that only when happen laser when ﬂuence laser exceeds fluence a thresholdexceeds a valuethreshold [72,73 value]. [72,73].

Figure 11. The fluence fluence pattern, temperature change and PA signal waveform of (a)) conventional shortshort laserlaser pulse-inducedpulse-induced PAPA effecteffect withwith bothboth stressstress andand thermalthermal confinementsconfinements satisfied;satisfied; ((bb)) longlong laserlaser pulsepulse inducedinduced dualdual PAPA nonlinearnonlinear effecteffect withwith onlyonly thermalthermal confinementconfinement satisfied,satisfied, andand ((cc)) eveneven longerlonger laserlaser pulsepulse withoutwithout thermalthermal confinement.confinement. ReproducedReproduced withwith permissionpermission fromfrom Ref.Ref. [[71].71].

This nonlinearnonlinear effect,effect, based based on on the the relationship relationship between between heat heat accumulation accumulation and and diffusion, diffusion, can alsocan bealso observed be observed at quasi-CW at quasi-CW pulses pulses excitation excitation when when the the pulse-pulse pulse-pulse interval interval is is smaller smaller thanthan thermalthermal relaxationrelaxation time.time. Figure 1212aa shows the increasing amplitudeamplitude of PA signalssignals inducedinduced byby the consecutive pulses.pulses. Comparing Comparing with with Figure Figure 12b,c, 12b,c, the the last last nonlin nonlinearear PA PAsignal signal is much is much greater greater than the than first the linear ﬁrst linearPA signal. PA signal. This Thiskind kindof characteristic of characteristic can canimprov improvee the the imaging imaging contrast contrast by by differential differential imaging. Figure 1212dd isis aa PAMPAMin in vivo vivoimaging imaging setup setup based based on on the the quasi-CW quasi-CW excitation. excitation. The The abdomen abdomen region region of aof 7-week-old a 7-week-old rat wasrat was injected injected with with a kind a ofkind NIR of organic NIR organic dye—IR-820 dye—IR-820 which haswhich quite has a lowquite thermal a low conductivitythermal conductivity and a good and thermal a good conﬁnement thermal confinem property.ent Theproperty. imaging The results imaging are results shown inare Figure shown 12 ine. DueFigure to the12e. existence Due to the of theexistence blood spotof the at blood the surface spot at of the abdomensurface of skin, the abdomen the linear skin, PA imaging the linear shows PA theimaging strong shows blood the signals strong and blood weak signals IR-820 and signals. weak InIR-820 nonlinear signals. PA In imaging, nonlinear the PA IR-820 imaging, signals the areIR- stronger820 signals due are to thestronger heat accumulation.due to the heat Furthermore, accumulation. the differentialFurthermore, imaging the differential can easily imaging identify can the positioneasily identify of IR-820. the positi Theon image of IR-820. contrast The can image also contrast be improved can al fromso be 0.49improved (linear from PAimaging) 0.49 (linear to PA 1.7 (nonlinearimaging) to PA 1.7 imaging) (nonlinear and PA 12.3 imagin (differentialg) and 12.3 imaging). (differential imaging).

Sensors 2018, 18, 2264 11 of 24 Sensors 2018, 18, x FOR PEER REVIEW 11 of 24

FigureFigure 12. 12.(a ()a A) A typical typical 500 500 PAPA signal,signal, ((b) the first first linear linear PA PA signal, signal, and and (c ()c the) the last last nonlinear nonlinear PA PA signal. signal. (d) In vivo experimental setup based on the LDPA technique by quasi-CW laser excitation. (e) The (d) In vivo experimental setup based on the LDPA technique by quasi-CW laser excitation. (e) The reconstructed linear PA, nonlinear PA and differential imaging results of a rat with subcutaneous reconstructed linear PA, nonlinear PA and differential imaging results of a rat with subcutaneous injection of IR820 in the abdomen. Strong PA signals from high-absorptive blood background are injection of IR820 in the abdomen. Strong PA signals from high-absorptive blood background are shown in the linear and nonlinear PA images. By subtracting, the signal from the IR820 pops up with shown in the linear and nonlinear PA images. By subtracting, the signal from the IR820 pops up with linear background suppression. Reproduced with permission from Ref. [71]. linear background suppression. Reproduced with permission from Ref. [71]. 2.4. Coded Excitation 2.4. Coded Excitation Although the LDs or LEDs can serve as a compact and economical light source for the PA system, theAlthough large number the LDs of averaging or LEDs can of the serve received as a compact signals andis required economical to improve light source the SNR, for the which PA system,will thelimit large the number imaging of averagingspeed. Coded of the excitation received is signals an alternative is required technique to improve to induce the SNR, PA which signals will and limit theimprove imaging SNR speed. instead Coded of excitationdata averaging. is an alternativeMoreover, techniqueunder the toconventional induce PA signalsaveraging and method, improve the SNR insteadrepetition of data rate averaging. of lasers will Moreover, be limited under by the the time conventional of flight of th averaginge acoustic method,signal. The the coded repetition excitation rate of laserscan willbe used be limited to overcome by the time this oflimitation. flight of theIn acousticthis method, signal. the The consecutive coded excitation coded cansequence be used is to overcometransmitted this from limitation. the light In source this method, and the the correspondi consecutiveng induced coded sequence PA signals is are transmitted properly filtered from the and light sourcedecoded and to the generate corresponding the desired induced PA signals PA signals [74]. The are desired properly PA filtered signals andexhibit decoded a higher to generatemainlobe- the desiredto-sidelobe PA signals ratio leading [74]. The to desiredthe improvement PA signals of exhibit the SNR. a higher mainlobe-to-sidelobe ratio leading to the improvementGolay codes of is the one SNR. of the coded excitation schemes. There are two main advantages in Golay codes [75]. One, is the improvement of SNR of PA signals as a function of √ where N denotes the Golay codes is one of the coded excitation schemes. There are two main√ advantages in Golay codescode [ 75length.]. One, Another is the improvementone is the simultaneous of SNR of acquir PA signalsement asof the a function PA signals of atN multi-wavelengthwhere N denotes by the exploiting the different sets of orthogonal codes. Golay codes consist of two biphase orthogonal codes code length. Another one is the simultaneous acquirement of the PA signals at multi-wavelength but the light source is unable to emit the negative signals, so that every biphase codes will be split up by exploiting the different sets of orthogonal codes. Golay codes consist of two biphase orthogonal into two unipolar parts that contain positive parts and negative parts, respectively. Thus, the single- codes but the light source is unable to emit the negative signals, so that every biphase codes will be wavelength Golay codes excitation consist of four excitation sequences. The elimination of sidelobes split up into two unipolar parts that contain positive parts and negative parts, respectively. Thus, of PA signals benefits from the complementary code pairs. In Reference [76], the multispectral thephotoacoustic single-wavelength coded Golay excitation codes (MS-PACE) excitation consist based ofon four unipolar excitation orthogonal sequences. Golay The eliminationcodes was of sidelobesdiscussed. of PA It contains signals benefits eight unip fromolar the coded complementary sequences and code simultaneous pairs. In Referencely acquires [76 ],the the PA multispectral data sets photoacousticfrom two irradiation coded excitation wavelengths (MS-PACE) and then based separa ontes unipolar them. In orthogonal Figure 13a, Golay a 652 codes nm custom-made was discussed. It containshigh power eight quasi-continuous unipolar coded laser sequences diode bar and (diode simultaneously 1) with a maximum acquires optical the PA power data of sets 100 fromW and two irradiationan 808 nm wavelengths laser diode bar and with then 225 separates W were them.used in In this Figure experimental 13a, a 652 setup nm custom-madefor MS-PACE. highThe light power quasi-continuousfrom diode 1 was laser mostly diode absorbed bar (diode by the 1) withgreena dye maximum so that the optical PA response power of from 100 Wthe andblack an dye 808 is nm lasermainly diode related bar with to the 225 light W werefrom diode used in2. thisFor these experimental experiments, setup the for code MS-PACE. length is The 512 lightbit and from the diodelight 1 wasrepetition mostly absorbedrate is 500 bykHz. the The green total dye time so to that send the the PA codes response is 4.2 ms, from which the black can obtain dye is 63 mainly waveforms related toper the wavelength light from diodefor the 2.conventional For these experiments, averaging method. the code Consequently, length is 512 the bitimaging and the results light by repetition coded rateexcitation is 500 kHz. (Figure The 13c,e) total are time compared to send with the that codes by is63 4.2 averaging ms, which times can (Figure obtain 13b,d). 63 waveforms The results per wavelengthshow that the for noise the conventional floor of the averaging averaging method method. is approximately Consequently, 9 dB the higher imaging than resultsthat of the by MS- coded excitationPACE method (Figure 13forc,e) each are comparedwavelength, with which that bycorresponds 63 averaging to timesthe calculated (Figure 13 b,d).coding The gain results (SNR show thatimprovement) the noise floor of ofroughly the averaging 9 dB. Furthermore, method is approximatelythere is no crosstalk 9 dB and higher correlation than that sidelobes of the MS-PACE in the

Sensors 2018, 18, 2264 12 of 24 method for each wavelength, which corresponds to the calculated coding gain (SNR improvement) Sensors 2018, 18, x FOR PEER REVIEW 12 of 24 of roughly 9 dB. Furthermore, there is no crosstalk and correlation sidelobes in the images based on MS-PACE.images based In addition on MS-PACE. to Golay code,In addition another to coding Golay strategy code, foranother multiple coding wavelength, strategy pseudorandomfor multiple codes,wavelength, is discussed pseudorandom in References codes, [77– is79 discussed], whose coding in References gain exceeds [77–79], that whose of orthogonal coding gain Golay exceeds codes forthat ﬁnite of codeorthogonal lengths. Golay codes for finite code lengths.

Figure 13. a b e Figure 13.( )(a Experimental) Experimental setup setup for for MS-PACE.MS-PACE. ((b–e) Comparison of of MS-PACE MS-PACE and and time time equivalent equivalent averagingaveraging for for both both excitation excitation wavelengths. wavelengths. ( bb)) PA PA image image for for averaging averaging as as many many acquisitions acquisitions as as possiblepossible during during the the coding coding procedure procedure for for W =W 1, = termed 1, termed time-equivalent time-equivalent averaging. averaging. (c) PA(c) imagePA image using MS-PACEusing MS-PACE for W = for 1. (Wd) = PA 1. image(d) PA usingimage time-equivalentusing time-equivalent averaging averaging for W for = 2.W (=e )2. PA (e) imagePA image using MS-PACEusing MS-PACE for W = 2.for Reproduced W = 2. Reproduced with permission with permission from Ref.from [76 Ref.]. [76].

TheThe Legendre Legendre sequences sequences have have simpler simpler codingcoding procedure than than unipolar unipolar Golay Golay codes codes since since it only it only consistsconsists of of two two unipolar unipolar sequencessequences [80]. [80]. It Itcan can slightly slightly exceed exceed the gain the gainof Golay of Golay codes for codes finite for code ﬁnite codesending sending time, time, and and the the range range sidelobes sidelobes remain remain invisible invisible if the if the code code length length is issufficiently sufﬁciently large. large. However,However, it is it not is suitablenot suitable for multi-wavelength for multi-wavelength imaging imaging since thesince multiple the multiple signals signals cannot becannot separated. be separated. Moreover, the further simple coding procedure is achieved in periodically perfect Moreover, the further simple coding procedure is achieved in periodically perfect sequences (PPS), sequences (PPS), as only a single sequence needs to be repeated continuously [81]. PPS is suitable for as only a single sequence needs to be repeated continuously [81]. PPS is suitable for continuous continuous acquisition without the artifact and the performance is better than the above coding acquisition without the artifact and the performance is better than the above coding strategies. strategies. Some modulation techniques can also be used in coded excitation such as frequency Some modulation techniques can also be used in coded excitation such as frequency modulation modulation[82] and pulse [82] andposition pulse modulation position modulation (PPM) [83]. (PPM)The frequency [83]. The modula frequencytion is modulation to tune the interval is to tune thebetween interval two between adjacent two laser adjacent pulses laser by changing pulses by the changing period of the pules. period The of PPM pules. is to The tune PPM the is pulses to tune theposition pulses positionat the same at theperiod. same Both period. two Bothtechniques two techniques will induce will the inducesidelobes the because sidelobes they because consist theyof consistonly ofa single only acode single sequence. code sequence. However, However, the coding the gain coding of both gain methods of both performs methods well performs and exceeds well and that of previous code strategies. PPM codes are suitable for short code applications since the performance will drop for long codes.

Sensors 2018, 18, 2264 13 of 24 exceeds that of previous code strategies. PPM codes are suitable for short code applications since the performance will drop for long codes. Sensors 2018, 18, x FOR PEER REVIEW 13 of 24 3. Continuous Modulation 3. ContinuousCompared withModulation the pulsed excitation that is used in time-domain photoacoustic imaging, the continuousCompared modulation with the is mostlypulsed usedexcitation in the that frequency-domain is used in time-domain (FD) photoacoustic photoacoustic (FDPA) imaging, imaging, the whichcontinuous employs modulation light intensity is mostly modulation used in the to frequency-domain generate the PA signals(FD) photoacoustic [84]. It could (FDPA) be considered imaging, as anotherwhich kindemploys of code light excitation intensity methodmodulation [85]. to It generate is more economicalthe PA signals than [84]. pulsed It could modulation be considered because as it usesanother the low-cost kind of code continuous excitation laser me diodesthod [85]. with It is a more specific economical modulation than scheme. pulsed modulation In addition, because the SNR willit uses be remedied the low-cost by employingcontinuous thelaser cross-correlation diodes with a specific operation modulation [86]. scheme. In addition, the SNR will be remedied by employing the cross-correlation operation [86]. 3.1. Chirp Modulation 3.1. Chirp Modulation The chirp modulation is implemented by changing the excitation frequency linearly, which is a typicalThe modulation chirp modulation method inis implemented frequency domain by changing [87–89 the]. The excitation correlation frequency processing linearly, (matched which is filter a compression)typical modulation and heterodyne method in mixing frequency with domain coherent [87–89]. detection The are correlation the two mainprocessing processing (matched approaches filter forcompression) FD signals [84and]. Theheterodyne FD imaging mixing has with been cohe implementedrent detection in x-yare scan the mode,two main whereas processing Stephan Kellnbergerapproaches et for al. FD proposed signals [84]. a novelThe FD scan imaging mode has that been requires implemented multiple in x-y projections scan mode, (angles) whereas and mathematicalStephan Kellnberger inversion et [al.90 ].proposed In Figure a novel 14e, thescan optical mode that fiber requires (OF) and multiple transducer projections (T) are (angles) rotating and mathematical inversion [90]. In Figure 14e, the optical fiber (OF) and transducer (T) are rotating simultaneously by the step sizes of 2◦, resulting in 180 projections and ~10 min for a whole image. simultaneously by the step sizes of 2°, resulting in 180 projections and ~10 min for a whole image. A A 808 nm CW laser diode with a peak power of 500 mW and a focused transducer with 3.5 MHz 808 nm CW laser diode with a peak power of 500 mW and a focused transducer with 3.5 MHz central central frequency and 76% bandwidth were used. The frequency sweeps from 1 to 5 MHz with a frequency and 76% bandwidth were used. The frequency sweeps from 1 to 5 MHz with a duration of duration of 1 ms. To validate the ability of tissue imaging in vivo by this FD optoacoustic tomography 1 ms. To validate the ability of tissue imaging in vivo by this FD optoacoustic tomography setup, the setup,mouse the tail mouse was imaged tail was at imageda height atof a~4 height cm from of ~4the cmdistal from end the with distal the injection end with of the Indocyanine injection of Indocyaninegreen (ICG). green The (ICG).lateral Thecaudal lateral veins, caudal dorsal veins, vein, dorsaland the vein, ventral and caudal the ventral artery caudal are visible artery in are Figure visible in14a. Figure The 14 taila. Thein Figure tail in 14b Figure was injected14b was with injected ICG withand after ICG ~10 and min after its ~10 imaging min itsresult imaging was record result in was recordFigure in 14c Figure to investigate 14c to investigate the dynamic the dynamic change of change ICG in of the ICG tail. in The the comparison tail. The comparison of these two of thesefigures two figuresshows shows that the that light the absorption light absorption intensity intensity reduces reduces in Figure in Figure14c as the14c ICG as the fading. ICG fading.The proposed The proposed scan scanmode mode can can realize realize diffraction-limited diffraction-limited resolution resolution through through retrieving retrieving much much higher higher effective effective apertures. apertures. ThisThis potentially potentially leads leads to to superior superior imaging imaging performanceperformance and SNR. SNR.

FigureFigure 14. 14. ((aa)) FD-optoacoustic FD-optoacoustic tomography tomography (FD-OAT) (FD-OAT) in vivoin mouse vivo mouse tail image tail without image ICG, without (b) FD- ICG, (b)OAT FD-OAT tail image tail during image ICG during injection, ICG ( injection,c) tail image (c )~10 tail min image after ~10initial min ICG after injection, initial (d) ICG cryoslice injection, of (d)the cryoslice mouse of(LV, the lateral mouse caudal (LV, lateralveins; DV, caudal dorsal veins; caud DV,al vein; dorsal VA, caudal ventral vein; caudal VA, artery; ventral dashed caudal circle artery; represents approximate tail surface), (e) schematic diagram of the frequency domain optoacoustic dashed circle represents approximate tail surface), (e) schematic diagram of the frequency domain tomography system. An optical fiber (OF) guides the laser chirps onto the object (OB). Signals from optoacoustic tomography system. An optical fiber (OF) guides the laser chirps onto the object (OB). the photodetector (PD) and the transducer (T) are acquired simultaneously by the data acquisition Signals from the photodetector (PD) and the transducer (T) are acquired simultaneously by the data (DAQ) and afterwards are cross-correlated. Reproduced with permission from Ref. [90]. acquisition (DAQ) and afterwards are cross-correlated. Reproduced with permission from Ref. [90]. In PA imaging, the measurement of hemoglobin oxygen saturation is one of the significant applicationsIn PA imaging, [91]. The the laser measurement fluence is ass ofumed hemoglobin to be known oxygen in the saturationtissue when is uses one two of wavelengths the significant applicationsto measure[ the91]. hemoglobin The laser fluence oxygen is saturation. assumed toIn bereality, known the influence the tissue is unknown when uses because two wavelengthsthe optical scatter is not wavelength-independent, so this assumption will cause some error [92]. Bahman Lashkari et al. proposed a method which employs the phase of the FD photoacoustic signals to measure the absorption coefficient and quantitatively probe the hemoglobin oxygen saturation

Sensors 2018, 18, 2264 14 of 24 to measure the hemoglobin oxygen saturation. In reality, the fluence is unknown because the optical scatter is not wavelength-independent, so this assumption will cause some error [92]. Bahman Lashkari et al. proposed a method which employs the phase of the FD photoacoustic signals Sensors 2018, 18, x FOR PEER REVIEW 14 of 24 to measure the absorption coefficient and quantitatively probe the hemoglobin oxygen saturation withoutwithout thethe influenceinfluence of fluencefluence [[92]92].. However, thethe prerequisiteprerequisite forfor usingusing thisthis methodmethod isis toto knowknow the chromophore geometry. In Fourier domain, the generated PA signals will be affected by the fluence, fluence, and the functionfunction of thethe correspondingcorresponding cross-correlationcross-correlation signals contains the fluencefluence parts, while the phase of cross-correlation signalssignals doesdoes not.not. That yields the absorption coefficient coefficient of the chromophore inin anan analyticalanalytical form,form, whichwhich cancan bebe usedused toto attainattain thethe absorptionabsorption coefficientcoefficient quantitatively.quantitatively. ItIt alsoalso shows thatthat onlyonly oneone wavelengthwavelength is is required. required. Through Through this this method, method, the the error error of of phase phase measurement measurement of oxygenof oxygen saturation saturation was was 5.3% 5.3% while while the th amplitudee amplitude measurement measurement error error was was 12.2%. 12.2%. To provideprovide sufficient sufficient diagnostic diagnostic information, information, combining combining different different imaging imaging modalities modalities is necessary, is whichnecessary, can which attain can the attain complementary the complementary information. information. Pervious Pervious investigations investigations have have combined combined FD photoacousticFD photoacoustic imaging imaging and and ultrasound ultrasound imaging imaging for for intravascular intravascular applications applications [ 93[93].]. ThisThis paper employed 1 W 1210 nm continuous LD with the intensity-modulated chirp frequency of 4–12 MHz. The agaragar vesselvessel phantomphantom with graphitegraphite and lipid targets waswas employedemployed to validatevalidate the feasibilityfeasibility of thisthis dualdual modalitymodality imagingimaging system.system. Moreover, suchsuch dual-modalitydual-modality imagingimaging systems make sense forfor tumortumor detection [94]. [94]. For For this this literature, literature, the the linea linearr frequency-modulated frequency-modulated chirp chirp signals signals (0.5 (0.5 MHz MHz to to 4 4MHz, MHz, 1 1ms ms duration) duration) was was produced produced by by an an 808 808 nm nm LD LD with with output output power power of of 6 6 W, W, and and the PA signals detection asas wellwell as as the the ultrasound ultrasound imaging imaging were were implemented implemented by aby 64 a element, 64 element, 3.3 MHz 3.3 MHz phased phased array transducer.array transducer. The imaging The imaging object was object a mouse was that a mouse was injected that withwas ainjected cultured with human a cultured hypopharyngeal human headhypopharyngeal and neck squamous head and cell neck carcinoma squamous FaDu cell cells carcinoma in the right FaDu thigh cells to causein the theright tumor thigh to to grow cause in the mouse.tumor to The grow results in the show mouse. thatthe The ultrasound results show imaging that the cannot ultrasound distinguish imaging the tumor cannot as thedistinguish interference the oftumor the surroundingas the interference soft issue of the such surrounding as the dash soft line issu ine such Figure as 15thea. dash PA imaging line in Figure is sensitive 15a. PA for imaging tumor detectionis sensitive due for totumor the increaseddetection blooddue to flowthe increased in tumor, blood which flow can in be tumor, seen inwhich Figure can 15 beb. seen Figure in Figure 15c is the15b. co-registered Figure 15c is image the co-registered of Figure 15a,b, image which of providesFigure 15a,b, complete which diagnostic provides information complete diagnostic for tumor detection.information It for indicates tumor thedetection. relative It position indicates of the the relative tumor comparedposition of with the tumor the other compared soft issue with in the mouseother soft thigh. issue in the mouse thigh.

Figure 15. (a) Pure ultrasoundultrasound imageimage showing the bottom edgeedge of the plastic seat (arrow) and region of interest (oval); (b) improved post-ampliﬁcationpost-amplification PA PA image of tumor in right thigh of nude mouse. ((cc)) ImprovedImproved ﬁlteredfiltered PAPA imageimage superimposedsuperimposed onon thethe purepure ultrasoundultrasound imageimage ofof thethe rightright thighthigh ofof thethe mouse. Reproduced with permissionpermission from Ref.Ref. [[94].94].

Sensors 2018, 18, 2264 15 of 24

3.2. Sinusoidal Modulation Driving the LDs or LEDs by sinusoidal modulation with single frequency to implement PA imaging is a special case of FDPA. In the case of sinusoidal excitation, SNR of the PA signals will be smaller than that of the pulsed mode, which has been discussed in Reference [95]. There are two main factors also presented in Reference [95] to remedy this difference. Firstly, because the noise amplitudeSensors 2018, is18, proportional x FOR PEER REVIEW to the square of the signal bandwidth, the narrowband signals generated15 of by 24 sinusoidal modulation could increase the SNR to some extent. Secondly, the transducer used in pulsed modulationpulsed modulation mode is mode required is required to have a to broad have bandwidth, a broad bandwidth, which will which cause insertionwill cause loss insertion because loss of thebecause low electromechanical of the low electromechanical coupling coefﬁcient coupling of piezoelectric coefficient materialof piezoelectric or the high material acoustic or impedance. the high However,acoustic impedance. a resonant andHowever, low-loss a transducer resonant canand be low-loss used in sinusoidaltransducer modulation can be used mode, in whichsinusoidal can exhibitmodulation much mode, higher which detection can exhibit sensitivity much than higher wideband detection transducers. sensitivity than wideband transducers. TheThe characteristiccharacteristic of of phase phase delay delay of PA of signalsPA sign inducedals induced by sinusoidal by sinusoid modulational modulation can be exploited can be toexploited achieve to photoacoustic achieve photoacoustic viscoelasticity viscoelasticity imaging (PAVEI) imaging since (PAVEI) different since viscoelasticitydifferent viscoelasticity will lead towill a differentlead to a phasedifferent of thephase PA of wave the [PA96]. wave The thermal[96]. The stress thermal will stress change will periodically change periodically when there when is cyclical there heatingis cyclical in theheating local region.in the The local corresponding region. The generated corresponding strain willgenerated also change strain periodically will also butchange will haveperiodically a delay but with will the have thermal a delay stress with due tothe the thermal viscoelasticity stress due of biologicalto the viscoelasticity tissues. By of establishing biological thetissues. relationship By establishing of the PA the phase relationship delay and of the the PA viscoelasticity phase delay ofand soft the issue, viscoelasticity the PAVEI of of soft tissue issue, could the bePAVEI implemented. of tissue could An 808be implemente nm CW laserd. An was 808 used nm CW with laser the was modulation used with frequency the modulation of 50 kHz frequency and a lock-inof 50 kHz ampliﬁer and a waslock-in used amplifier to calculate was the used phase to calculate delay between the phase the dominantdelay between frequency the dominant PA wave andfrequency the reference PA wave signal. and Figurethe reference 16b shows signal. the Figure PAVEI 16b results shows in Reference the PAVEI [96 results], which in showsReference that [96], the viscoelasticitywhich shows that distribution the viscoelasticity of the sample distribution can be mapped. of the sample This technique can be mapped. could possibly This technique be applied could to thepossibly detection be applied of atherosclerotic to the detection plaque of and atherosclerotic skin tumor since plaque the changeand skin in viscoelasticitytumor since the could change herald in theviscoelasticity diseases. could herald the diseases.

FigureFigure 16.16. ((aa)) PhotographPhotograph ofof biologicalbiological tissuestissues (fat(fat andand liver).liver). ((bb)) ViscoelasticityViscoelasticity distributiondistribution ofof thethe samplesample detecteddetected byby PAVEI.PAVEI. Reproduced Reproduced with with permission permission from from Ref. Ref. [ 96[96].].

As mentioned above, the FDPA imaging has combined with ultrasound imaging to provide the As mentioned above, the FDPA imaging has combined with ultrasound imaging to provide the complementary information. For Sinusoidal modulation, it has combined with fluorescence complementary information. For Sinusoidal modulation, it has combined with fluorescence microscopy microscopy to map the fluorescence position in blood smear [97]. In this imaging system, a to map the fluorescence position in blood smear [97]. In this imaging system, a continuous LD of continuous LD of 405 nm wavelength with an output power of 120 mW was modulated with the 405 nm wavelength with an output power of 120 mW was modulated with the frequency of 10 MHz frequency of 10 MHz and the FWHM of the laser spot was ~750 nm. The shape of blood cell on blood and the FWHM of the laser spot was ~750 nm. The shape of blood cell on blood smear is clear in PA smear is clear in PA imaging (Figure 17a) and some fluorescent spots can be seen in the luminescence imaging (Figure 17a) and some fluorescent spots can be seen in the luminescence image (Figure 17b). image (Figure 17b). The overlay of these two results (Figure 17c) reveal that the fluorescent spot is The overlay of these two results (Figure 17c) reveal that the fluorescent spot is from the blood cell. from the blood cell. In general, the normal blood cell does not express the characteristics of In general, the normal blood cell does not express the characteristics of fluorescence. The origin of fluorescence. The origin of the fluorescence is unknown. However, this system validates the the fluorescence is unknown. However, this system validates the important combination of these two important combination of these two imaging modalities as they can provide complementary imaging modalities as they can provide complementary information for each other. information for each other.

Sensors 2018, 18, 2264 16 of 24 Sensors 2018, 18, x FOR PEER REVIEW 16 of 24

Figure 17. ((aa)) OR-PAM OR-PAM image ( a) and simultaneously obtain obtaineded luminescence image ( b) of human red red blood cells. ( cc)) Overlay Overlay of of the the PA PA and luminescence image. image. ( (dd)) Bright-field Bright-ﬁeld image image of of the the same same region. region. Reproduced with permission from Ref. [[97].97].

Furthermore, the image contrast under sinusoidal modulation could be enhanced significantly signiﬁcantly due to the PA resonance effect [[98].98]. The expression of acoustic pressure is thethe equation of motion of a forced harmonic oscillator with damping. Thus Thus the the amplitude amplitude of of PA PA signals exists at a maximum value when the resonance phenomenon happens. Ba Basedsed on on this resonance effect, a novel imaging modality, namednamed photoacousticphotoacoustic resonance resonance imaging imaging (PARI), (PARI), is proposedis proposed in Reference in Reference [99] and[99] theand great the greatimprovement improvement of image of image contrast contrast was demonstratedwas demonstrated using using tissue-mimicking tissue-mimicking phantom phantom and exand vivo ex vivotissue. tissue. The exThe vivo ex tissuevivo tissue shown shown in Figure in Figure 18a consists 18a consists of the of porcine the porcine muscle muscle and liver and issues liver issues of which of whichresonance resonance spectra spectra are shown are inshown Figure in 18 Figureb. It can 18b. be It seen can thatbe seen the highestthat the amplitude highest amplitude of PA signals of PA of signalsporcine of muscle porcine and muscle liver are and at liver the frequency are at the of freq 910uency KHz andof 910 1300 KHz KHz, and respectively. 1300 KHz, respectively. The following The PA followingresonant imaging PA resonant experiments imaging regard experiments these two regard frequencies these two as frequencies the resonance as the frequency resonance of muscle frequency and ofliver muscle respectively. and liver However, respectively. it should However, be noticed it should that from be noticed Figure 18thatb, thefrom signal Figure from 18b, the the liver signal increases from theand liver that fromincreases themuscle and that decreases, from the which muscle means decreases, that the which resonance means frequency that the resonance could not befrequency directly coulddistinguished not be directly because distinguished there is no Lorentz-shape because there function is no Lorentz-shape in this image. function Due to a in lack this of image. enough Due broad to amodulation lack of enough frequency broad range modulation in this paper, frequency the resonance range in frequencythis paper, may the be resonance distinguished frequency if broadening may be distinguishedthe modulation if broadening frequency range. the modulation Figure 18 c,dfrequency shows range. that the Figure amplitude 18c,d shows of PA signalsthat the at amplitude resonant ofmodulation PA signals frequency at resonant will modulation be higher than frequency that at off-resonantwill be higher modulation than that frequency.at off-resonant The imaging modulation area frequency.is the white The dotted imaging square area shown is the in Figurewhite 18dotteda. Figure square 18e isshown the conventional in Figure 18a. single-pulse-induced Figure 18e is the conventionalPA imaging. Thesingle-pulse-induced differential image PA (Figure imaging. 18h) isThe obtained differential by subtracting image (Figure the liver 18h) resonance is obtained image by (Figuresubtracting 18g) the from liver the resonance muscle resonance image (Figure image 18g) (Figure from 18 thef). muscle As shown resonance in Figure image 18i, (Figure the differential 18f). As shownimage showsin Figure a much 18i, higherthe differential image contrast image than show thes othera much images. higher image contrast than the other images.

Sensors 2018, 18, 2264 17 of 24 Sensors 2018, 18, x FOR PEER REVIEW 17 of 24

FigureFigure 18.18. ((aa)) PhotographPhotograph ofof porcineporcine tissue.tissue. ((bb)) TheThe PAPA resonanceresonance spectraspectra ofof porcineporcine musclemuscle (blue(blue dashed)dashed) andand liverliver (red(red solid).solid). ((cc)) PAPA waveforms waveforms ofof liverliver atat resonanceresonance (1300(1300 kHz,kHz, dasheddashed redred line)line) andand off-resonanceoff-resonance (910 (910 kHz, kHz, solid solid blue blue line). line). (d )(d PA) PA waveforms waveforms of muscle of muscle at resonance at resonance (910 kHz,(910 kHz, solid bluesolid line)blue andline) off-resonance and off-resonance (1300 kHz,(1300 dashed kHz, dashed red line). red ( eline).) Conventional (e) Conventional single-pulse-induced single-pulse-induced PA image PA (fimage) PARI (f at) PARI muscle at muscle resonance resonance (910 kHz). (910 (g kHz).) PARI (g at) PARI liver resonanceat liver resonance (1300 kHz). (1300 (h kHz).) Differential (h) Differential image. (iimage.) Contrast (i) Contrast comparison comparison of Figure of5e–h. Figure Reproduced 5e–h. Reproduced with permission with permission from Ref. from [99]. Ref. [99].

3.3.3.3. HybridHybrid ModulationModulation TheThe hybridhybrid modulation combines combines different different modulation modulation types types to to ex excitecite a single a single laser laser source. source. To Tothe the best best of ofour our knowledge, knowledge, there there is is an an article article to to utilize the hybridhybrid modulationmodulation forfor temperaturetemperature monitoringmonitoring andand regulationregulation inin photothermalphotothermal therapytherapy [[100].100]. ThisThis articlearticle discussesdiscusses aa laser-sharedlaser-shared methodmethod toto achieveachieve temperaturetemperature monitoring,monitoring, regulationregulation andand photothermalphotothermal therapytherapy simultaneously.simultaneously. FigureFigure 1919aa isis thethe modulationmodulation schemescheme ofof thethe laserlaser diode,diode, whichwhich showsshows thatthat thethe FDPAFDPA temperaturetemperature measurementmeasurement withwith chirpchirp modulationmodulation andand photothermalphotothermal heatingheating withwith constantconstant laserlaser intensityintensity areare alternatingalternating quicklyquickly inin thethe timetime domain.domain. To estimateestimate thethe temperaturetemperature increaseincrease accordingaccording toto thethe amplitudeamplitude ofof PAPA signals,signals, thethe PAPA signalssignals werewere correlatedcorrelated withwith thethe referencereference signalssignals andand thenthen averaged. AA proportional-integral-derivativeproportional-integral-derivative (PID)(PID) controllercontroller isis adoptedadopted inin thethe feedbackfeedback looploop toto implementimplement thethe self-temperatureself-temperature regulation.regulation. TheThe acousto-opticacousto-optic modulatormodulator (AOM)(AOM) isis usedused toto modulatemodulate thethe lightlight intointo chirpchirp pulsepulse forfor PAPA imaging.imaging. TheThe heatingheating dosedose ofof thethe laserlaser isis controlledcontrolled byby aa mirror driven by a motor. AA 405405 nmnm laserlaser diodediode withwith aa maximummaximum powerpower ofof 11 WW isis usedused andand thethe phantomphantom isis aa blackblack wirewire withwith aa radiusradius ofof 11 mm.mm. TheThe chirpchirp frequencyfrequency sweepssweeps fromfrom 0.50.5 toto 1.51.5 MHzMHz inin PAPA temperature-measurement temperature-measurement mode.mode. BeforeBefore measuring measuring the the temperature, temperature, this this temperature temperature measurement measurement system system is calibrated is calibrated by using by ausing K-type a K-type thermocouple. thermocouple. It can beIt can seen be from seen Figure from Figure19b that 19b the that temperature the temperature can be maintainedcan be maintained stably onstably the pre-seton the temperaturepre-set temperature except forexcept the openfor the loop op photothermalen loop photothermal therapy. therapy. Figure 19 Figurec shows 19c that shows the accuracythat the accuracy of temperature of temperature regulation regulation is 0.9 ◦C, andis 0.9 the °C, accuracy and the ofaccuracy temperature of temperature measurement measurement is 0.75 ◦C. Thisis 0.75 system °C. This can achievesystem ultrafastcan achieve (4 KHz) ultrafast temperature (4 KHz) measurement. temperature measurement. The accuracy improvementThe accuracy requiresimprovement the enhancement requires the ofenhancement signals’ SNR of thatsignals’ will SNR be achieved that will bybe achieved a longer chirpby a longer signal. chirp Therefore, signal. thereTherefore, is a tradeoff there is between a tradeoff the between measurement the measurement speed and accuracy. speed and accuracy.

Sensors 2018, 18, 2264 18 of 24 Sensors 2018, 18, x FOR PEER REVIEW 18 of 24

FigureFigure 19. 19. (a(a) )Modulation Modulation scheme scheme of of the the laser laser diode. diode. (b (b) )Self-temperature Self-temperature regula regulationtion with with closed-loop closed-loop pre-setpre-set values values at at 29 29 °C,◦C, 44 44 °C,◦C, and and 56 56 °C,◦C, respectively. respectively. (c (c) )Temperature-regulation Temperature-regulation accuracy accuracy analysis analysis results.results. Reproduced Reproduced with with pe permissionrmission from from Ref. Ref. [100]. [100].

4.4. Discussion Discussion and and Conclusions Conclusions TheThe LDs LDs and and LEDs LEDs prompt prompt the the low-cost low-cost and and comp compactact PA PA sensing sensing and and imaging imaging system. system. However, However, thethe weak weak SNR SNR of of the the PA PA signals signals needs needs to to be be paid paid ex extratra attention, attention, due due to to the the much much low low energy energy emitted emitted byby LDs LDs or LEDs. ToTo solvesolve this this challenge, challenge, firstly, firstly, the the most most direct direct method method is to is improve to improve the output the output power powerof these of semiconductorthese semiconductor diodes diodes to enhance to enhance the SNR. th Toe SNR. the best To ofthe our best best of knowledge,our best knowledge, the maximum the maximumoutput peak output power peak of 325 power W 905 of nm325 laser W 905 diode nm with laser 50 diode ns pulsed with duration 50 ns pulsed was demonstrated duration was to demonstratedexcite the PA signalsto excite with the sufficientPA signals SNR with (18 sufficie dB fornt imaging SNR (18 porcine dB for ovaryimaging ex porcine vivo) and ovary no averaging ex vivo) andwas no required averaging [34]. was Secondly, required the [34]. most Secondly, commanding the most and commanding simple method and to simple increase method SNR is to to increase average SNRhundreds is to average of PA signals hundreds at the of samePA signals irradiation at the point, same whichirradiation will limitpoint, the which imaging will orlimit detection the imaging speed. orThirdly, detection as mentionedspeed. Thirdly, above, as mentioned the coded above, excitation the coded could excitation enhance the could SNR enhance through the the SNR correlation through theoperation correlation between operation the PA between signals andthe referencePA signals signals, and reference and the enhancement signals, and isthe proportional enhancement to theis proportionalcoded length. to the Moreover, coded length. the continuous Moreover, modulation the continuous could modulation facilitate thecould more facilitate low-cost the imagingmore low- or costsensing imaging system or sensing since the system required since output the required power of output the light power source of the is much light source lower thanis much the pulsedlower thanmodulation. the pulsed Fourthly, modulation. the contrast Fourthly, agent the could contrast be used agent to generatecould be largerused to PA generate signal amplitude, larger PA suchsignal as amplitude,gold and silversuch as nanoparticles gold and silver and nanoparticles gold nanorods and [gold101– 109nanorods]. Moreover, [101–109]. the Moreover, double-walled the double- carbon wallednanotubes carbon could nanotubes be used could as nontoxic be used and as biodegradable nontoxic and biodegradable contrast agents contrast [110]. Fifthly, agents some [110]. algorithm Fifthly, somealso couldalgorithm be exploited also could to improve be exploited the SNR to such improv as thee the adaptive SNR such line enhancers as the ad algorithmaptive line [111 enhancers]. Finally, algorithmother methods [111]. suchFinally, as the other empirical methods mode such decomposition as the empirical method mode and decomposition wavelet-based method methods and are wavelet-basedwidely investigated methods by manyare widely research investigated groups [112 by, 113many]. research groups [112,113]. ToTo conclude, conclude, Table Table 1 is thethe summarysummary of the differentdifferent modulations with different light sources thatthat have have been been discussed discussed in in this this review. review. This This pape paperr has has discussed discussed the the recent recent development development of of low-cost low-cost photoacousticphotoacoustic sensing sensing and and imaging imaging systems systems based based on on LDs LDs or or LEDs LEDs according according to to different different modulation modulation schemes.schemes. Through Through the the development development of of these these systems, systems, it it can can be be considered considered that that the the clinical clinical diagnosis diagnosis andand application application of of photoacoustic photoacoustic imaging imaging requires requires that that compact compact and and affordable affordable light light sources sources replace replace bulkybulky and expensiveexpensive light light sources. sources. Fortunately, Fortunately, many many prototype prototype systems systems have validated have validated the feasibility the feasibilityfor specific for functions specific functions such as hemoglobinsuch as hemoglobin oxygen saturationoxygen saturation detection detection [92], in [92], vivo insmall vivoanimal small animalbrain imagingbrain imaging [30], relative [30], relative temperature temperature detection detec [tion100], [100], etc. Furthermore, etc. Furthermore, the commercializedthe commercialized and andportable portable PA imagingPA imaging systems systems have have been been developed, developed, by which by which the synovitis the synovitis detection detection [31] and [31] rabbit and rabbit eye imaging [68] were achieved. It heralds a promising scenario for clinical application of PA imaging and sensing.

Sensors 2018, 18, 2264 19 of 24 eye imaging [68] were achieved. It heralds a promising scenario for clinical application of PA imaging and sensing.

Table 1. The summary of different modulation modes based on different light sources. PLDs, pulsed laser diodes; CW-LDs, continuous laser diodes.

Light Source Modulation Mode Advantages Disadvantages Application Example Simple excitation Portable imaging Single pulse Low SNR technique device [31] PLDs High speed and high Complicated code Multi-wavelength Code SNR sequence imaging [76] Provides another Nonlinear PA imaging Quasi-CW Too much heating dose image contrast [71] Portable imaging Single pulse Lower cost Low ﬁber couple LEDs device [68] efﬁcient High speed and high Multi-wavelength Code SNR imaging [27] High speed and high Complicated data Dual modality Chirp SNR processing imaging [94] CW-LDs High speed and higher Viscoelasticity Sinusoidal No depth information SNR imaging [96] Extra equipment to Hybrid Real-time PA sensing PA sensing [100] modulate the light Simple excitation Human skin imaging Single pulse Low SNR technique [34]

Funding: This research was funded by Natural Science Foundation of Shanghai, grant number 18ZR1425000. Conﬂicts of Interest: The authors declare no conﬂicts of interest.

References

1. Gao, F.; Zhang, R.; Feng, X.; Liu, S.; Ding, R.; Kishor, R.; Qiu, L.; Zheng, Y. Phase-domain photoacoustic sensing. Appl. Phys. Lett. 2017, 110, 033701. [CrossRef] 2. Gao, F.; Feng, X.; Zheng, Y. Photoacoustic elastic oscillation and characterization. Opt. Express 2015, 23, 20617–20628. [CrossRef][PubMed] 3. Kishor, R.; Gao, F.; Sreejith, S.; Feng, X.; Seah, Y.P.; Wang, Z.; Stuparu, M.C.; Lim, T.T.; Chen, X.; Zheng, Y. Photoacoustic induced surface acoustic wave sensor for concurrent opto-mechanical microﬂuidic sensing of dyes and plasmonic nanoparticles. RSC Adv. 2016, 6, 50238–50244. [CrossRef] 4. Gao, F.; Feng, X.; Zheng, Y. Photoacoustic phasoscopy super-contrast imaging. Appl. Phys. Lett. 2014, 104, 213701. [CrossRef] 5. Feng, X.; Gao, F.; Zheng, Y. Photoacoustic-Based-Close-Loop Temperature Control for Nanoparticle Hyperthermia. IEEE Trans. Biomed. Eng. 2015, 62, 1728–1737. 6. Feng, X.; Gao, F.; Zheng, Y. Magnetically mediated thermoacoustic imaging. Proc. SPIE 2014, 8943.[CrossRef] 7. Gao, F.; Feng, X.; Zheng, Y. Micro-Doppler Photoacoustic Effect and Sensing by Ultrasound Radar. IEEE J. Sel. Top. Quantum Electr. 2016, 22, 1–6. [CrossRef] 8. Gao, F.; Feng, X.; Zheng, Y. Microwave-acoustic correlated imaging and circuit modelling of biological tissues. In Proceedings of the 2013 IEEE Mtt-S International Microwave Workshop Series on RF and Wireless Technologies for Biomedical and Healthcare Applications (IMWS-BIO), Singapore, 9–11 December 2013; pp. 356–358. 9. Li, G.; Gao, F.; Qiu, Y.; Feng, X.; Zheng, Y. Multiple stimulated emission ﬂuorescence photoacoustic sensing and spectroscopy. Appl. Phys. Lett. 2016, 109, 013701. [CrossRef] 10. Gao, F. Multi-Wave Electromagnetic-Acoustic Sensing and Imaging; Springer: Singapore, 2017. 11. Gao, F.; Feng, X.; Miao, X.; Zheng, Y. A Prototype for a Palm-sized Photoacoustic Sensing Unit. X-Acoust. Imaging Sens. 2015, 1, 73–77. [CrossRef] Sensors 2018, 18, 2264 20 of 24

12. Gao, F.; Zheng, Y.; Feng, X.; Ohl, C.D. Thermoacoustic resonance effect and circuit modelling of biological tissue. Appl. Phys. Lett. 2013, 102, 063702. [CrossRef] 13. Li, C.; Wang, L.V. Photoacoustic tomography and sensing in biomedicine. Phys. Med. Biol. 2009, 54, R59–R97. [CrossRef][PubMed] 14. Gao, F.; Kishor, R.; Feng, X.; Liu, S.; Ding, R.; Zhang, R.; Zheng, Y. An analytical study of photoacoustic and thermoacoustic generation efficiency towards contrast agent and film design optimization. Photoacoustics 2017, 7, 1–11. [CrossRef][PubMed] 15. Feng, X.; Gao, F.; Kishor, R.; Zheng, Y. Coexisting and mixing phenomena of thermoacoustic and magnetoacoustic waves in water. Sci. Rep. 2015, 5, 11489. [CrossRef][PubMed] 16. Gao, F.; Feng, X.; Zheng, Y. Coherent Photoacoustic-Ultrasound Correlation and Imaging. IEEE Trans. Biomed. Eng. 2014, 61, 2507–2512. [CrossRef][PubMed] 17. Gao, F.; Zheng, Q.; Zheng, Y. Electrical circuit modeling and analysis of microwave acoustic interaction with biological tissues. Med. Phys. 2014, 41.[CrossRef][PubMed] 18. Gao, F.; Ong, Y.H.; Li, G.; Feng, X.; Liu, Q.; Zheng, Y. Fast photoacoustic-guided depth-resolved Raman spectroscopy: A feasibility study. Opt. Lett. 2015, 40, 3568–3571. [CrossRef][PubMed] 19. Ding, R.; Liu, H.; Zhang, X.; Xiao, J.; Kishor, R.; Sun, H.; Zhu, B.; Chen, G.; Gao, F.; Feng, X.; et al. Flexible Piezoelectric Nanocomposite Generators Based on Formamidinium Lead Halide Perovskite Nanoparticles. Adv. Funct. Mater. 2016, 26, 7708–7716. [CrossRef] 20. Xu, M.H.; Wang, L.H.V. Photoacoustic imaging in biomedicine. Rev. Sci. Instrum. 2006, 7, 041101. [CrossRef] 21. Wang, L.V. Prospects of photoacoustic tomography. Med. Phys. 2008, 35, 5758–5767. [CrossRef][PubMed] 22. Hu, S.; Wang, L.V. Photoacoustic imaging and characterization of the microvasculature. J. Biomed. Opt. 2010, 15, 011101. [CrossRef][PubMed] 23. Wang, D.; Wang, Y.; Wang, W.; Luo, D.; Chitgupi, U.; Geng, J.; Zhou, Y.; Wang, L.; Lovell, J.F.; Xia, Y. Deep tissue photoacoustic computed tomography with a fast and compact laser system. Biomed. Opt. Expr. 2017, 8, 112–123. [CrossRef][PubMed] 24. Canal, C.; Laugustin, A.; Kohl, A.; Rabot, A. Portable multiwavelength laser diode source for handheld photoacoustic devices. Proc. SPIE 2016, 9887.[CrossRef] 25. Allen, T.J.; Beard, P.C. Pulsed near-infrared laser diode excitation system for biomedical photoacoustic imaging. Opt. Lett. 2006, 31, 3462–3464. [CrossRef][PubMed] 26. Kolkman, R.G.; Steenbergen, W.; van Leeuwen, T.G. In vivo photoacoustic imaging of blood vessels with a pulsed laser diode. Lasers Med. Sci. 2006, 21, 134–139. [CrossRef][PubMed] 27. Allen, T.J.; Beard, P.C. High power visible light emitting diodes as pulsed excitation sources for biomedical photoacoustics. Biomed. Opt. Expr. 2016, 7, 1260–1270. [CrossRef][PubMed] 28. Allen, T.J.; Cox, B.T.; Beard, P.C. Generating photoacoustic signals using high-peak power pulsed laser diodes. In Photons Plus Ultrasound Imaging Sensing 2005; International Society for Optics and Photonics: Belingham, WA, USA, 2005; Volume 5697, pp. 233–242. 29. Nissinen, J.; Kostamovaara, J. A 4 a peak current and 2 ns pulse width CMOS laser diode driver for high measurement rate applications. In Proceedings of the ESSCIRC (ESSCIRC), Bucharest, Romania, 16–20 September 2013. 30. Upputuri, P.K.; Pramanik, M. Dynamic in vivo imaging of small animal brain using pulsed laser diode-based photoacoustic tomography system. J. Biomed. Opt. 2017, 22, 1–4. [CrossRef][PubMed] 31. Van den Berg, P.J.; Daodi, K.; Moens, H.J.B.; Steenbergen, W. Feasibility of photoacoustic/ultrasound imaging of synovitis in finger joints using a point-of-care system. Photoacoustics 2017, 8, 8–14. [CrossRef][PubMed] 32. Daoudi, K.; van der Berg, P.J.; Rabot, O.; Kohl, A.; Tisserand, S.; Brands, P.; Steenbergen, W. Handheld probe for portable high frame photoacoustic/ultrasound imaging system. In Photons Plus Ultrasound: Imaging and Sensing 2013; International Society for Optics and Photonics: Belingham, WA, USA, 2013. 33. Van den Berg, P.J.; Daoudi, K.; Steenbergen, W. Pulsed photoacoustic flow imaging with a handheld system. J. Biomed. Opt. 2016, 21, 26004. [CrossRef][PubMed] 34. Erfanzadeh, M.; Kumavor, P.D.; Zhu, Q. Laser scanning laser diode photoacoustic microscopy system. Photoacoustics 2018, 9, 1–9. [CrossRef][PubMed] 35. Zeng, L.; Liu, G.; Yang, D.; Ji, X. 3D-visual laser-diode-based photoacoustic imaging. Opt. Express. 2012, 20, 1237–1246. [CrossRef][PubMed] Sensors 2018, 18, 2264 21 of 24

36. Zeng, L.; Liu, G.; Yang, D.; Ji, X. Cost-efﬁcient laser-diode-induced optical-resolution photoacoustic microscopy for two-dimensional/three-dimensional biomedical imaging. J. Biomed. Opt. 2014, 19, 076017. [CrossRef][PubMed] 37. Zeng, L.; Liu, G.; Yang, D.; Ji, X. Portable optical-resolution photoacoustic microscopy with a pulsed laser diode excitation. Appl. Phys. Lett. 2013, 102, 053704. [CrossRef] 38. Upputuri, P.K.; Sivasubramanian, K.; Pramanik, M. High speed photoacoustic tomography system with low cost portable pulsed diode laser. In Proceedings of the International Conference on Optical and Photonic Engineering (icOPEN 2015), Singapore, 14–16 April 2015. 39. Beckmann, M.F.; Schwab, H.M.; Schmitz, G. Optimized SNR simultaneous multispectral photoacoustic imaging with laser diodes. Opt. Expr. 2015, 23, 1816–1828. [CrossRef][PubMed] 40. Leggio, L.; De Varona, A.; Escudero, P.; del Barrio, G.C.; Osinski, M.; Rivera, H. A compact multi-wavelength optoacoustic system based on high-power diode lasers for characterization of double-walled carbon nanotubes (DWCNTs) for biomedical applications. Proc. SPIE 2015, 9531.[CrossRef] 41. Cherkashin, M.; Brenner, C.; Goring, L.; Dopke, B.; Gerhardt, N.C.; Hofman, M.R. Laser-diode-based photoacoustic setup to analyze Grüneisen relaxation-effect induced signal enhancement. In Opto-Acoustic Methods and Applications in Biophotonics II; International Society for Optics and Photonics: Belingham, WA, USA, 2015. 42. Upputuri, P.K.; Pramanik, M. Performance characterization of low-cost, high-speed, portable pulsed laser diode photoacoustic tomography (PLD-PAT) system. Biomed. Opt. Express 2015, 6, 4118–4129. [CrossRef] [PubMed] 43. Friedrich, C.-S.; Mienkina, M.P.; Brenner, C.; Gerhardt, N.C.; Jörger, M.; Strauß, A.; Beckmann, M.F.; Schmitz, G.; Hofmann, M.R. Photoacoustic blood oxygenation imaging based on semiconductor lasers. Photonics Optoelectr. 2012, 1, 48–54. 44. Wang, T.; Nandy, S.; Salehi, H.S.; Kumavor, P.D.; Zhu, Q. A low-cost photoacoustic microscopy system with a laser diode excitation. Biomed. Opt. Expr. 2014, 5, 3053–3058. [CrossRef][PubMed] 45. Arabul, M.U.; Heres, H.M.; Rutten, M.C.M.; van Sambeek, M.R.H.M.; van de Vosse, F.N.V.; Lopata, R.G.P. Ex vivo photoacoustic imaging of atherosclerotic carotid plaques. In Proceedings of the 2015 IEEE International Ultrasonics Symposium (IUS), Taipei, Taiwan, 21–24 October 2015. 46. Asundi, A.K.; Fu, Y.; Pramanik, M. Deep imaging with low-cost photoacoustic tomography system with pulsed diode laser. In Proceedings of the International Conference on Optical and Photonic Engineering (icOPEN 2015), Singapore, 14–16 April 2015. 47. Zeng, L.; Piao, Z.; Huang, S.; Jia, W.; Chen, Z. Label-free optical-resolution photoacoustic microscopy of superﬁcial microvasculature using a compact visible laser diode excitation. Opt. Expr. 2015, 23, 31026–31033. [CrossRef][PubMed] 48. Erfanzadeh, M.; Salehi, H.S.; Kumavor, P.; Zhu, Q. Improvement and evaluation of a low-cost laser diode photoacoustic microscopy system for ovarian tissue imaging. In Photons Plus Ultrasound: Imaging and Sensing; International Society for Optics and Photonics: Belingham, WA, USA, 2016. 49. Zeng, L.; Piao, Z.; Huang, S.; Jia, W.; Chen, Z. A practical optical-resolution photoacoustic microscopy prototype using a 300 mW visible laser diode. In Photons Plus Ultrasound: Imaging and Sensing; International Society for Optics and Photonics: Belingham, WA, USA, 2016. 50. Hariri, A.; Fatima, A.; Mohammadian, N.; Mahmoodkalayeh, S.; Ansari, M.A.; Bely, N.; Avanaki, M.R.N. Development of low-cost photoacoustic imaging systems using very low-energy pulsed laser diodes. J. Biomed. Opt. 2017, 22, 75001. [CrossRef][PubMed] 51. Allen, T.J.; Beard, P.C. Dual wavelength laser diode excitation source for 2D photoacoustic imaging. In Photons Plus Ultrasound: Imaging and Sensing, Proceedings of the Eighth Conference on Biomedical Thermoacoustics, Optoacoustics, and Acousto-Optics, San Jose, CA, USA, 20–25 January 2007; International Society for Optics and Photonics: Belingham, WA, USA, 2007; Volume 6437. [CrossRef] 52. Heres, H.M.; Arabul, M.U.; Rutten, M.C.; Van de Vosse, F.N.; Lopata, R.G. Visualization of vasculature using a hand-held photoacoustic probe: Phantom and in vivo validation. J. Biomed. Opt. 2017, 22, 41013. [CrossRef] [PubMed] 53. Petrova, I.Y.; Petrov, Y.Y.; Prough, D.S.; Esenaliev, R.O. Clinical tests of highly portable, 2-lb, laser diode-based, noninvasive, optoacoustic hemoglobin monitor. In Proceedings of the Photons Plus Ultrasound: Imaging and Sensing 2009, San Jose, CA, USA, 24–29 January 2009. Sensors 2018, 18, 2264 22 of 24

54. Upputuri, P.K.; Kalva, S.K.; Moothanchery, M.; Pramanik, M. Pulsed laser diode photoacoustic tomography (PLD-PAT) system for fast in vivo imaging of small animal brain. In Proceedings of the Photons Plus Ultrasound: Imaging and Sensing 2017, San Francisco, CA, USA, 28 January–2 February 2017. 55. Yin, X.; Dong, L.; Wu, H.P.; Ma, W.G.; Zhang, L.; Yin, W.B.; Xiao, L.T.; Jia, S.T.; Tittel, F.K. Ppb-level

H2S detection for SF6 decomposition based on a fiber-amplified telecommunication diode laser and a background-gas-induced high-Q photoacoustic cell. Appl. Phys. Lett. 2017, 111.[CrossRef] 56. Yin, X.; Dong, L.; Wu, H.P.; Zheng, H.D.; Ma, W.G.; Zhang, L.; Yin, W.B.; Jia, S.T.; Tittel, F.K. Sub-ppb nitrogen dioxide detection with a large linear dynamic range by use of a differential photoacoustic cell and a 3.5 W blue multimode diode laser. Sens. Actuator B Chem. 2017, 247, 329–335. [CrossRef] 57. Stylogiannis, A.; Prade, L.; Buehler, A.; Aguirre, J.; Sergiadis, G.; Ntziachristos, V. Continuous wave laser diodes enable fast optoacoustic imaging. Photoacoustics 2018, 9, 31–38. [CrossRef][PubMed] 58. Willert, C.; Stasicki, B.; Klinner, J.; Moessner, S. Pulsed operation of high-power light emitting diodes for imaging flow velocimetry. Meas. Sci. Technol. 2010, 21.[CrossRef] 59. Hansen, R.S. Using high-power light emitting diodes for photoacoustic imaging. In Proceedings of the Medical Imaging 2011: Ultrasonic Imaging, Tomography, and Therapy, Lake Buena Vista, FL, USA, 13–14 February 2011. 60. Dai, X.; Yang, H.; Jiang, H. In vivo photoacoustic imaging of vasculature with a low-cost miniature light emitting diode excitation. Opt. Lett. 2017, 42, 1456–1459. [CrossRef][PubMed] 61. Bui, D.A.; Hauser, P.C. Analytical devices based on light-emitting diodes-A review of the state-of-the-art. Anal. Chim. Acta 2015, 853, 46–58. [CrossRef][PubMed] 62. Mitterer, S.; Gannot, I.; Beck, T.; Erhardt, A.; Betz, C.; Schachenmayr, H.; Stepp, H. Transmission of LED-light through optical fibers for optical tissue diagnostics. In Proceedings of the Optical Fibers and Sensors for Medical Diagnostics and Treatment Applications VIII, San Jose, CA, USA, 19–21 January 2008; Volume 6852. 63. LED-Based Sun-Simulator Design: Technical and Commercial Considerations. Available online: https://www.photonics.com/a57253/LED-Based_Sun-Simulator_Design_Technical_and (accessed on 1 March 2015). 64. Lay-Ekuakille, A.; Vendramin, G.; Trotta, A. LED-based sensing system for biomedical gas monitoring: Design and experimentation of a photoacoustic chamber. Sens. Actuator B Chem. 2009, 135, 411–419. [CrossRef] 65. Böttger, S.; Köhring, M.; Willer, U.; Schade, W. Off-beam quartz-enhanced photoacoustic spectroscopy with LEDs. Appl. Phys. B 2013, 113, 227–232. [CrossRef] 66. Rey, J.M.; Sigrist, M.W. New differential mode excitation photoacoustic scheme for near-infrared water vapour sensing. Sens. Actuator B Chem. 2008, 135, 161–165. [CrossRef] 67. Adachi, Y.; Hoshimiya, T. Photoacoustic imaging with multiple-wavelength light-emitting diodes. Jpn. J. Appl. Phys. 2013, 52.[CrossRef] 68. Hariri, A.; Lemaster, J.; Wang, J.; Jeevarathinam, A.S.; Chao, D.L.; Jokerst, J.V. The characterization of an economic and portable LED-based photoacoustic imaging system to facilitate molecular imaging. Photoacoustics 2018, 9, 10–20. [CrossRef][PubMed] 69. Xia, W.; Singh, M.K.A.; Maneas, E.; Sato, N.; Shigeta, Y.; Agano, T.; Ourselin, S.; West, S.J.; Desjardins, A.E. Handheld Real-Time LED-Based Photoacoustic and Ultrasound Imaging System for Accurate Visualization of Clinical Metal Needles and Superficial Vasculature to Guide Minimally Invasive Procedures. Sensors 2018, 18, 1394. [CrossRef][PubMed] 70. Maneas, E.; Xia, W.; Singh, M.K.; Sato, N.; Agano, T.; Ourselin, S.; West, S.J.; David, A.L.; Vercauteren, T.; Desjardins, A.E. Human placental vasculature imaging using an LED-based photoacoustic/ultrasound imaging system. Proc. SPIE 2018, 10494.[CrossRef] 71. Gao, F.; Feng, X.; Zhang, R.; Liu, S.; Ding, R.; Kishor, R.; Zheng, Y. Single laser pulse generates dual photoacoustic signals for differential contrast photoacoustic imaging. Sci. Rep. 2017, 7, 626. [CrossRef] [PubMed] 72. Simandoux, O.; Prost, A.; Gateau, J.; Bossy, E. Influence of nanoscale temperature rises on photoacoustic generation: Discrimination between optical absorbers based on thermal nonlinearity at high frequency. Photoacoustics 2015, 3, 20–25. [CrossRef][PubMed] 73. Zharov, V.P. Ultrasharp nonlinear photothermal and photoacoustic resonances and holes beyond the spectral limit. Nat. Photonics 2011, 5, 110. [CrossRef][PubMed] Sensors 2018, 18, 2264 23 of 24

74. Mienkina, M.P.; Friedrich, C.S.; Gerhardt, N.C.; Wilkening, W.G.; Hofmann, M.R.; Schmitz, G. Experimental evaluation of photoacoustic coded excitation using unipolar golay codes. IEEE Trans. Ultrason Ferroelectr. Freq. Control 2010, 57, 1583–1593. [CrossRef][PubMed] 75. Allen, T.J.; Beard, P.C. Light emitting diodes as an excitation source for biomedical photoacoustics. In Proceedings of the Photons Plus Ultrasound: Imaging and Sensing 2013, San Francisco, CA, USA, 3–5 February 2013. 76. Mienkina, M.P.; Friedrich, C.S.; Gerhardt, N.C.; Beckmann, M.F.; Schiffner, M.F.; Hofmann, M.R.; Schmitz, G. Multispectral photoacoustic coded excitation imaging using unipolar orthogonal Golay codes. Opt. Express. 2010, 18, 9076–9087. [CrossRef][PubMed] 77. Beckmann, M.F.; Friedrich, C.-S.; Mienkina, M.P.; Gerhardt, N.C.; Hofmann, M.R.; Schmitz, G. Multispectral photoacoustic coded excitation using pseudorandom codes. In Proceedings of the Photons Plus Ultrasound: Imaging and Sensing 2012, San Francisco, CA, USA, 21–26 January 2012. 78. Zhang, H.; Kondo, K.; Yamakawa, M.; Shiina, T. Simultaneous multispectral coded excitation using periodic and unipolar M-sequences for photoacoustic imaging. In Proceedings of the Photons Plus Ultrasound: Imaging and Sensing 2013, San Francisco, CA, USA, 3–5 February 2013. 79. Azuma, M.; Zhang, H.K.; Kondo, K.; Namita, T.; Yamakawa, M.; Shiina, T. High frame rate photoacoustic computed tomography using coded excitation. In Proceedings of the Photons Plus Ultrasound: Imaging and Sensing 2015, San Francisco, CA, USA, 7–12 February 2015. 80. Beckmann, M.F.; Mienkina, M.P.; Schmitz, G.; Friedrich, C.S.; Gerhardt, N.C.; Hofmann, M.R. Monospectral photoacoustic imaging using Legendre sequences. In Proceedings of the 2010 IEEE International Ultrasonics Symposium, San Diego, CA, USA, 11–14 October 2010; pp. 386–389. 81. Beckmann, M.F.; Mienkina, M.P.; Schmitz, G.; Friedrich, C.S.; Gerhardt, N.C.; Hofmann, M.R. Photoacoustic coded excitation using periodically perfect sequences. In Proceedings of the 2011 IEEE International Ultrasonics Symposium (IUS), Orlando, FL, USA, 18–21 October 2011. 82. Su, S.-Y.; Li, P.-C. Coded excitation for photoacoustic imaging using a high-speed diode laser. Opt. Expr. 2011, 19, 1174–1182. [CrossRef][PubMed] 83. Beckmann, M.F.; Schmitz, G. Photoacoustic coded excitation using pulse position modulation. In Proceedings of the 2013 Ieee International Ultrasonics Symposium (IUS) 2013, Prague, Czech Republic, 21–25 July 2013. 84. Telenkov, S.A.; Mandelis, A. Fourier-domain biophotoacoustic subsurface depth selective amplitude and phase imaging of turbid phantoms and biological tissue. J. Biomed. Opt. 2006, 11.[CrossRef][PubMed] 85. Chiao, R.Y.; Hao, X. Coded excitation for diagnostic ultrasound: A system developer’s perspective. IEEE Trans. Ultrason Ferroelectr. Freq. Control 2005, 52, 160–170. [CrossRef][PubMed] 86. Fan, Y.; Mandelis, A.; Spirou, G.; Alex Vitkin, I. Development of a laser photothermoacoustic frequency-swept system for subsurface imaging: Theory and experiment. J. Acoust. Soc. Am. 2004, 116, 3523–3533. [CrossRef] [PubMed] 87. Gao, F.; Feng, X.; Zhang, R.; Liu, S.; Zheng, Y. Adaptive Photoacoustic Sensing Using Matched Filter. IEEE Sens. J. 2017, 1, 1–3. [CrossRef] 88. Gao, F.; Feng, X.; Zheng, Y. Advanced photoacoustic and thermoacoustic sensing and imaging beyond pulsed absorption contrast. J. Opt. 2016, 18, 074006. [CrossRef] 89. Li, G.; Gao, F.; Feng, X.; Zheng, Y. Analysis of stimulated Raman photoacoustics in frequency domain: A feasibility study. J. Appl. Phys. 2016, 120, 083105. [CrossRef] 90. Kellnberger, S.; Deliolanis, N.C.; Queiros, D.; Sergiadis, G.; Ntziachristos, V. In vivo frequency domain optoacoustic tomography. Opt. Lett. 2012, 37, 3423–3425. [CrossRef][PubMed] 91. Gao, F.; Peng, Q.; Feng, X.; Gao, B.; Zheng, Y. Single-Wavelength Blood Oxygen Saturation Sensing With Combined Optical Absorption and Scattering. IEEE Sens. J. 2016, 16, 1943–1948. [CrossRef] 92. Lashkari, B.; Choi, S.S.S.; Dovlo, E.; Dhody, S.; Mandelis, A. Frequency-domain photoacoustic phase spectroscopy: A ﬂuence-independent approach for quantitative probing of hemoglobin oxygen saturation. IEEE J. Sel. Top. Quantum Electr. 2016, 22, 127–136. [CrossRef] 93. Castelino, R.F.; Hynes, M.; Munding, C.E.; Telenkov, S.; Foster, F.S. Combined frequency domain photoacoustic and ultrasound imaging for intravascular applications. Biomed. Opt. Expr. 2016, 7, 4441–4449. [CrossRef][PubMed] 94. Dovlo, E.; Lashkari, B.; Mandelis, A. Co-registered frequency-domain photoacoustic radar and ultrasound system for subsurface imaging in turbid media. Int. J. Thermophys. 2016, 37.[CrossRef] Sensors 2018, 18, 2264 24 of 24

95. Maslov, K.; Wang, L.V. Photoacoustic imaging of biological tissue with intensity-modulated continuous-wave laser. J. Biomed. Opt. 2008, 13, 024006. [CrossRef][PubMed] 96. Zhao, Y.; Yang, S.H. Photoacoustic Viscoelasticity imaging of biological tissues with intensity-modulated continuous-wave laser. J. Innov. Opt. Health Sci. 2013, 6.[CrossRef] 97. Langer, G.; Buchegger, B.; Jacak, J.; Klar, T.A.; Berer, T. Frequency domain photoacoustic and ﬂuorescence microscopy. Biomed. Opt. Expr. 2016, 7, 2692–2702. [CrossRef][PubMed] 98. Gao, F.; Feng, X.; Zheng, Y.; Ohl, C.D. Photoacoustic resonance spectroscopy for biological tissue characterization. J. Biomed. Opt. 2014, 19, 067006. [CrossRef][PubMed] 99. Zhang, R.; Gao, F.; Feng, X.; Liu, S.; Ding, R.; Zheng, Y. Photoacoustic resonance imaging. IEEE J. Sel. Top. Quantum Electr. 2019, 25, 1–7. [CrossRef] 100. Feng, X.; Gao, F.; Xu, C.; Gaoming, L.; Zheng, Y. Self temperature regulation of photothermal therapy by laser-shared photoacoustic feedback. Opt. Lett. 2015, 40, 4492–4495. [CrossRef][PubMed] 101. Cunningham, V.; Lamela, H. Laser optoacoustic spectroscopy of gold nanorods within a highly scattering medium. Opt. Lett. 2010, 35, 3387–3389. [CrossRef][PubMed] 102. Gutrath, B.S.; Beckmann, M.F.; Buchkremer, A.; Eckert, T.; Timper, J.; Leifert, A.; Richtering, W.; Schmitz, G.; Simon, U. Size-dependent multispectral photoacoustic response of solid and hollow gold nanoparticles. Nanotechnology 2012, 23, 225707. [CrossRef][PubMed] 103. Beckmann, M.F.; Gutrath, B.S.; Buchkremer, A.; Eckert, T.; Timper, J.; Leifert, A.; Richtering, W.; Simon, U.; Schmit, G. Size dependent photoacoustic signal response of gold nanoparticles using a multispectral laser diode system. In Proceedings of the 2012 IEEE International Ultrasonics Symposium, Dresden, Germany, 7–10 October 2012. 104. Leggio, L.; Gawali, S.; Gallego, D.; Rodriguez, S.; Sanchez, M.; Carpintero, G.; Lamela, H. Optoacoustic response of gold nanorods in soft phantoms using high-power diode laser assemblies at 870 and 905 nm. Biomed. Opt. Expr. 2017, 8, 1430–1440. [CrossRef][PubMed] 105. Ding, R.; Zhang, X.; Chen, G.; Wang, H.; Kishor, R.; Xiao, J.; Gao, F.; Zeng, K.; Chen, X.; Sun, X.W. High-performance piezoelectric nanogenerators composed of formamidinium lead halide perovskite nanoparticles and poly(vinylidene ﬂuoride). Nano Energy 2017, 37, 126–135. [CrossRef] 106. Feng, X; Gao, F.; Zheng, Y. Modulatable magnetically mediated thermoacoustic imaging with magnetic nanoparticles. Appl. Phys. Lett. 2015, 106, 153702. [CrossRef] 107. Zhang, R.; Gao, F.; Feng, X.; Liu, S.; Kishor, R.; Luo, Y.; Zheng, Y. Noninvasive photoacoustic measurement of glucose by data fusion. Analyst 2017, 142, 2892–2896. [CrossRef][PubMed] 108. Gao, F.; Bai, L.; Liu, S.; Zhang, R.; Zhang, J.; Feng, X.; Zheng, Y.; Zhao, Y. Rationally Encapsulated Gold Nanorods Improving both Linear and Nonlinear Photoacoustic Imaging Contrast In Vivo. Nanoscale 2016, 9, 79–86. [CrossRef][PubMed] 109. Gao, F.; Bai, L.; Feng, X.; Tham, H.P.; Zhang, R.; Zhang, Y.; Liu, S.; Zhao, L.; Zheng, Y.; Zhao, Y. Remarkable In Vivo Nonlinear Photoacoustic Imaging Based on Near-Infrared Organic Dyes. Small 2016, 12, 5239–5244. [CrossRef][PubMed] 110. Leggio, L.; de Varona, O.; Escudero, P.; del Barrio, G.C.; Osinski, M.; Rivera, H.L. Improvement of signal-to-noise ratio of optoacoustic signals from double-walled carbon nanotubes by using an array of dual-wavelength high-power diode lasers. In Proceedings of the Opto-Acoustic Methods and Applications in Biophotonics II, Munich, Germany, 21–25 June 2015. 111. Hariri, A.; Hosseinzadeh, M.; Noei, S.; Nasiriavanaki, M.R. Photoacoustic signal enhancement: Towards utilization of very low-cost laser diodes in photoacoustic imaging. In Proceedings of the Photons Plus Ultrasound: Imaging and Sensing 2017, San Francisco, CA, USA, 28 January–2 February 2017. 112. Sun, M.; Feng, N.; Shen, Y.I.; Shen, X.; Li, J. Photoacoustic signals denoising based on empirical mode decomposition and energy-window method. Adv. Adapt. Data Anal. 2012, 4, 1250004. [CrossRef] 113. Ermilov, S.A.; Khamapirad, T.; Conjusteau, A.; Leonard, M.H.; Lacewell, R.; Mehta, K.; Miller, T.; Oraevsky, A.A. Laser optoacoustic imaging system for detection of breast cancer. J. Biomed. Opt. 2009, 14, 024007. [CrossRef][PubMed]

© 2018 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).