Photoacoustic Spectroscopy in Trace Gas Monitoring · 2019. 4. 3. · Photoacoustic Spectroscopy in Trace Gas Monitoring Frans J.M. Harren and Simona M. Cristescu RadboudUniversityNijmegen,Nijmegen,

Photoacoustic Spectroscopy inTrace Gas Monitoring

Frans J.M. Harren and Simona M. CristescuRadboud University Nijmegen, Nijmegen,The Netherlands

1 Introduction 12 Historical Overview 13 Devices and Equipment 2

3.1 Nonlaser Light Sources 23.2 Laser Light Sources 43.3 Photoacoustic Cells 53.4 Photoacoustic Detectors 83.5 Special Designs 93.6 Sensitivity, Selectivity, Limitations,

Interference, and Detection Limits 114 Applications 14

4.1 Atmospheric Applications 144.2 Detection of Chemical Warfare Agents and

Explosives 174.3 Biological and Agricultural Applications 184.4 Human Health, Noninvasive Breath

Analysis 205 Comparison with Other Spectroscopic Methods in

Trace Gas Monitoring 23Acknowledgments 24Abbreviations and Acronyms 24Related Articles 24References 24

Gas-phase spectroscopy is nowadays very common in awide variety of fields next to chemistry and physics, rangingfrom atmospheric chemistry to biology, medical sciences,and safety. Spectroscopic gas sensors have proven to beindispensable tools for these areas. There are various waysof utilizing gas sensors, and each application puts differentdemands. Some applications require a very high sensitivityfor one specific gas compound, while others benefit morefrom a sensor that has the ability to measure a wide rangeof gases or benefit from a miniaturized footprint. A high

Update based on the original article by Frans J.M. Harren, Gina Cotti, JosOomens, Sacco te Lintel Hekkert, Encyclopedia of Analytical Chemistry, ©2000, John Wiley & Sons, Ltd.

time resolution is also desirable, as well as selectivity,robustness, and little or no need for sample preparationand maintenance; photoacoustic spectroscopy and theirsensors meet a large number of these demands. This articlediscusses photoacoustic spectroscopy as sensitive, real-time, and noninvasive tool for trace gas monitoring. Afteran introduction and historic overview, attention is focusedonto the description of devices and equipment; they deter-mine the detection limits and selectivity. Applications arediscussed with emphasis on atmospheric applications,safety, and biological/agricultural applications followedby human health examples. Finally, a comparison is madewith other spectroscopy methods.

1 INTRODUCTION

A gaseous molecule that absorbs electromagnetic radi-ation is excited to a higher electronic, vibrational, orrotational quantum state. Generally, depopulation of thisquantum state to lower lying states occurs either via fluo-rescence or collisions; the latter gives rise to a temperatureincrease of the gas, owing to energy transfer to transla-tion. This nonradiative relaxation process occurs whenthe relaxation time can compete with the radiative life-time of the excited energy levels. Radiative decay has acharacteristic lifetime of 10−7 s at visible wavelengths ascompared to 10−2 s at 10 μm. For nonradiative decay, thesevalues depend on the pressure (decay time τ inverselyproportional to the pressure) and can vary strongly atatmospheric pressures (10−3 to 10−8 s).

By modulating the radiation source at an acousticfrequency, the temperature changes periodically, givingrise to a periodical pressure change, which can beobserved as an acoustic signal; in the gas phase the effectcan be detected with a sensitive microphone. Laser-basedphotoacoustic detectors are able to monitor trace gasconcentrations at atmospheric conditions with ordersof magnitude better sensitivity per centimeter absorp-tion path length, as compared to conventional scientificinstrumentation (see Section 5). In addition, they areable to monitor noninvasively and on-line under dynamicchanging conditions.

2 HISTORICAL OVERVIEW

The photoacoustic effect was first reported by AlexanderGraham Bell in 1880(1); he discovered that: ‘… thindiscs emitted sound when exposed to a rapidly inter-rupted beam of sunlight… ’. In a later experiment heremoved the eyepiece of a commercial spectroscope andplaced absorbing substances at the focal point of the

Encyclopedia of Analytical Chemistry, Online © 2006–2019 John Wiley & Sons, Ltd.This article is © 2019 John Wiley & Sons, Ltd.This article was published in the Encyclopedia of Analytical Chemistry in 2019 by John Wiley & Sons, Ltd.DOI: 10.1002/9780470027318.a0718.pub3

2 ENVIRONMENT: TRACE GAS MONITORING

(a)

(b)

B

CDA

Figure 1 (a) The eyepiece of a spectroscope is removed and substances are placed in the focal point of the instrument behind aslit. These substances are put in contact with the ear by means of a hearing tube.(2) (b) Sunlight is modulated with chopper B andfocused onto a glass bulb A.

instrument.(2) The substances were put in contact withthe ear by means of a hearing tube (Figure 1a and b) andhe found: ‘… good sounds in all parts of the visible andinvisible electromagnetic spectrum of the sun… ’. Otherpublications on this phenomenon followed this first work;we mention here the works of Röntgen,(3) Tyndall,(4) andPreece.(5) However, owing to the lack of a quantitativedescription and sensitive microphones, the interest in thephotoacoustic effect soon declined.

In 1938, Viegerov(6) refined the photoacoustic methodfor the first spectroscopic gas analysis; thereafter, Pfund(7)

and Luft(8) measured trace gas absorption spectra, withan infrared broadband light source down to the parts-per-million level. After the invention of the laser, atthe end of the 1960s, the scientific interest expandedagain. In 1968, Kerr and Atwood(9) utilized laser photoa-coustic detection to obtain the absorption spectrumof small gaseous molecules. Owing to the high spec-tral brightness of lasers and improved phase-sensitivelock-in techniques to amplify the acoustic signal, theywere able to determine low concentrations of air pollu-tants. Kreuzer(10) demonstrated that it was possible todetect concentrations of 10 parts per billion volume(1 ppbv= 1:109 mol/mol) of methane in nitrogen, usingan intensity-modulated infrared (3 μm) He–Ne laser.Patel et al.(11) demonstrated the application poten-tial by measuring the NO and H2O concentrationsat an altitude of 28 km with a balloon-borne spin-flip Raman laser. From there on, the photoacousticeffect was introduced into the field of spectroscopywith all its environmental, biological, and medicalapplications.

3 DEVICES AND EQUIPMENT

3.1 Nonlaser Light Sources

Lasers are not essential to operate photoacoustic gasdetection systems. Although the spectral power densityof broadband infrared lamps is orders of magnitudelower as compared to lasers, they have substantial advan-tages such as wide wavelength coverage, reliability, andcost-effectiveness. Such infrared lamp sources, in combi-nation with various photoacoustic detection schemes,are commercially available for trace gas detection atparts-per-million-volume (ppmv) levels for a wide rangeof molecular gasses, such as CO, CO2, NO, N2O, SO2,CH4, C2H4, and C3H8. Such design is a direct result ofthe developments made by Luft et al. in the 1950s and1960s.(8,12)

The infrared lamp-based gas analyzer uses a photoa-coustic detection scheme, which is able to detect a specificgas out of a multicomponent gas mixture, avoiding cross-interferences. In this instrument, selectivity is achieved bycomparing the direct absorption in a sample cell to thatof a reference cell (Figure 2). After passing the samplingcells, each attenuated light beam enters the detectioncell, filled only with the gas of interest (e.g. CO2); thedetection cells are interconnected via a membrane thatis connected to a capacitance. Since the dual beamis modulated by a chopper, the difference in acousticenergy reflects the difference in absorption and thusthe concentration difference between the sample celland the reference cell. Selection of wavelength occursby the species under investigation itself, in such a waythat all wavelengths at which absorption occurs are


PHOTOACOUSTIC SPECTROSCOPY IN TRACE GAS MONITORING 3

Equalizer

Membrane

M2

M1

Chopper

Filter

Splitter

IR source

Reference

Figure 2 Infrared gas analyzer with photoacoustic detection scheme to detect a specific gas out of a multicomponent gas mixture,thereby avoiding cross interference (ABB GmbH, Frankfurt). Light from the infrared source is split into two paths. The choppermodulates the intensity for both paths. A filter volume in each path serves to filter out light of wavelengths not needed for thedetection process; they can be filled with gases of which the absorption spectra do not overlap with those of the species under scrutiny.M1 and M2 serve as measuring cell and reference cell, respectively. With the help of the equalizer both light intensities become equalbefore entering the last cell. The last cell consists of two compartments with a membrane in between. Both compartments are filledwith the gas under investigation so that all wavelength characteristics for this gas contribute to the signal. If the attenuation differsin M1 from that in M2 the membrane starts to oscillate with the frequency of the chopper. This oscillation (typically a few hertz) isdetected capacitively.

Infrared source

Opticalfilter wheel

Chopper wheel

Mirror

Optical filter

Optical window

Microphone 2

Microphone 1

Pump

Analysis cell

Samplingtube

Inlet valve

Outletvalve

Flush valve

Internalfine air-filter

Air outlet

Air-shunt

Shunt valve

Pressure sensor

Air inlet

Figure 3 Light from an infrared light source is reflected off a mirror and is passed through a mechanical chopper and then throughone of the optical filters in the filter wheel (LumaSense Technologies, Denmark). The gas being monitored, causing the temperatureof the gas to increase selectively, absorbs the light transmitted by the optical filter. The filter wheel turns so that light is transmittedthrough the next optical filter, and the new signal is measured. The number of times this step is repeated is dependent on the numberof gases being measured. The response time is 30 s if five gases and water vapor are measured.

simultaneously active. When other gases absorb in thesame spectral region, additional absorptions at otherwavelengths than the gas of interest will not contributeto the acoustic signal; the light passes the detection cellunattenuated. When a specific compound, e.g. H2O,causes strong spectral overlap, an extra cell can be placedin the light path filled with the interfering gas. This cellattenuates completely the wavelengths where this inter-fering molecule absorbs, including the spectral overlapregions. Thus, these wavelengths cannot contributeto the photoacoustic signal, and a single component

can be detected out of complex multicomponent gasmixtures.

Spectral selectivity can also be achieved throughthe use of optical filters, see Figure 3. By installing anumber of optical filters in a rotation cartwheel beforethe light passes through the photoacoustic gas samplingcell, it selects these specific wavelengths at which thegas of interest has it strongest absorption bands. Sucha system can measure mixtures of concentrations inany air sample. Detection limit is gas dependent, butis typically in the sub-ppmv region. Such instruments



2 3 4 5 6 7 8 9

Re

lative

ab

so

rptio

n

Re

lative

po

we

r

1.2

(a)

(b)

1

0.8

0.6

0.4

0.2

0

0 3 4 5 6 7 8 9

Wavelength (μm)

CH4 CO2 SO2

LED 3.4 μm LED 4.2 μm

LED 7.0 μm

Figure 4 (a,b) Relative powers of the LEDs used in experi-ments as a function of the wavelength. The dip in the spectrumof the LED light at 4.2 μm is due to absorption of CO2, andthree dips at between 5 and 9 μm are due to absorption of water.Panel (a) shows the absorption spectrum of methane (CH4),carbon dioxide (CO2), and sulfur dioxide (SO2), all based onthe HITRAN database.(16) (Reproduced with permission fromRef. 15. © Elsevier, 2009.)

require no consumables and very little maintenance andare therefore ideally suited for permanent monitoringtasks (environment and industrial).

Instead of using optical filters, higher spectral resolu-tion can be achieved if photoacoustic detection is alsocombined with Fourier transformed infrared (FTIR)spectroscopy. FTIR is a widely used method for obtaininginfrared spectra of a sample. Infrared radiation is split ina Michelson interferometer where half of the light passesthrough to a fixed mirror and the other half is reflectedtoward a moving mirror. The two beams recombine andpass through a photoacoustic gas cell where the sampleabsorbs light at molecule-specific frequencies. Usingphotoacoustics, a step-scan interferometer is used. Themoving mirror is moved in discrete steps and haltedat each retardation position at which the interfero-gram is sampled. After each point is sampled, a Fouriertransformation is performed. Within the FTIR, a short,small volume photoacoustic cell is used making use ofa Helmholtz resonance. In such a way, sensitivity ofgas-phase FTIR is improved to low-ppmv levels for awide range of gases.(13,14)

Nowadays, light-emitting diodes (LEDs) are an inter-mediate alternative with reduced spectral range, butincreased power density as compared to infrared lamps.LEDs are very convenient sources for photoacoustics,since they have small size and low power consumption.The full width at half maximum (FWHM) of the emissionspectrum is typically 10–20% of its peak wavelength,which is sufficient for most single gas measurement. TheLED output is easily modulated by current pulsing, withfrequencies up to several megahertz; its brightness andon/off contrast do not decline as modulation frequencyis increased, as with traditional blackbody emitters. Withmid-infrared LEDs, photoacoustic trace gas detection isperformed in the 3–7 μm wavelength range,(15) using acantilever instead of a microphone. Detection at the low-ppmv levels can be achieved for gases such as methane,propane, CO2, and SO2, with 1-s sample integrationtime, see Figure 4. Using UV LEDs at 285 nm as a lightsource, ozone could be sensed down to 1 ppmv basedon quartz-enhanced photoacoustic spectroscopy.(17) Withthe same method, an NO2 sensor was developed using ablue LED down to a detection limit of 1.3 ppbv.(18)

3.2 Laser Light Sources

High spectral brightness renders continuous wave (cw)laser sources ideally, suitable for photoacoustic tracegas detection. At small absorptions, the photoacousticsignal is proportional to the laser power; from theBeer–Lambert law, one finds:

P = P0e−σ⋅N⋅l ⇒ ΔP = P0 − P ≈ P0 ⋅ σ ⋅ N ⋅ l (1)

with ΔP the difference of laser power before and afterthe photoacoustic cell, σ the absorption cross-section permolecule (in cm2), N the number of absorbing moleculesper cm3, and l the absorption path length (in cm). Theabsorbed energy (P0 −P) is converted into acousticenergy and recorded by a microphone. As can be seenfrom this equation, for small absorptions the generatedacoustical signal is proportional to the incoming laserpower (high laser powers are advantageous) and thegas concentration (linearity of the signal). A nonlinearabsorption response can occur in specific cases, suchas using focused, high-power laser beams. In that casepumping rate by the laser light to a higher rovibrationallevel exceeds the collisional de-excitation rates, andsaturation occurs. Another nonlinear response is causedby kinetic cooling in the gas; the latter is explained later.

At present, photoacoustic detection goes together withbright laser light sources, ranging from the infrared tothe visible and UV. The high-power, cw molecular COand CO2 lasers were the first lasers to detect trace gasesphotoacoustically, with high sensitivity and selectivity.(19)



Both the lasers were for a long time the most used lasersfor infrared photoacoustic spectroscopy.(20–24) Althoughthese lasers have a wide wavelength coverage and highpower, owing to their dependence on molecular emis-sion lines, they do not have a full wavelength coverage.The developments in solid-state physics and nonlinearoptics have expanded the potential for photoacousticspectroscopy to other wavelength regions. The recentavailability of high-power (cw and pulsed) quantumcascade lasers (QCLs), operating at room temperaturewith power levels up to 1 W, has boosted photoacousticspectroscopy in the mid-infrared wavelength region.Commercially available external cavity QCLs, operatingsingle-mode tuning ranges, approach a 300-cm−1 wave-length coverage; QCLs and the related interband cascadelasers (ICLs) operate in the mid-infrared wavelengthregion from 3 to 24 μm.(25,26)

For the 2.5–5 μm wavelength region, high-power,periodically poled lithium niobate (PPLN) optical para-metric oscillators (OPOs) are available. Nowadays, suchOPOs are pumped by high-power, near-infrared (fiber-)lasers with excellent spectroscopic properties (narrowlinewidth, Gaussian beam profile, and fast wavelengthscanning), which the OPO can convert efficiently intothe mid-infrared wavelengths, owing to its instantaneousnonlinear effects.(27) In the near-infrared wavelengthregion, lasers are used for high-resolution studies ofweak overtone absorption spectra of molecular gases.(28)

Overtone spectroscopy profits from sensitive photoa-coustic detection, since one deals with high vibrationalexcitation, where fluorescence-based techniques do notwork. Electronic spectroscopy with light sources in thevisible or in the ultraviolet does not immediately askfor photoacoustic detection because other techniquesprovide more powerful means (e.g. laser-induced fluo-rescence (LIF) or resonance-enhanced multiphotonionization).

In the ‘telecom’ wavelength region, around 1.5 μm,diode lasers and diode pumped fiber lasers have theadvantage of the technical developments, resulting in asmall footprint, reliable tunability, low costs, large life-time, and low maintenance. Diode lasers are not alwaysin favor to be combined with photoacoustic spectroscopy,owing to their low power; however, in combination withfiber amplifiers this disadvantage can be compensated.Larger tuning ranges can be obtained by external cavityset-ups.(29)

By using pulsed lasers, the wavelength region can beextended and in combination with gated detection theacoustic signal can be isolated from the surroundingacoustical noise and acoustic background signal anddetected by a piezo transducer (or microphone).However, in general, pulsed lasers have high inten-sity, pulse-to-pulse intensity fluctuations, and a varying

intensity profile causing nonlinear effects in molecularabsorption, thereby reducing detection sensitivity. Thehighest photoacoustic detection sensitivities have beenachieved for modulated cw lasers. If no other argumentsprevail, the use of cw lasers is recommended.

3.3 Photoacoustic Cells

3.3.1 Nonresonant Cells

To perform trace gas detection, the ideal photoacousticcell should amplify the generated sound, originatingfrom the molecular gas absorption, meanwhile rejectingacoustic (and electric) noise and in-phase infraredabsorption from other materials. Interfering gases shouldbe distinguished by spectroscopic or physical methods.For gas-phase measurements, mainly resonant cellsare combined with modulated cw lasers and lock-inamplifiers; pulsed lasers are combined with piezoelectricdetectors and boxcars. These selective amplifiers arisefrom the necessity to lower acoustic and electric noiselevels, thus improving the signal-to-noise ratio (SNR).Other requirements for photoacoustic cells are a lowgas consumption or a fast response; for this the activevolume of the cell should be small so that no dilutioncan take place when the trace gas and its carrier gasflow through the photoacoustic cell. If we consider anonresonant, cylindrical cell, its performance can beexpressed as its efficiency to convert absorbed photonenergy into acoustic energy, i.e. pgas =F ⋅ΔP, with pgas thegenerated acoustic pressure in the gas and ΔP derivedfrom Equation (1). The cell constant F in (Pa m W−1) canbe derived by:

Fnon resonant =G(γ − 1)lnr

2πνVnr(2)

with lnr and Vnr the length and the volume of the nonres-onant (nr) cell, respectively, γ the specific heat constant,ν the modulation frequency, and G a geometrical factorin the order of 1.(30) From the above formula, one canderive that F is independent of the cell length. The diam-eter should be as small as possible to generate high signals,just large enough to allow proper alignment of the inci-dent light beam. To minimize the detection of absorptionby the cell windows, one has to maximize the cell length;the energy absorbed by the windows is then distributedover a larger volume. The photoacoustic signal decreaseswith higher modulation frequencies; therefore, it is advan-tageous to use low modulation frequencies.

3.3.2 Resonant Cylindrical Cells

In a resonant cell, the modulation frequency is chosento match one of the resonant frequencies of the cell.



Longitudinal

L

r

z

2 R

Azimuthal Radial

k = 0,1, 2 .... m = 0,1, 2 .... n = 0,1, 2 ....

Figure 5 Resonant acoustic modes of a cylindrical closed chamber, the fundamental longitudinal, azimuthal, and radial modes.

This produces a standing, amplified sound wave in theresonator. For a gas-type resonant cell, two varietiesof acoustical resonators have found widespread use:the Helmholtz resonator and the cylindrical resonatorexcited in a longitudinal, radial, and/or azimuthal mode(Figure 5).

For cylindrical resonant cells, one has to multiply the cellconstant F (Equation 2) with the quality factor Q of thegenerated acoustic resonance:

Fresonant =G(γ − 1)lrQ

2πνVr(3)

with lr and Vr the length and the volume of the acousticresonator, respectively; Q equals the ratio of the energystored in the acoustical standing wave over the energylosses per cycle. For resonant cylindrical cells, one canderive for the cell constant:

Fresonant ∝√

lrR

(4)

with R the radius of the resonator.(30) Cell geometrieswith large diameter-to-length ratios, to excite the reso-nance in the radial or azimuthal acoustic mode, possesshigh Q-values and high resonance frequencies. However,they have low F-values. Photoacoustic cells with highQ-values are sensitive to long-term drifts, owing totemperature fluctuations (speed of sound is temperaturedependent). Such cells require active locking of the modu-lation frequency on the resonance frequency of the cell.Besides, such cells pickup easily acoustic surroundingnoise.

In longitudinally excited resonators, a smaller acousticgain, as a consequence of a relatively low Q-value, iscompensated for by the signal gain due to the smallerdiameter (Equation 4, Figure 6). The resonant acousticalamplification is limited by various dissipation processes,which can be divided into surface and volume effects. The

main losses are viscous and thermal losses at the resonatorsurface, microphone losses, and acoustic wave scatteringlosses at obstacles inside the cell. The less-important volu-metric losses are radiation losses, free-space viscous andthermal losses, and dissipative relaxation processes withinthe absorbing gases. For a cylindrical resonator, the reso-nant frequencies are given by(32) (Figure 5):

νk,m,n = Vs

[{αm,n

d

}2

+{

k2l

}2]1∕2

(5)

with Vs the sound velocity of the gas inside the cavity,d the cell diameter, l the axial length, k= 0,1, 2,3, …the axial mode numbers, and αm,n the nth root of thederivative of the Bessel function: dJm/dr= 0 at r = R(e. g. α0,0 = 0, α0,1 = 1.2197, α1,0 = 0.5861, and α1,1 = 1.6970).

3.3.3 Helmholtz Resonant Cells

Another type of resonant cell is a Helmholtz resonator;it consists of a closed volume (cavity) connected via along narrow tube to the microphone (Figure 7). It is theacoustic equivalent of a mechanical oscillator composedof a mass (long tube) and a spring (cavity). The oscillationfrequency (νH) of the system depends on the length (lH),the cross-section (AH) of the tube (or duct), and thevolume (VH) of the cavity via the following relation:

νH = VS

{AH

VHlH

}1∕2

(6)

with Vs the velocity of sound.(21) Although the acousticalresponse of such a resonator is somewhat lower ascompared to cylindrical resonators, it has advantageswhen low or high temperatures are needed. Micro-phones have a limited operating temperature. Whengases are investigated at low (e.g. 80 K) or high (350 K)temperatures, it is advantageous to have a homogeneous



Gi

Go

GoGo

BW

M

BW

M

TACNF Gi

Laser beam

0 5 10 cm

0 5 10 cm

(a)

(b)

Figure 6 Two photoacoustic cell types successfully used in trace gas detection. (a) A banana-shaped cell, where the laser beamenters from the left through a Brewster window (BW) to leave the central part (l = λ/2) of the resonator at the right Brewsterwindow. The position of the microphone M is indicated. The two pieces of length λ/4 before and after the bends serve to suppressthe window signal, which are at the pressure nodes of the standing acoustic wave. The total length of the banana cell amounts to λ.Also, at the bends are the gas inlet and outlet (Gi and Go). (Reproduced with permission from Ref. 31. © Springer-Verlag, 1987.)(b) The open-organ-tube cell; the main part is manufactured of a block of massive material (e.g. brass) to suppress pickup fromthe surrounding noise. The central part acts as an open acoustical resonator with length λ/2; the resonator tube itself has a highlypolished, gold-coated inner surface to minimize wall heating by stray light and reduce surface losses. The central position of the gasinlet (Gi) is essential to obtain short response times using a gas flow. Only the resonator volume (here 2.8 mL) has to be replenishedbefore an independent concentration measurement can be performed; TAC : λ/4 tunable air columns to suppress window signal;NF : λ/4 notch filter to suppress acoustic in-coupling of acoustic noise via the gas inlet. (Reproduced with permission from Ref. 30. ©AIP Publishing, 1996.)

Lightsource

Windows

Modulatedlight beam

(a) (b)

Duct

Sample

BaF2 window

102Tap

Microphone holder

cavity

Microphonecavity Microphone

Figure 7 Two types of photoacoustic Helmholtz resonators. (a) A normal Helmholtz resonator. (Reproduced with permission fromRef. 33. © Mariusz Suchenek, 2011.) (b) A differential Helmholtz resonator. (Reproduced with permission from Ref. 34. © Elsevier,1999.)



Cooling/heading

0 5 10 cm

Microphone

Gas

Gas

LaserV1

Water bath (278–350 K)

Stainless steel Aluminum

Figure 8 Photoacoustic cell applied for temperature-dependent investigations on fatty acids. The temperature ofthe water bath was varied between 278 and 350 K with a coldfinger and two immersion heaters, while the temperature of themicrophone was kept constant with a cooling/heating device.(Reproduced with permission from Ref. 36. © Springer-Verlag,1996.)

temperature distribution across the gas volume whilethe microphone operates at room temperature.(35) Thetemperature gradient is over the long narrow tube of theresonator at which end the microphone is placed.

A way to improve the sensitivity for a specific gas isa temperature change. By increasing the temperature,the vapor pressure of, e.g. oil with low vapor pressure isincreased, thereby improving the sensitivity. A Helmholtzresonant photoacoustic cell is an ideal configuration forthis, since the microphone can be at room temperature(Figure 8).(36)

3.4 Photoacoustic Detectors

3.4.1 Microphones

Pressure-detection methods are widely available rangingfrom the large membranes as in the earliest designs ofLuft(8) to piezoelectric transducers operating at mega-hertz frequencies. Normally, for photoacoustic detectionof gases, audio microphones are used. Microphones rangefrom very small size (1-mm membrane) for hearing aidpurposes behind the ear to large-size (25-mm membrane)very sensitive (1000 mV Pa−1) professional microphoneswith a spectrally flat response to sound pressure (Brueland Kjaer microphones). The small-size microphones(such as from Knowles) have mostly a peak in acousticamplitude response around 1 kHz with a typical responseof 10 mV Pa−1.(37) Because of the low costs, mass-produced micro-electro-mechanical system (MEMS)microphones, commonly built into smartphones, are

nowadays also incorporated in photoacoustic cells.(38)

Besides the fact that the latter are inexpensive, they havethe advantage that they do not take up so much acousticenergy as compared to large-size microphones.(30) Thisis important in the case of resonant acoustic cells, forwhich increased friction losses will lower the qualityfactor of the resonator and, thus, the generated acousticamplitude.

3.4.2 Quartz Tuning Fork

In gas-phase spectroscopy, the common approach is touse a resonant photoacoustic cell and accumulate theabsorbed energy in the gas. However, an alternativelydeveloped approach is to store the absorbed energy ina sensitive element. A well-suited material for this is ahigh-Q quartz crystal element with piezoelectric proper-ties. Originally, these elements are mass produced andinexpensive; high-Q quartz crystals are used as frequencystandard in clocks, watches, or smart phones. The mechan-ical and electrical properties of tuning forks (TFs) wereextensively studied in relation with their application inultrahigh resolution scanning microscopy.(39) Usually, aquartz TF has a resonant frequency close to 32 768 Hz.A typical TF has a Q-value of 100 000 in vacuum; atnormal atmospheric pressure this is lower (∼10 000),owing to the viscous properties of air.(40,41) The mode atthis frequency corresponds to a symmetric vibration (theprongs move in opposite directions), while the antisym-metric vibration is piezoelectrically inactive. This makesthe TF an acoustic quadrupole, which provides good envi-ronmental noise immunity. In addition, with its narrowresonance (∼4 Hz) only a few frequency components canproduce efficient excitation of the TF vibrations. Thetypical energy accumulation time at atmospheric pres-sure is ∼300 ms, which is quite longer than any prac-tical gas-filled resonator. Another property is that the gassampling volume in between the prongs is extremely small(∼1 mm3).

Nowadays there are several approaches developed toincrease the absorption path length, and thus the detec-tion sensitivity. One way is to add microresonators (mRs)to the TF (Figure 9a). This ‘on-beam’ configuration hasacoustic mRs with two thin tubes aligned at micrometerdistance near the TF. The laser beam will enter via themRs, where it is critical that the laser beam does nottouch the walls of the resonator; the latter will producecoherent acoustic background effects. The mRs will beacoustically coupled to the TF, and the combined acousticbehavior will result in a lower Q-factor, a slightly changedresonant frequency and an increased signal. The on-axisconfiguration has the disadvantage that the open-endedmRs introduce sound radiation losses. Furthermore, thegap between the prongs of the TF is ∼300 μm; the inner



0.71 mm

4.40 mm

0.30 mm

3.20 mm

0.7 mm

8.00 mm

0.30 mm

3.20 mm

(a) (b)

Figure 9 A tuning fork (TF) with additional microresonators to enhance the acoustic signal. (a) The ‘on-beam’ configuration, withthe two microresonator tubes aligned perpendicular to the TF. (b) The ‘off-beam’ configuration, in which the microresonator tubehas an aperture close to the prongs of the TF. (Source: Pietro, https://www.mdpi.com/1424-8220/14/4/6165/html.(41))

diameter of the mR has to be of the same dimensions,which limits the size of the laser beam that passes throughthe tubes.

An alternative configuration is the ‘off-beam’approach,(42) where the TF measures the pressure inthe mR through a small aperture in its center. Now themR is a single tube and its length determines the firstlongitudinal mode of the acoustic wave. A small slitis made in the middle of the mR for optimal acousticcoupling to the TF (Figure 9b). The off-beam-QEPASconfiguration results in certain technical advantages asit facilitates the optical alignment and allows more flex-ibility in the selection of the dimensions of the quartzTF. In off-beam-QEPAS, the acoustic oscillations of thegas are excited in the mR by the intensity modulationinduced by the externally located laser source.

3.4.3 Cantilever-type Optical Microphone

Microphones have a small gap below their membrane,which cannot be decreased below a certain level; thislimits microphone sensitivity since the air flow in and outof this gap requires energy and therefore creates signifi-cant damping in the membrane. Moreover, the membranestretches out radially, which limits the dynamical rangewith a nonlinear response.(15) A cantilever-type pressuresensor has been proposed as shown in Figure 10.(43) Athin (t= 10 μm) cantilever portion moves like a flexibledoor, owing to the pressure variations in the surroundinggas. There is a narrow (Δ< 5 μm) gap between thethicker frame and the cantilever on three sides (l= 6 mm,w= 4 mm).

As the pressure varies, the cantilever only bendsand does not stretch. Therefore, the movement ofthe cantilever’s free end can be about two orders ofmagnitude greater than the movement of the middlepoint of a tightened membrane under the same pres-sure variation. The displacement of the cantilever ismeasured via a compact Michelson-type laser interfer-ometer (Figure 10). The laser beam is focused on the freeend of the cantilever. Both arms of the interferometerare set such that there is a half of an interference fringeover each of the four photodiodes.

The abovementioned cantilever system has a large sizeand operates at low frequencies in a nonresonant cell withrelative low SNR. Alternatively, cantilever-based reso-nant photoacoustic spectroscopy is proposed on top ofa fiber using a small-size Fabry–Perot interferometer.(44)

The schematic structure and the image of the fiber-tipFabry–Perot acoustic sensor head, with stainless steelshell, ceramic ferrule, and stainless steel cantilever,are shown in Figure 11. The air gap between thecantilever and the cleaved end facet of the fiber forms theFabry–Perot cavity (length∼ 1.3 mm). The rectangularstainless steel cantilever is 1.9 mm× 0.8 mm and has athickness of 10 μm.

3.5 Special Designs

In the past, special designs have been developed forlongitudinally, azimuthally, and radially resonant photoa-coustic cells(30,32) even without windows.(45–47) In order toimprove the selectivity, the combination of the Stark orZeeman effect with photoacoustic detection represents



l

l

t

w

Δ

Modulated IR radiation

Side viewt = 5 μmI = 4 mm

Top viewΔ = 30 μmW = 2 mm

(a)

(b)

Reference cell N2

Sample cell cx, p

IR windows

L = 147 mm l = 95 mm

Cantilever

A

B

Mirrors

Michelsson-type interferometer

c0x, p0

Photoacoustic cell

P(ν)

P(ν)

Figure 10 (a) The dimensions of the silicon cantilever and the gap between the cantilever and the frame. (b) The set-up containingthe cantilever is in principle the same set-up as the nondispersive set-up from Figure 2, but now with a cantilever detection set-up.It consists of three cells: the reference cell, the sample cell, and the photoacoustic cell. The photoacoustic cell acts as a detectorand includes high concentration of the gas to be detected. The cell is divided into two parts, A and B, which are separated by thecantilever and its frame. The position of the cantilever is proportional to the pressure difference between the cell parts A and B. Thedisplacement of the cantilever end is measured continuously with a laser Michelson interferometer. (Reproduced with permissionfrom Ref. 43. © Elsevier, 2006.)

Glue Ceramic ferrule

Stainless steel shell(a) (b) (c)

Fiber

Cantilever

Figure 11 Cantilever-enhanced resonant photoacoustic spectroscopy. (a) Schematic structure. (b) Image of the cantilever sensorhead. (c) Image of the stainless steel cantilever. (Reproduced with permission from Ref. 44. © Elsevier, 2018.)

an interesting solution for specific molecules such asammonia (NH3)(48) and NO.(49)

Other designs are based on optical power enhancement.Since the photoacoustic signal goes proportional with theapplied laser power, it is favorable to use a photoacousticcell inside an optical cavity. This can be either by using anintracavity set-up, in which the photoacoustic cell is withinthe optical laser cavity or by using an external cavity.Intracavity photoacoustic set-ups are demonstrated forCO2 and CO lasers in the mid-infrared wavelength region

gaining up to 100× in power,(20,22,50) or using externalcavity diode lasers and fiber lasers in the near-infraredwavelength region.(51,52)

Another approach is to use an optical cavity locked tothe laser in such a way that laser power is built up in theoptical cavity.(53,54) An example is shown, using quartz-enhanced photoacoustic spectroscopy and a distributedfeedback QCL emitting at 4.33 μm (Figure 12). Thelaser optical power build-up factor is ∼500×, whichcorresponds to an intracavity laser power of 0.75 W. The



QCLReference cell

Optical resonator

Functiongenerator

× 30Lock-inamplifier

Currentdriver

HgCdTedetector

PZT

Lock-in+ PI

QTF

CEU

IM

MML

LMHM

M

DC +LM

λ/4

OC

Oscilloscope

InSb detector

PBS

Figure 12 Schematic representation of the intracavity quartz tuning fork, enhanced photoacoustic set-up. Owing to tight lockingof the laser and optical cavity, a power enhancement of ∼500 is achieved. PBS, polarizing beam splitter; λ/4, quarter-wave plate; M,mirror; MML, mode-matching lens; IM, input mirror; OC, output coupler; PI, proportional-integral controller; LM, low-frequencymodulation (linear ramp); HM, high-frequency modulation (sinusoidal dither); CEU, control electronics unit. (Reproduced withpermission from Ref. 53. © AIP Publishing, 2014.)

normalized noise equivalent absorption (NNEA) forsuch a set-up is typically ∼5× 10−10 W cm−1 Hz−1/2.

3.6 Sensitivity, Selectivity, Limitations, Interference,and Detection Limits

Laser-based absorption spectroscopy performs well interms of sensitive and selective detection of trace gases,and it allows on-line measurements. In contrast to othersensing methods, such as mass spectrometry, the set-upcan be compact, affordable, and with low maintenancecosts. The high sensitivity and selectivity are due toseveral important factors. Firstly, the narrow linewidth oflasers gives a high spectral power density as comparedto broadband light sources. Especially for cw lasers, thislinewidth is typically much narrower than the molecularabsorption lines. This causes the total laser emission tobe attenuated by the gas sample under investigation,instead of only a small fraction, as it is the case in broad-band absorption. Thus, the absorption signal increasesstrongly due to its narrow linewidth. In addition, theselectivity improves because it enables a wavelength scanover an absorption feature. This helps to distinguish thetarget molecule from the interfering compounds andbackground signals. Selectivity enhances further by theunique spectral absorption spectrum of each molecule,which corresponds to different rovibrational transitionsof the molecule. Spectroscopic gas detection probesthe internal structure of the molecule, whereas other

techniques such as mass spectrometry only assess a globalproperty of the investigated compound.

In the infrared fingerprint wavelength region(2.5–25 μm) small molecules have their characteristicstrong absorption lines (Figure 13). Combined withhighly sensitive spectroscopic methods, gases can bedetected at trace levels below the parts-per-billion-volume range (1 ppbv= 1:109 mol/mol). A disadvantageof laser-based spectroscopy is that larger moleculeshave more vibrational modes (3N-6, with N the numberof atoms in the molecule) and a smaller rotationalconstant. At room temperature, this results in broadstructures of overlapping rotational lines with muchlower spectroscopic selectivity.

Initially, molecular CO2 and CO lasers dominatedmolecular detection of gases in the mid-infrared due totheir high-power (photoacoustic signal is proportionalto the laser power) and wide wavelength coverage. Inthe past decades, the selective detection of moleculargasses improved strongly for a wide range of gasses withthe development of QCLs and OPOs. The inventionof the QCL made it possible to operate current-drivendiode-like lasers at room temperature in the mid-infraredwavelength region from 3.5 to as long as 19 μm.(25)

The first cw singly-resonant OPOs, suitable for trace gasdetection, were developed in 1996; they contained a quasi-phase-matched PPLN crystal with a long interactionlength (several centimeters) to lower the threshold foroperation.(27,55)



1800

800 900 1000 1100 1200 1300

1e–20

1e–21

1e–20

1e–19

1e–18

1e–19

1e–18

Lin

e s

trength

(cm

−1/m

ole

c.c

m−2

)Lin

e s

trength

(cm

−1/m

ole

c.c

m−2

)1e–17

NO: 5.26 μm

OCS

CONO

CH4

CH4SO2

NH3O3N2O

C2H4

CH2O

CO2

N2O

HCl

2000 2200 2400 2600

Wavenumber (cm−1)

Wavenumber (cm−1)

CO2: 4.3 μm

O3: 10 μmN2O, CH4: 7.66 μmNH3: 10.6 μm

CH2O: 3.6 μm

COS: 4.86 μm

3.1–5.5 μm

7.6–12.5 μm

CO: 4.66 μm

CH4: 3.3 μm

2800 3000 3200

Figure 13 Molecular line strength for a number of molecules in the mid-infrared wavelength region. It demonstrates the spectralselectivity and sensitivity for a number of interesting gasses Line strength data can be found at the Hitran database. (Reproducedwith permission from Ref. 16. © Frank Tittel, 2018.)

In the near-infrared wavelength region, photoacousticsis not favorable for trace gas detection. In this region thereare mainly weak vibrational overtone absorption bands,in comparison to the (100–1000 times) stronger funda-mental bands in the mid-infrared region. However, othertrace gas sensing methods, based on direct absorptionspectroscopy, are successfully used in this region (e.g.cavity ring-down spectroscopy and cavity-enhancedabsorption spectroscopy).(56,57) They profit from the high-quality commercially available telecom lasers, the long(cavity-enhanced) absorption path lengths (mirrors withextreme high reflectivity: >99.99%), and the excellentdetectivity (D*) of the available detectors to detect asmall ΔP (Equation 1). In addition, telecom lasers havemostly low power output; therefore, photoacoustics isnot competitive. In the mid-infrared wavelength region,this advantage does not exist, because of the less well-developed infrared mirrors and less-sensitive opticaldetectors.

In order to have a maximum SNR, one has to take intoaccount the noise contribution. The theoretical minimumacoustical noise results from random pressure fluctua-tions and their frequency distribution due to Brownianmotion. The total power of these pressure fluctuations is

constant, but the noise spectrum depends on the Q-valueof the acoustical resonance of the photoacoustic cell; ahigh Q-value will significantly amplify noise power atthese resonances. However, often this Brownian noiselevel is below what is produced by other noise sources,such as electronic amplifier noise and external acousticalnoise from the surroundings; their power decreases withincreasing frequency. It is then advantageous to utilizea resonant cell to improve the SNR. The resonancefrequency might be chosen such that contributions fromBrownian noise and amplifier/external noise becomecomparable. At resonance, external acoustical noisewithin the cell will be amplified at a high Q-value.Here acoustical shielding helps, that is a proper cellwall construction (massive), material choice for a highacoustic impedance (such as brass), and good design ofin- and outlet ports are required. For the latter, λ/4 notchfilters (λ the acoustical wavelength) are positioned in linewith the gas in- and outlet (Figure 6). Several types ofmicrophones are used for a photoacoustic resonant cell,but the most used are hearing aid microphones, whichare mass produced, small, and relatively cheap.(37,54)

By definition, noise components do not possess a fixedphase relation with the periodical modulation of the light



intensity. However, external perturbing sources may havea fixed phase relation, for example the directly gener-ated acoustical sound caused by amplitude modulatingthe light beam. They must be minimized in the same wayas the external acoustical noise. A more serious problem isformed by the photoacoustic background signal. Absorp-tion of the light beam in the window material or lightscattered/reflected from the resonator wall causes a signalat the frequency of the modulated light beam. In resonantcells, using λ/4 tubes next to the windows can diminishwindow signals. These tubes, placed perpendicular to theresonator axis close to the windows, are tuned to theresonator frequency and act as notch interference filtersfor the window signals (Figure 6). Influence of scatteredlight on the photoacoustic background signal can be mini-mized by use of highly reflecting, polished resonator wallmaterial with a good thermal conductivity. In the case ofinfrared light, a polished gold-coated copper wall of theacoustical resonator has been found to work satisfactorily.

Another way to diminish the photoacoustic backgroundsignal is to frequency modulate (FM) the laser light overthe molecular absorption line. As the wall/window mate-rial will have broad absorption features, they will hardlycontribute to the 1f of 2f generated signal of the molecularabsorption line.(58,59)

3.6.1 Kinetic Cooling Effect

Using infrared absorption spectroscopy, the absorptionprocess involves vibrational transitions. For photoa-coustic spectroscopy, the increase in vibrational energy is

transferred by collisional relaxation processes to trans-lational energy, causing the gas to eventually be heated,generating a pressure increase. In the case of moleculeswith many low-lying energy states (such as H2O), thisrelaxation process is rapid. When the relaxation processbecomes slow, as compared to the modulation frequencyof the light, the generated photoacoustic signal changesin phase and decreases in amplitude, thereby severelyaffecting the linearity and sensitivity of the photoacousticsignal for trace gas detection.(60–63)

In the atmosphere, N2 has a long vibrational lifetime ofits first excited vibrational mode (≈1 ms at 1 atmosphereat ν1 = 2200 cm−1). When atmospheric CO2 is detectedusing a CO2 laser in the mid-infrared wavelength region(9–11 μm), CO2 will absorb via a hot-band absorptionand is excited to the ν3-vibrational level (Figure 14). Thislevel is almost resonant with the first vibrational level ofN2. Since N2 is much more abundant than CO2, all thevibrational energy will go into N2. Under atmosphericconditions the absorbed energy by CO2 will only slowlybe converted into a temperature and pressure increase ofthe gas. Using kilohertz modulation of the laser beam, aphase shifts occurs in the generated photoacoustic signal.

N2 is not unique; similar effects can be observed withoxygen and CH4.(64) Oxygen has a near-resonant vibra-tional level (1554 cm−1) with the lowest vibrational levelof methane; only 170 collisions are needed to transferenergy from methane to the vibrational mode of O2.(65)

The average number of collisions required for O2 to relaxto the ground state is 8.3× 107. The much larger numberdensity as compared to CH4 creates a buffer of vibrational

CO2(ν1) CO2(ν2) CO2(ν3) H2O(ν2)N2 O2

33803353

2880

2295224119971920

960

Energ

y (

K)

Equilib.

τ6 = 7.5 × 10−7 s

τ13 = 1.9 × 10−4 s

τ4 = 6.3 × 10−7 s

τ11 = 2.9 × 10−7 s

τ8 = 1.2 × 10−2 s

τ5 = 2.2 × 10−4 s

τ7 = 1.0 × 10−7 s

τ9 = 3.2 × 10−2 s

τ10,eff = 1.4 × 10−2 s

(Flux = 1kw cm−2)

Figure 14 Vibrational energy level diagram for atmospheric species important in the transfer of energy absorbed by CO2 at 10.6 μm.The indicated relaxation times have been determined for the 1962 US Standard Atmosphere at sea level, assuming a relative humidityof 30%. Note that the energy is given in Kelvin, 1 K= 0.695030 cm−1. (Reproduced from Ref. 60. © Optical Society of America, 1971.)



2

1

0 0

40

80

0 2 4 6 8

SF6 (%)

Mic

rophone s

ignal R

50 lin

e (

mV

W−1

)

Mic

rophone s

ignal R

16 lin

e (

mV

W−

1)

R16 line

R50 line

Figure 15 Increase of photoacoustic signal as a consequenceof the absorption of CO2 is shown as a function of the addedSF6 concentration. The effect is shown for two CO2 lines near4.23 μm. (Reproduced with permission from Ref. 66. © Springer-Verlag, 2006.)

energy leading to an effective transient cooling of thetranslational degrees of freedom. Owing to a high photoa-coustic modulation frequency (1010 Hz), we observed thistransient cooling by O2 as an amplitude decrease and aphase change relative to the signal caused by the sameamount of CH4 in N2.

Addition of H2O or SF6 molecules (notoriousrelaxers) to the gas under investigation can reversethe situation. An example is shown with the strongrotational–vibrational absorption band of CO2 at4.23 μm.(66) To overcome the slow vibrational relax-ation of the excited CO2 levels, 4% of SF6 gas was addedto the atmospheric mixture (Figure 15). In this way, it waspossible to detect CO2 down to 7 ppbv using 20 mW ofthe OPO light.

3.6.2 Interference of Other Gasses

In other cases, limitations are caused by the presenceof a complex mixture of gases to be investigated. Gasesspectroscopically interfere if molecular absorption linesare close to each other. The most abundant infraredabsorbing gases are water and carbon dioxide. To over-come such interferences, detection can be performed atreduced pressure, thereby reducing the pressure broad-ening. With the nowadays tunable lasers, one can chosethe appropriate wavelength to minimize the effect ofother gases. An example is given in Figure 16.(67) Whenmeasuring biological samples such as plants, seeds,bacteria, or human breath, attention has to be paidto interfering gases such as water and CO2. To detecthydrogen cyanide (HCN) at low concentration in exhaled

breath, an optimal wavelength has to be found in thewavelength region around 3287 cm−1. To estimate theinterference of HCN absorption lines with water andCO2 absorption lines, the ratio is calculated between theabsorption of 10 ppbv of HCN and the sum of absorptioncoefficients of 1% water and 5% CO2 referring to theHITRAN database(16) (Figure 16a). This ratio helped todetermine the best spectral regions for detecting HCNfrom biological samples. It plays a role of a first stepof an analysis to limit the wavelength choice. The fullanalysis requires consideration of a few other factors,such as interfering gases in a gas mixture, concentrationsof compounds, pressure, absolute intensity of a particularabsorption line, and availability of a laser source. Afew of the spectral regions are plotted in the panels ofFigure 16, where the P8 transition in the ν3 band of HCN,at 3287.247603 cm−1, was found to be the best for ourbiological experiments, because it has less interferencewith CO2.

Gases can also be separated by gas chromatographicmethods, selective trapping inside a cold trap (e.g.water),(50) or by a specific chemical reaction (e.g. removalof CO2 by making KOH to react with K2CO3, and waterwith CaCl2). In most cases, a small amount of the inter-fering compound remains present. To address this issue,a multicomponent analysis is necessary.

Also, the gas under investigation can prevent stableand reliable long-term measurements. To perform aproper sampling, it is important to avoid any metal partsin the gas flow system from the sample to the photoa-coustic detection cell; wall adsorption on metal partssignificantly reduces an initial gas mixing concentra-tion for polar molecules such as NH3, HCN, and H2O.By replacing all gas sampling lines and gas connectorswith Teflon lines and connectors and Teflon valves, anormal response time (as for nonpolar molecules) can beachieved. The photoacoustic cell is normally made out ofmetal, but can be coated with Silcosteel coating to reduceadsorption of polar gases.

4 APPLICATIONS

4.1 Atmospheric Applications

Early attempts were made to measure stack gas emissionsfrom power plants with laser photoacoustics. Owing tothe high amount of NO and NO2 compounds in theemissions, they have a significant contribution in photo-chemical smog formation and acidification of the soil. Toreduce this, NH3 is added in the exhaust gas toward thechimney. Using a voluminous catalyst, NOx reduction(within the stack gas) takes place. In order to checkthe performance and to avoid an excess of ammonia



(a)

0

0.0

H2O, 1%

HCN, 10 ppbv

P14 HCN transitionin the ν3 band

P8 HCN transitionin the ν3 band

P25 HCN transitionin the ν3 bandCO2, 5%

H2O, 1%

HCN, 10 ppbv

CO2, 5%

H2O, 1%

HCN, 10 ppbv

CO2, 5%

3265 3266 3267 3268

Wavenumbers (cm−1) Wavenumbers (cm−1) Wavenumbers (cm−1)

3270 32903289328832873286 3329 3330 3331 3332 3333 333432853269

1

2.0 × 10−7

1.5 × 10−7

1.0 × 10−7

5.0 × 10−8

0.0

2.0 × 10−7

1.5 × 10−7

1.0 × 10−7

5.0 × 10−8

0.0

2.0 × 10−7

1.5 × 10−7

1.0 × 10−7

5.0 × 10−8

2

3

4

5

3200H

CN

/(w

ate

r +

CO

2)

Absorp

tion c

oeffic

ient (1

/cm

)

Absorp

tion c

oeffic

ient (1

/cm

)

3250 3300

Wavenumbers (cm−1)

3350 3400

(b) (c) (d)

Figure 16 (a) Ratio between 10 ppbv of HCN and a sum of 1% of water and 5% of CO2 for a pressure of 200 mbar in the wavelengthrange 3200–3400 cm−1. (b–d) Simulated spectra at a 200 mbar pressure inside the photoacoustic for 1% of water (black-dashed line),5% of CO2 (red-dotted line), and 10 ppbv of HCN (blue solid line). (Reprinted with permission from D. D. Arslanov, M. P. P. Castro,N. A. Creemers, A. H. Neerincx, M. Spunei, J. Mandon, S. M. Cristescu, P. Merkus, and F. J. M. Harren, “Optical parametric oscillator-based photoacoustic detection of hydrogen cyanide for biomedical applications,” Journal of Biomedical Optics 18 (2013). Copyright2013 SPIE.(67))

injection, the ammonia concentration in the chimney ismonitored. For this, a CO2-laser-based photoacousticdetection system was successfully applied for in situmonitoring of ammonia concentrations.(68) Owing tothe difficulty, that is hostile environment halfway upthe chimney (vibrations, temperature fluctuations, etc.),concessions have to be made as to the sensitivity of theapparatus. With a nonresonant photoacoustic cell at125 ∘C, a detection limit of 1 ppmv NH3 was achieved ina multicomponent gas mixture containing 10–15% CO2.

Another example is the mobile system for fieldcampaigns by Sigrist et al. Installed in a small trailer,the system has been operational for years and hasbeen applied for several field campaigns in urban andrural environments.(24,69–71) For example, as part of anatmospheric pollution measurement campaign, theyperformed in situ measurements of NH3, C2H4, andCO2 concentrations at the exit of a freeway tunnel witha time resolution of 1 min over a period of 5 weeks.

Peak concentrations were observed at 600 ppbv (NH3),400 ppbv (C2H4), and 1000 ppmv (CO2). Owing to thesemulticomponent gas mixtures, cross-sensitivities in theabsorption coefficients are induced, and as a result detec-tion limits are higher, as compared to those based onextrapolation from larger quantities of trace gases in aclean buffer gas. A mathematical analysis of the photoa-coustic spectra of multicomponent mixtures is based onthe weighted least squares fit of the measured spectrawith spectra from pure compound spectra. The latterspectra are separately taken from trace gas mixtures inpure nitrogen or air.

Numerous practical applications, in industrial processcontrol and environmental monitoring, require sensi-tive and selective concentration measurement over longtime periods by compact and automatic monitoringinstruments. This became available with the develop-ment of tunable diode, quantum cascade, and othercompact lasers systems, resulting in the development



of compact and reliable photoacoustic instruments.Very compact laser-based TF photoacoustic sensorshave been demonstrated for detection of chemicalspecies, such as NH3, NO, CO2, N2O, CO, CH2O, andC3H6O, with a record detection limit of 50 parts pertrillion for SF6, corresponding to a NNEA value of2.7× 10−10 W cm−1 Hz−1/2.(72) For these publications, see,e.g. Refs. 32, 40, 41, 73, 74 and references therein.

For the detection of NO2, photoacoustic cell functionsat its best in the visible or infrared wavelength regions,where it has a low-lying electronic transition in theblue-violet region around 400 nm, next to its vibrationalabsorption band in the infrared. A sub-ppbv detectionlimit has been achieved with external cavity QCLs at6 μm in combination with photoacoustic detection.(75) Amuch cheaper, smaller, and cost-effective approach is touse blue LEDs(18) or blue diode lasers.(38) The latter isbased on the implementation of low-cost components,i.e. a mass-produced blue diode laser and a standardMEMS microphone, which is commonly built into smart-phones. A photoacoustic cell design was realized bymeans of 3D printing (Figure 17). The detection limitwas determined to be 33 pptv NO2 and the NNEA7.0× 10−10 W cm−1 Hz−1/2.

Ammonia is the third most abundant nitrogencompound in the atmosphere and plays a significantrole in cloud and rainwater chemistry. Typical ammoniasources include livestock, fertilizers, soils, forest fires,humans, animals, oceans, industry, and traffic. The mainemission source is the use of animal manure and fertilizersfor agricultural soils (responsible for more than 80% ofthe total ammonia emission). In Figure 18, a QCL-based

(a)

(b)

(c) (e)

(d)

(c)

Figure 17 3D printed photoacoustic cell for the detection ofNO2 with a blue diode laser and MEMS microphone. The mate-rial is acrylate polymer, composed of 12 different substances. Theresonator tube is of stainless steel. (a) Gas inlet, (b) gas outlet,(c) rotatable Brewster window mounts, (d) socket for mountingvarious detectors, and (e) laser beam. (Reproduced with permis-sion from Ref. 38. © Elsevier, 2017.)

photoacoustic set-up is shown with the QCL and powermeter mounted into the photoacoustic cell. This commer-cial photoacoustic system is able to detect ammonia downto a concentration of 2 ppbv in realistic conditions.(76)

Another laser-based photoacoustic instrument wasdeveloped and used for aircraft measurements of ethenefrom industrial sources near Houston (Figure 19). Thisinstrument provided 20-s measurements with a detectionlimit of ∼0.7 ppbv.(77) In flight, data collected from this

0

20

40

60

80

Vredepeel 24/04/2010100

200 400 600 800 1000 1200 1400

Time (min)(b)(a)

NH

3 c

oncentr

ation (

ppb)

Figure 18 (a) QCL-based photoacoustic set-up with laser and power meter mounted onto the photoacoustic cell. (b) 24-h-longmeasurements in the air near a livestock farm in the Netherlands. (Reproduced with permission from Ref. 76. © LSE Monitors,2018.)



30.0

29.5

29.0

28.5

−96.5 −95.5 −94.5−96.0

20100

−95.0

Ethene (ppbv) Emission (kg h−1)

Wind

direction

Houston Houston

ship channel

Mt. Belvieu

Bayport

Galveston bay

Texas city

Chocolate bayouFreeportSweeney

0.0 2.01.51.00.5

Figure 19 Flight track from 19 September 2006: size- and color coded by the ethene mixing ratios measured by laser photoacousticspectroscopy below 1000 m. The color scale is truncated at 2 ppbv to bring out the smaller plumes in the graph. The white trianglesshow ethene point sources in the 2004 Texas Commission on Environmental Quality (TCEQ) inventory, with the size proportionalto the source strength. (Reprinted with permission from J. A. De Gouw, S. T. L. Hekkert, J. Mellqvist, C. Warneke, E. L. Atlas, F. C.Fehsenfeld, A. Fried, G. J. Frost, F. J. M. Harren, J. S. Holloway, B. Lefer, R. Lueb, J. F. Meagher, D. D. Parrish, M. Patel, L. Pope, D.Richter, C. Rivera, T. B. Ryerson, J. Samuelsson, J. Walega, R. A. Washenfelder, P. Weibring, and X. Zhu, “Airborne measurements ofethene from industrial sources using laser photo-acoustic spectroscopy,” Environmental Science & Technology 43, 2437–2442 (2009).Copyright 2009 American Chemical Society.(77))

instrument were compared with gas chromatographicmeasurements, which agreed within 15% on an average.Ethene fluxes from a chemical complex near Houstoncould be quantified during 10 different flights. Theaverage measured fluxes were one order of magnitudehigher than the regulatory emission inventories indicate.

4.2 Detection of Chemical Warfare Agents andExplosives

Over the past decades, the global rise in terrorism andits threats against public places (markets and airports)has increased the awareness for the development ofsensitive detectors for the early detection of the presenceof chemical warfare agents and explosives, such as Sarinaffecting the nerve system.(78) Such chemicals havea large amount of rotational and vibrational modes

that are populated at room temperature; therefore,these molecules have no spectrally resolved lines,but broadband absorption structures (Figure 20). Forsuch broadband structures, broadly tunable lasers areneeded. CO2 lasers in combination with photoacousticspectroscopy are used for this detection of vapors.(79)

However, these laser sources are only line tunable, andit is difficult to access the strong absorption features.The development of QCLs can fulfill these criteria, asthey have continuous tunability and can be designed forspecific wavelength regions ranging from 3 to 20 μm.(25) Inan external cavity configuration they have a wide wave-length coverage (∼300 cm−1) at high power (>100 mW upto several Watt).(80) However, high sensitivity alone is notsufficient. Such a sensor must have very few false alarmsto avoid unacceptable social and economic disruptions.



7 × 10−3

6

5

4

3

Decadic

absorb

ance (

ppm

-mete

r)−1

2

1

0

9.0 9.5 10.0 10.5

Wavelength (μm)

11.0

13CO2 laser

TEPVX

SarinDMMPDIMPCycloSarin

Spectra:

SomanTabun

11.5

Figure 20 Spectral absorption features of a number of chemical warfare agents in the 9–11.5 μm region, within the lower level thepositions and relative power levels of the 13CO2 laser output on the lasing transitions as a function of the wavelength. (Reproducedwith permission from Ref. 79. © AIP Publishing, 2006.)

The problem of interference rejection is severe because ofthe broad absorption features of many chemical warfareagents. In normal indoor or outdoor environments,hundreds of gases are present, of which a few dozen needspecial consideration because of their ambient concen-trations, absorption magnitudes, and spectral overlap.(79)

4.3 Biological and Agricultural Applications

4.3.1 Plant Physiology

For a long time, plants were known to emit ethylene(C2H4); it is considered a gaseous phytohormone regu-lating various growth and development processes, startingfrom seed germination through growth and flowering,until fruit ripening and plant senescence, synergisti-cally or antagonistically with other hormones.(81,82)

The biosynthesis in plants is well understood, leadingfrom the amino acid methionine over two inter-mediates, S-adenosylmethionine and amino-cyclo-propane-carboxylic acid, to ethylene. Ethylene isproduced by plant tissue as a response to manyexternal factors, such as pathogen attack,(83) herbivorouspredation,(84) and abiotic environmental factors includingflooding,(85) chemical exposure,(86) day length and lightintensity,(87) temperature,(88) or nutrient availability.(89)

Ethylene emissions may vary according to the plantspecies, organ type (e.g. root, leaf, and flower), and

developmental stage of the plant. It is demonstrated thatethylene can stimulate fruit ripening even at low ppbvlevels.(90) Therefore, it is of paramount importance tomonitor and control the ethylene emission in growthchambers, greenhouses, and storage facilities to optimizefruit freshness. Controlled atmosphere was developed toallow longer storage periods and, subsequently, to providea wide variety of fruits to the consumers over the entireyear. Apart from the agricultural- and industrial-relatedinvolvement of ethylene in controlling fruit quality, a lotof effort has been dedicated to understanding the physi-ological mechanism of ethylene biosynthesis, action, andperception in plants. In spite of great achievements atboth the plant’s molecular and physiological levels, manyethylene-related plant events remain unrevealed, mainlydue to a lack of fast and sensitive ethylene detection.

Hence, there is an increasing demand for simple, afford-able, and reliable ethylene sensors that could be usedin conjunction with equipment for controlling ethyleneconcentrations. With laser-based photoacoustic detec-tion, the dynamics of many processes that take placein plants become easily visible. For example, ethylenereleased by Arabidopsis displays a circadian rhythm (highemission during the light and low emission during dark)with a peak in the mid-subjective day. When the plantswere grown for 6 days in a succession of light (16 h) anddark (8 h) and afterward were kept in continuous light for



0.0

0.4

0.8

1.2

5432

LightLight dark

Eth

yle

ne (

ppb)

Days

1

Figure 21 Influence of day length on ethylene production inArabidopsis. Seedlings were grown for 6 days in 16 h light and8 h dark and then tested for ethylene emission during onelight–dark cycle followed by constant light. Ethylene releasedby Arabidopsis displays a circadian rhythm with a peak in themid-subjective day.

several days, they showed an ethylene pattern as wouldfollow the light–dark periods (Figure 21).(87)

One of the major environmental factors influencingcrop productivity is flooding. Tolerant species oftenpossess ‘escape’ mechanisms such as fast underwatergrowth that minimizes the duration of total submergence.However, at early vegetative stages, even in ‘deep-water’types of rice, such escape is precluded by a lack ofsufficiently vigorous shoot elongation for resurfacing.Survival of submergence under these circumstancesdepends on intrinsic physiological and biochemical toler-ance. Screening of cultivars is needed by plant breeders,agrotechnology, and biotechnology to achieve sustainableimprovements in the productivity of rice farming in therain-fed lowland of Southeast Asia.(91) Submergence isnot an easily quantified stress compared with, for examplecold, heat, and pollution, since it cannot be assessed byone simple analysis. This is because it involves exclusionand entrapment of several key gases and interference withillumination. Normal rates of respiration and photosyn-thesis are affected; differences in submergence toleranceare reflected also in a different reaction to low oxygenconcentrations. Precise measurements of fermentationrates can result in a powerful tool to shed some lighton submergence stress. Fermentation in plant tissuewas studied by the use of photoacoustic techniques,through detection of ethanol and acetaldehyde. It wasshown that young rice seedlings of various species showdifferent degrees of tolerance toward submergence,which is expressed by different fermentation rates underanaerobic and submergence conditions.(92)

4.3.2 Postharvest Physiology

It is estimated that globally 33% of the edible partsof food produced for human consumption gets lost orwasted.(93) Worldwide, 1.6 billion tons of fruit and vegeta-bles are produced, from which 120 million tons are lostduring the postharvest process, i.e. losses due to spoilageand degradation during handling, storage, and transporta-tion between farm and distribution. Storage involvesstrict protocols with respect to timing of harvest and thegradual establishment and maintenance of the storageconditions. Often it involves storage at low temperature,decreased oxygen, and increased carbon dioxide concen-trations. During or after storage, physiological or patho-logical disorders may develop, which can greatly diminishthe commercial value of the fruits, owing to spoilage orquality classification downgrading. The optimization ofthe storage conditions and timing in order to maximallypreserve the quality is a challenge, owing to restrictedaccess to the products for quality inspection under suchstrict regimes.(94) In addition, to enable decision-making,it is important to have relevant information about thestatus of the stored products. Information about the actualripening stage and the possible occurrence of fungal infec-tions may influence the decisions about the best momentin time to open stores and sell the fruit.

The respiration quotient (ratio between CO2 producedand O2 consumed) indicating aerobic or anaerobic(fermentation) metabolism is an important parameterthat can serve for adjusting the oxygen concentration inthe storage room. Other important volatiles released bystored products include ethylene (marker for ripening),ethane (indicator of damage),(95) ethanol, acetaldehyde,and ethyl acetate (markers for fermentation), whilemethanol and acetone may serve as rotting processindicators. More specifically, the development of fungalinfections (potentially dangerous for the consumer’shealth via mycotoxins) during storage may be detectedby rotting volatiles in combination with low levels offermentation volatiles. Therefore, the continuous moni-toring of such volatiles is important not only for foodquality but also for food safety control. The noninva-sive nature of laser-based trace gas detection renders itsuitable for applications in this field of postharvest phys-iology. Moreover, the high sensitivity combined with fastanalysis allows one to study an individual piece of fruit sothat metabolic activity may be followed in real time.(96)

4.3.3 Microbiology

The growth of all organisms depends on the availabilityof mineral nutrients, and nitrogen is one of the mostrequired in many biological reactions. Although theearth’s atmosphere contains nearly 80% of nitrogen



gas, this is unavailable for use by most organisms, beingan inert molecule. In order for nitrogen to be used forgrowth, it must be ‘fixed’ (combined) in the form ofammonium or nitrate ions. This conversion process isknown as nitrogen fixation and is exclusively performedby prokaryotes (bacteria) using an enzyme complexcalled nitrogenase.(97) Acetylene reduction is the mostwidely used method for measuring the nitrogenaseactivity.(98) It is based on the nitrogenase property toreduce compounds with a triple bond. Thus, by reducingnitrogen to ammonia, the nitrogenase reduces acetyleneto ethylene. Both these gases can be measured with highsensitivity by gas chromatography with flame ioniza-tion detection. However, owing to its higher sensitivityand fast time response (about 20 s), laser photoacousticspectroscopy offers the advantage of monitoring smallchanges in the nitrogenase activity of low amounts ofbiomass in real time.(99) The on-line monitoring allowsmeasurements under constant conditions, which elimi-nates artifacts caused by changes in the concentrationsof O2 or CO2 due to respiration or photosynthesis duringincubation. Changes in light intensity and temperatureinduce fast response in the nitrogenase activity. Knowingthe level of saturation of the enzyme at different tempera-tures and light is important when rates of acetylene reduc-tion are converted to rates of nitrogen fixation (Figure 22).

4.3.4 Entomology

Methane has been recognized as one of the principalgreenhouse gases, next to CO2. Its estimated contribu-tion to the enhanced greenhouse effect varies between15% and 20%, depending on the time window for whichthe calculation is made. Natural sources are responsiblefor the major part of the global methane flux.(100) Some

arthropods (i.e. millipedes, cockroaches, termites, andscarab beetles) are believed to contribute up to 25%of the total budget.(101) Gas chromatography allowsmeasurement of methane release of single animals onlyafter incubation for several hours. Consequently, thesemeasurements cannot provide information about thedynamic character of the gas emission. Infrared absorp-tion measurements featuring fast response time permittedobservation of the dynamics of CO2 emissions. BesidesCH4 and CO2 emissions from insects, water vapor releasehas been the subject of many studies. Information aboutwater loss dynamics in insects is based on theoreticalmodeling, sensitive weighing, and relatively slow directwater vapor measurements.(102) Using laser photoacousticdetection, cockroaches at rest show a regular breathingpattern to optimize oxygen uptake and reduce waterloss. Respiration patterns of cockroaches and beetleshave been recorded in which microbiologically producedmethane was found to be co-emitted with CO2 and watervapor (Figure 23).

Combining an OPO with photoacoustic spectroscopy,CO2 emissions even from small insects could bedetected. For this the OPO was made tunable in the3.9–4.8 μm infrared wavelength region, covering thestrong rotational–vibrational absorption band of CO2 at4.23 μm. With this system, it was possible to detect CO2with a sensitivity of 7 ppbv using 20 mW OPO power.They demonstrated in real time the CO2 emission inexhaled air of a single ant (Lasius niger) and individualfruit flies (Drosophila melanogaster, see Figure 24).(66)

4.4 Human Health, Noninvasive Breath Analysis

Humans exhale hundreds of gases, including inorganiccompounds (e.g. NH3, NO, and H2S) and volatile organic

0 2 4 6

8

12

16

20

Eth

yle

ne (

ppbv)

Time (h)

10 °C

15 °C

20 °C

25 °C

30 °C

35 °C 37 °C

Figure 22 Response of cyanobacteria N. spumigena (see photo) to change of temperature from 10 to 37 ∘C under light conditions.Significant instability in the nitrogenase activity is observed at 35 ∘C, leading to a short inhibition tendency at 37 ∘C, followed by adrastic decrease.



1.0

0.5

0.0

75

50

25

0

Time (h)

CH4 (μmol h−1)

CO2 (μmol h−1)

5.2 5.6 6.0 6.4

Figure 23 Methane release together with the CO2 emission during a complete respiration sequence of breathing (starting at 5.15and 6.08 h) and constriction (starting at 5.48 and 6.32 h) for the cockroach species Gromphadorhina portentosa. A downshift inbreathing frequency can be observed at 5.36 h; incidental releases can be observed at 5.6–5.7 h. (Reproduced from Ref. 50. © OpticalSociety of America, 1996.)

compounds (VOCs, e.g. acetone), arising from theirnormal body metabolism.(103–105) Human breath alsocontains exogenous compounds that originate from foodand beverages or from current or previous environ-mental exposures.(106) Since ancient times some diseaseshave been recognized by their associated odor(107):uncontrolled diabetes produces a sweet, fruity odor;advanced liver diseases entail a musty, fishy reek; failingkidneys bring about an urinelike smell; and a lungabscess can be brought to light by its putrid stench. AfterPauling’s breakthrough in 1971,(108) modern breath anal-ysis has connected these odors with specific gas analytes(biomarkers) or a profile of VOCs. Although manyexhaled compounds potentially reflect the physiologicaland pathophysiological conditions related to variousdiseases, only some VOCs have been established asbiomarkers.(109) Several methods are currently employedfor exhaled volatiles detection, among which massspectrometry is the most widely used analytical tool.(110)

Next to analytical tools, there is a strong need forminiaturized devices, which are affordable, accurate, anduser-friendly, providing a fast response, preferably in realtime. This is where the strength is of optical methods:specific, reliable, quantitative real-time analysis of singleanalytes, and without the need for frequent calibration.

A list of representative molecules of medical interestdetermined with laser-based spectroscopic methods islisted elsewhere.(110) Depending on the specific experi-mental set-up, laser-based measurements can be done inreal time (subsecond time resolution and resolving indi-vidual breath cycles) or near real time.

Owing to its high sensitivity, laser photoacoustics allowssingle breath collection from a small sampling volume(few hundred milliliters) with no preconcentration stepsneeded. An early example of using laser photoacousticsfor breath analysis is the investigation of the effect ofUV radiation on the human skin. Under stress conditions(e.g. ionizing radiation, toxic chemical substances, anddiseases), the production of free radicals in the body issignificantly increased. Subsequently, the capacity of thefree radical scavengers in the body is overloaded, and achain of chemical reactions is activated.(111) This leadsultimately to cell membrane damage (i.e. lipid peroxi-dation and oxidation of fatty acids in cell membrane)that plays an important role in the aging processesand pathogenesis of some diseases. The cell damage isaccompanied by the production of small hydrocarbonssuch as ethane, pentane, and ethylene that can be easilymeasured, also with photoacoustics.(112,113) This investi-gation was recently extended to a clinical setting, with



30 6

420

10

10

5

0

0

0

5 10 15 20 25

Time (min)

CO

2 p

roduction (

μl h

−1)

CO

2 p

roduction (

μl h

−1)

CO

2 c

oncentr

ation (

ppm

v)

2

0

4

CO

2 c

oncentr

ation (

ppm

v)

2

00

4 8 12 16 20

Time (min)

CO2 emission ant (Lasius niger)

CO2 emission

Drosophila fruit fly

Empty cuvette

(a)

(b)

Figure 24 (a) Real-time CO2 emission of a small ant (Lasiusniger, weight 3.7 mg). Analysis of the ants’ respiration showsa periodic release of CO2, on an average every 90 s. (b) CO2release of a fruit fly (Drosophila melanogaster, weight 1.4 mg).The comparison of the CO2 trace from the fruit fly with thebackground (without fly) shows that the CO2 release does notdrop to zero. The accuracy of the CO2 detection shows that theCO2 peaks vary in amplitude and frequency. (Reproduced withpermission from Ref. 66. © Springer-Verlag, 2006.)

the assessment of lipid peroxidation by expired ethyleneduring cardiac surgery.(114) Patients were undergoingaortic or mitral valve surgery requiring cardiopulmonarybypass and off-pump coronary artery bypass surgery. Thestudy revealed for the first time the real-time signaturesof the surgical response triggered by oxidative stressincluding lipid peroxidation (Figure 25). This innovativeapproach has demonstrated the significant contributionof diathermy to overall oxidative stress and identifiedincreased ethylene associated with regional myocardialischemia on an individual basis. It can be concluded thatsensitive, totally noninvasive, and real-time analysis oflipid peroxidation using breath ethylene is feasible in theclinical and perioperative setting.(115)

40

30

20

Eth

yle

ne (

ppbv)

Eth

yle

ne (

ppbv)

10

0

0 60 120 180 240

00

2

4

6

8

10

12

14

60 80 1201004020

Time (min)

Time (min)

(a)

(b)

DT on

DT

DT

Figure 25 Intraoperative, real-time monitoring of ethylenemeasured in patients undergoing off-pump coronary arterybypass surgery. (a) Example of continuous monitoring duringthe entire operation. At t= 0 h, the patient was connected tothe sampling line; DT, diathermy electrocautery with high-frequency electric currents. (b) Example of a grafting procedure.The high-resolution peaks are associated with ethylene inducedby lipid peroxidation during the reperfusion events. (Repro-duced with permission from Ref. 114. © American PhysiologicalSociety, 2014.)

Cystic fibrosis (CF) is a multisystem disease, leading tochronic respiratory infections that result in progressiveloss of lung function and diminished life expectancy.Pseudomonas aeruginosa is one of the most commonpathogens causing these infections. The presenceof this opportunistic bacterium is associated with ahigh morbidity and mortality rate.(116,117) Initially, thebacterium can often be eradicated with aggressive antibi-otic treatment. However, over time the phenotypechanges, complicating eradication. To detect and monitor



P. aeruginosa, it is needed to screen patients frequentlyfor colonization, and to start with aggressive eradicationtherapy. The current ‘golden standard’ for detection ofP. aeruginosa is a sputum culture. However, young CFpatients are not able to expectorate adequate samples.Alternatives have therefore been sought and includeinvasive bronchoscopy, nonspecific chest X-ray, or unre-liable cough swabs methods.(117) For this reason, thereis a need for a quick, patient friendly, reliable, safe, andsensitive method to diagnose P. aeruginosa infection inan early stage. The ability of P. aeruginosa to synthesizeHCN can be the base of this diagnosis. Cyanogenesis orHCN synthesis occurs only in a few bacterial species,including P. aeruginosa.

Mid-infrared OPOs are considered among the mostuseful tools for sensitive laser photoacoustic gas sensing,owing to their high-power (Watt-level), wide tuningrange, and ease of tunability. Here, such a light source wassuccessfully applied for the detection of HCN in exhaledbreath of CF patients.(67,117,118) In high concentrations,HCN is toxic, but it may also originate endogenously (insmall amounts), from pathogens or ingestion of food.(67)

5 COMPARISON WITH OTHERSPECTROSCOPIC METHODS IN TRACEGAS MONITORING

Fast and sensitive trace gas detection is not exclusivelyreserved for photoacoustics. In general, all spectroscopictechniques are able to monitor gas absorptions on asubsecond timescale. However, depending on the type ofapplication not all techniques possess a high sensitivity.The sensitivity of a gas absorption measurement can beenhanced in two ways: either by increasing the absorptionpath length or by increasing the SNR.

Using direct absorption spectroscopy over shortdistances the most abundant species can only be detected(such as CO2 and water vapor). This can be improvedusing multipass cells, thereby increasing the path length toseveral tenths of meters, achieving sub-parts-per-milliongas concentrations levels. Direct absorption will have

minimal detectable absorption (MDA) of ∼10−2. MDA isdefined as:

MDA =αmin√

BHz−1∕2 (7)

with B the detection bandwidth (Hz), with the minimumabsorption, αmin =ΔI/I, equivalent to the ratio of laserintensity noise ΔI and laser intensity I.(57) The detec-tion bandwidth is important, as it normalizes the integra-tion time over the measurement. Wavelength modulationspectroscopy (WMS) improves the MDA to 10−5 to 10−6,because of the high modulation frequencies, reducing thenoise on the intensity detection.

Trace gases in the open air can be very well studiedalong long light paths, rendering low detection limits.To normalize and to be able to compare methods overpath lengths, the noise equivalent absorption sensitivity(NEAS) is defined:

NEAS = MDAL

=αmin

L√

Bcm−1 ⋅ Hz−1

/2 (8)

Cavity ring-down spectroscopy and integrated cavityoutput spectroscopy make use of a high finesse cavity tohave a very long path length (up to kilometers) in a cavitywith a small footprint, thanks to the use of high reflectivemirrors and sensitive infrared detectors. Thanks to thelong path length the NEAS can reach, 10−10 to 10−11,although the MDA stays at the level of 10−6 (Table 1).Laser-based photoacoustic detection has the advantagethat it is background free: it is not relying on a decreaseof the transmitted light (and thus the intensity noise),but on an increase from a zero baseline, i.e. on a colli-sional release of energy after absorption. As such theMDA will already be in the 10−8 Hz−1/2 region. At smallabsorptions, the photoacoustic signal is linear with thelaser power (Equation 1); to compare photoacousticsensitivities, the MDA is normalized on the laser powerinstead of the path length: the NNEA with typical valuesof 10−10 W cm−1 Hz−1/2. Using a laser power of 1 W andhaving an absorption path length in the photoacoustic cellof 10 cm, the NEAS is at the 10−9 cm−1 Hz−1/2. Anotherbackground-free and therefore strongly competitive

Table 1 Comparison between several spectroscopic methods and photoacoustic spectroscopy

αmin L (cm) t (s) MDA (Hz−1/2) NEAS (cm−1 Hz−1/2)

Direct absorption 4× 10−2 5 000 10 10−2 2× 10−6

Wavelength modulation 2× 10−4 5 000 10 5× 10−5 10−8

Cavity-enhanced methods: CRD/ICOS 1× 10−5 50 000 10 2× 10−6 5× 10−11

Photoacoustic (PA) (1 W power) 4× 10−8 10 10 10−8 10−9

The minimal absorption αmin is normalized on the bandwidth (B), which has the relation with the integration time t= 1/2B, using the Nyquist criterion.Measurement time and path length (L) are typical laboratory values to achieve αmin. Photoacoustic values are converted to NEAS using 1 W of laserpower (see text).



technique is LIF. LIF is extremely sensitive owing to thehigh collection efficiency of photomultipliers; they can beapplied locally within a small gas volume. LIF and laserphotoacoustics are complementary in energy release. Theadvantage of LIF is that it can also detect free radicals(e.g. OH) with relatively low laser powers. A restrictionof LIF is the lifetime of the excited levels. As stated inSection 1, the competition between collisional relaxationand radiative decay restricts LIF mainly to the visible andthe UV parts of the electromagnetic spectrum.

ACKNOWLEDGMENTS

The authors wish to thank J. Reuss for the helpful discus-sions. Also, the authors thank F. Bijnen, H. de Vries, S.Persijn, T. Groot, E. Santosa, I. Boamfa, M van Herpen,B. Moeskops, M. Steeghs, A. Ngai, E. Crespo, D. Arslanov,D. Marchenko, D. Samudrala, R. Centeno, Y. Jin, A. Neer-incx, S. te Lintel Hekkert, L.J. Laarhoven, P. Brown, andJ. Mandon for their experimental support and C. Sikkensfor technical assistance. Finally, they would like to thankthe Dutch Science foundation, Dutch Technology Foun-dation, Dutch Royal Academy of Science, Province ofGelderland, INTERREG North-West Europe, and theEuropean Union for their continuous financial supportvia various programs.

ABBREVIATIONS AND ACRONYMS

CF Cystic Fibrosiscw Continuous WaveFM Frequency ModulationFTIR Fourier Transformed InfraredFWHM Full Width at Half MaximumHCN Hydrogen CyanideICL Interband Cascade LaserLED Light-emitting DiodeLIF Laser-induced FluorescenceMDA Minimal Detectable AbsorptionMEMS Micro-electro-mechanical SystemmR MicroresonatorNEAS Noise Equivalent Absorption SensitivityNNEA Normalized Noise Equivalent AbsorptionOPO Optical Parametric OscillatorPPLN Periodically Poled Lithium NiobateQCL Quantum Cascade LaserSNR Signal-to-noise RatioTF Tuning ForkVOC Volatile Organic CompoundWMS Wavelength Modulation Spectroscopy

RELATED ARTICLES

Biomolecules AnalysisInfrared Spectroscopy of Biological Applications

Chemical Weapons Chemicals AnalysisFourier Transform Infrared in On-site and Off-site Anal-ysis of Chemicals Related to the Chemical WeaponsConvention

Clinical ChemistryInfrared Spectroscopy in Clinical Chemistry

Environment: Trace Gas MonitoringEnvironmental Trace Species Monitoring: Introduction •Diode Laser Spectroscopic Monitoring of Trace Gases •Optical Gas Sensors in Analytical Chemistry: Applica-tions, Trends and General Comments • Fourier Trans-form Infrared Spectrometry in Atmospheric and TraceGas Analysis • Laser- and Optical-Based Techniques forthe Detection of Explosives

Environment: Water and WasteDetection and Quantification of Environmental Pollu-tants

Infrared SpectroscopyCavity Ringdown Laser Absorption Spectroscopy •Instrumentation and Applications of Cascade LaserSpectroscopy

REFERENCES

1. A.G. Bell, ‘On the Production and Reproduction ofSound by Light’, Am. J. Sci., 20, 305–324 (1880).

2. A.G. Bell, ‘Upon The Production of Sound by RadiantEnergy’, London Edinburgh Dublin Philos. Mag. J. Sci.,11, 510–528 (1881).

3. W.C. Röntgen, ‘Über Töne, welche durch intermittirendeBestrahlung eines Gases entstehen’, Ann. Phys., 248,155–159 (1881).

4. J. Tyndall, ‘Action of an Intermittent Beam of RadiantHeat Upon Gaseous Matter’, Science (New York, N.Y.),2, 110–114 (1881).

5. W.H. Preece, ‘I. On the Conversion of Radiant Energyinto Sonorous Vibrations’, Proc. R. Soc. Lond., 31,506–520 (1881).

6. M.L. Viegerov, ‘Eine Methode der gasanalyse, beruhendauf der optisch-akustischen Tyndall-Röntgen erschei-nung’, Comptes Rendus (Doklady) de l’Académie desSciences de l’URSS, 19, 687–688 (1938).

7. A.H. Pfund, ‘Atmospheric Contamination’, Science (NewYork, N.Y.), 90, 326–327 (1939).



8. K.F. Luft, ‘Über eine neue Methode der registrierendenGasanalyse mit Hilfe der Absorption ultraroter Strahlenohne spektrale Zerlegung’, Z. Tech. Phys., 24, 97–104(1943).

9. E.L. Kerr, J.G. Atwood, ‘The Laser Illuminated Absorp-tivity Spectrophone: A Method for Measurement of WeakAbsorptivity in Gases at Laser Wavelengths’, Appl. Opt.,7, 915–921 (1968).

10. L.B. Kreuzer, ‘Ultralow gas Concentration InfraredAbsorption Spectroscopy’, J. Appl. Phys., 42, 2934 (1971).

11. C.K.N. Patel, E.G. Burkhardt, C.A. Lambert, ‘Spectro-scopic Measurements of Stratospheric Nitric Oxide andWater Vapor’, Science, 184, 1173–1176 (1974).

12. K.F. Luft, W. Schaefer, G. Wiegleb, ‘50 Years NDIR Gas-Analysis’, Tech. Mess., 60, 363–371 (1993).

13. L. Liu, A. Mandelis, H. Huan, K.H. Michaelian,‘Step-Scan Differential Fourier Transform InfraredPhotoacoustic Spectroscopy (DFTIR-PAS): A SpectralDeconvolution Method for Weak Absorber Detectionin the Presence of Strongly Overlapping BackgroundAbsorptions’, Opt. Lett., 42, 1424–1427 (2017).

14. R.M. Dittmar, J.L. Chao, R.A. Palmer, ‘PhotoacousticDepth Profiling of Polymer Laminates by Step-ScanFourier-Transform Infrared-Spectroscopy’, Appl. Spec-trosc., 45, 1104–1110 (1991).

15. T. Kuusela, J. Peura, B.A. Matveev, M.A. Remennyy,N.M. Stus, ‘Photoacoustic Gas Detection Using aCantilever Microphone and III-V Mid-IR LEDs’, Vib.Spectrosc., 51, 289–293 (2009).

16. Harvard-Smithsonian Center for Astrophysics, ‘TheHITRAN Database’, http://hitran.org/home

17. S. Bottger, M. Kohring, U. Willer, W. Schade, ‘Off-BeamQuartz-Enhanced Photoacoustic Spectroscopy withLEDs’, Appl. Phys. B Lasers Optics, 113, 227–232 (2013).

18. H.D. Zheng, L. Dong, X.K. Yin, X.L. Liu, H.P. Wu,L. Zhang, W.G. Ma, W.B. Yin, S.T. Jia, ‘Ppb-Level QEPASNO2 Sensor by Use of Electrical Modulation CancellationMethod with a High Power Blue LED’, Sens. ActuatorsB Chem., 208, 173–179 (2015).

19. L.B. Kreuzer, N.D. Kenyon, C.K.N. Patel, ‘Air-Pollution- Sensitive Detection of 10 Pollutant Gases by Carbon-Monoxide and Carbon-Dioxide Lasers’, Science, 177, 347(1972).

20. F. Harren, F. Bijnen, J. Reuss, L. Voesenek, C. Blom,‘Sensitive Intracavity Photoacoustic Measurements with aCO2 Waveguide Laser’, Appl. Phys. B, 50, 137–144 (1990).

21. S. Bernegger, M.W. Sigrist, ‘CO-Laser Photoacoustic-Spectroscopy of Gases and Vapors for TraceGas-Analysis’, Infrared Phys. Technol., 30, 375–429(1990).

22. F.G.C. Bijnen, H. Zuckermann, F.J.M. Harren, J. Reuss,‘Multicomponent Trace-Gas Analysis by Three Intra-cavity Photoacoustic Cells in a CO Laser: Observation

of Anaerobic and Postanaerobic Emission of Acetalde-hyde and Ethanol in Cherry Tomatoes’, Appl. Opt., 37,3345–3353 (1998).

23. S.T. Persijn, R.H. Veltman, J. Oomens, F.J.M. Harren,D.H. Parker, ‘CO Laser Absorption Coefficients forGases of Biological Relevance: H2O, CO2, Ethanol,Acetaldehyde, and Ethylene’, Appl. Spectrosc., 54, 62–71(2000).

24. M. Nagele, M.W. Sigrist, ‘Mobile Laser Spectrometer withNovel Resonant Multipass Photoacoustic Cell for Trace-Gas Sensing’, Appl. Phys. B, 70, 895–901 (2000).

25. R.F. Curl, F. Capasso, C. Gmachl, A.A. Kosterev,B. McManus, R. Lewicki, M. Pusharsky, G. Wysocki, F.K.Tittel, ‘Quantum Cascade Lasers in Chemical Physics’,Chem. Phys. Lett., 487, 1–18 (2010).

26. I. Vurgaftman, R. Weih, M. Kamp, J.R. Meyer, C.L.Canedy, C.S. Kim, M. Kim, W.W. Bewley, C.D. Merritt,J. Abell, S. Hoefling, ‘Interband Cascade Lasers’, J. Phys.D Appl. Phys., 48, 123001 (2015).

27. L.E. Myers, W.R. Bosenberg, ‘Periodically Poled LithiumNiobate and Quasi-Phase-Matched Optical ParametricOscillators’, IEEE J. Quantum Electron., 33, 1663–1672(1997).

28. D.L. Howard, H.G. Kjaergaard, ‘Influence of Intramolec-ular Hydrogen Bond Strength on OH-Stretching Over-tones’, J. Phys. Chem. A, 110, 10245–10250 (2006).

29. K. Liu, W.X. Zhao, L. Wang, T. Tan, G.S. Wang, W.J.Zhang, X.M. Gao, W.D. Chen, ‘Quartz-EnhancedPhotoacoustic Spectroscopy of HCN from 6433 to6613 cm−1’, Opt. Commun., 340, 126–130 (2015).

30. F.G.C. Bijnen, J. Reuss, F.J.M. Harren, ‘Geometrical Opti-mization of a Longitudinal Resonant Photoacoustic Cellfor Sensitive and Fast Trace Gas Detection’, Rev. Sci.Instrum., 67, 2914–2923 (1996).

31. S. Bernegger, M.W. Sigrist, ‘Longitudinal Resonant Spec-trophone for CO-Laser Photoacoustic-Spectroscopy’,Appl. Phys. B, 44, 125–132 (1987).

32. A. Miklos, P. Hess, Z. Bozoki, ‘Application of AcousticResonators in Photoacoustic Trace Gas Analysis andMetrology’, Rev. Sci. Instrum., 72, 1937–1955 (2001).

33. M. Suchenek, ‘Influence of the Duct Shape on the Q-Factor of a Photoacoustic Helmholtz Resonator’, Int.J. Thermophys., 32, 886–892 (2011).

34. V. Zeninari, V.A. Kapitanov, D. Courtois, Y.N. Pono-marev, ‘Design and Characteristics of a DifferentialHelmholtz Resonant Photoacoustic Cell for Infrared GasDetection’, Infrared Phys. Technol., 40, 1–23 (1999).

35. O. Nordhaus, J. Pelzl, ‘Frequency-Dependence of Reso-nant Photoacoustic Cells – The Extended HelmholtzResonator’, Appl. Phys., 25, 221–229 (1981).

36. R. Kastle, M.W. Sigrist, ‘Temperature-DependentPhotoacoustic Spectroscopy with a HelmholtzResonator’, Appl. Phys. B, 63, 389–397 (1996).



37. S.C. Thompson, J.L. LoPresti, E.M. Ring, H.G. Nepo-muceno, J.J. Beard, W.J. Ballad, E.V. Carlson, ‘Noisein Miniature Microphones’, J. Acoust. Soc. Am., 111,861–866 (2002).

38. T. Rueck, R. Bierl, F.-M. Matysik, ‘Low-Cost Photoa-coustic NO2 Trace gas Monitoring at the pptv-Level’,Sens. Actuators A Phys., 266, 369 (2017).

39. D.V. Serebryakov, A.P. Cherkun, B.A. Loginov, V.S.Letokhov, ‘Tuning-Fork-Based Fast Highly Sensi-tive Surface-Contact Sensor for Atomic ForceMicroscopy/Near-Field Scanning Optical Microscopy’,Rev. Sci. Instrum., 73, 1795–1802 (2002).

40. A.A. Kosterev, F.K. Tittel, D.V. Serebryakov, A.L. Mali-novsky, I.V. Morozov, ‘Applications of Quartz TuningForks in Spectroscopic Gas Sensing’, Rev. Sci. Instrum.,76, 9 (2005).

41. P. Patimisco, G. Scamarcio, F.K. Tittel, V. Spagnolo,‘Quartz-Enhanced Photoacoustic Spectroscopy: AReview’, Sensors, 14, 6165–6206 (2014).

42. K. Liu, X.Y. Guo, H.M. Yi, W.D. Chen, W.J. Zhang, X.M.Gao, ‘Off-Beam Quartz-Enhanced Photoacoustic Spec-troscopy’, Opt. Lett., 34, 1594–1596 (2009).

43. V. Koskinen, J. Fonsen, J. Kauppinen, I. Kauppinen,‘Extremely Sensitive Trace Gas Analysis with ModernPhotoacoustic Spectroscopy’, Vib. Spectrosc., 42, 239–242(2006).

44. K. Chen, Q.X. Yu, Z.F. Gong, M. Guo, C. Qu, ‘Ultra-HighSensitive Fiber-Optic Fabry-Perot Cantilever EnhancedResonant Photoacoustic Spectroscopy’, Sens. ActuatorsB Chem., 268, 205–209 (2018).

45. R. Gerlach, N.M. Amer, ‘Brewster Window and Window-less Resonant Spectrophones for Intra-Cavity Operation’,Appl. Phys., 23, 319–326 (1980).

46. S. Tranchart, I.H. Bachir, T.R. Huet, A. Olafsson,J.L. Destombes, S. Naim, A. Fayt, ‘High-ResolutionLaser Photoacoustic Spectroscopy of OCS in the12000–13000 cm−1 Region’, J. Mol. Spectrosc., 196,265–273 (1999).

47. T. Starecki, A. Geras, ‘Improved Open PhotoacousticHelmholtz Cell’, Int. J. Thermophys., 35, 2023–2031(2014).

48. A. Thony, M.W. Sigrist, ‘Novel Photoacoustic Stark Cell’,Rev. Sci. Instrum., 66, 227–231 (1995).

49. M. Seaver, T.J. Manuccia, J.R. McDonald, ‘EnhancedSpecies Selectivity in Optoacoustic Detection’, Chem.Phys. Lett., 123, 164–168 (1986).

50. F.G.C. Bijnen, F.J.M. Harren, J.H.P. Hackstein, J. Reuss,‘Intracavity CO Laser Photoacoustic Trace gas Detection:Cyclic CH4, H2O and CO2 Emission by Cockroaches andScarab Beetles’, Appl. Opt., 35, 5357–5368 (1996).

51. Z. Bozoki, J. Sneider, G. Szabo, A. Miklos, M. Serenyi,G. Nagy, M. Feher, ‘Intracavity Photoacoustic Gas Detec-tion with an External Cavity Diode Laser’, Appl. Phys.B Lasers Opt., 63, 399–401 (1996).

52. Q. Wang, Z. Wang, J. Chang, W. Ren, ‘Fiber-Ring Laser-Based Intracavity Photoacoustic Spectroscopy for TraceGas Sensing’, Opt. Lett., 42, 2114–2117 (2017).

53. S. Borri, P. Patimisco, I. Galli, D. Mazzotti, G. Gius-fredi, N. Akikusa, M. Yamanishi, G. Scamarcio, P. DeNatale, V. Spagnolo, ‘Intracavity Quartz-EnhancedPhotoacoustic Sensor’, Appl. Phys. Lett., 104, 4 (2014).

54. A. Kachanov, S. Koulikov, F.K. Tittel, ‘Cavity-EnhancedOptical Feedback-Assisted Photo-Acoustic Spectroscopywith a 10.4 μm External Cavity Quantum Cascade Laser’,Appl. Phys. B, 110, 47–56 (2013).

55. D.D. Arslanov, M. Spunei, J. Mandon, S.M. Cristescu, S.T.Persijn, F.J.M. Harren, ‘Continuous-Wave Optical Para-metric Oscillator Based Infrared Spectroscopy for Sensi-tive Molecular Gas Sensing’, Laser Photonics Rev., 7,188–206 (2013).

56. D.A. Long, A. Cygan, R.D. van Zee, M. Okumura, C.E.Miller, D. Lisak, J.T. Hodges, ‘Frequency-StabilizedCavity Ring-Down Spectroscopy’, Chem. Phys. Lett., 536,1–8 (2012).

57. E.J. Moyer, D.S. Sayres, G.S. Engel, J.M.S. Clair, F.N.Keutsch, N.T. Allen, J.H. Kroll, J.G. Anderson, ‘DesignConsiderations in High-Sensitivity Off-Axis IntegratedCavity Output Spectroscopy’, Appl. Phys. B, 92, 467–474(2008).

58. S. Schilt, L. Thevenaz, ‘Wavelength Modulation Photoa-coustic Spectroscopy: Theoretical Description and Exper-imental Results’, Infrared Phys. Technol., 48, 154–162(2006).

59. Y.F. Ma, Y. He, X. Yu, C. Chen, R. Sun, F.K. Tittel, ‘HClppb-Level Detection Based on QEPAS Sensor Using aLow Resonance Frequency Quartz Tuning Fork’, Sens.Actuators B Chem., 233, 388–393 (2016).

60. A.D. Wood, M. Camac, E.T. Gerry, ‘Effects of 10.6 μmLaser Induced Air Chemistry on Atmospheric RefractiveIndex’, Appl. Opt., 10, 1877 (1971).

61. R.A. Rooth, A.J.L. Verhage, L.W. Wouters, ‘Photoa-coustic Measurement of Ammonia in the Atmosphere -Influence of Water Vapor and Carbondioxide’, Appl. Opt.,29, 3643–3653 (1990).

62. G. Wysocki, A.A. Kosterev, F.K. Tittel, ‘Influence ofMolecular Relaxation Dynamics on Quartz-EnhancedPhotoacoustic Detection of CO2 at λ= 2 μm’, Appl.Phys. B, 85, 301–306 (2006).

63. S. Schilt, J.P. Besson, L. Thevenaz, ‘Near-Infrared LaserPhotoacoustic Detection of Methane: The Impact ofMolecular Relaxation’, Appl. Phys. B, 82, 319–328 (2006).

64. S.T. Persijn, E. Santosa, F.J.M. Harren, ‘A VersatilePhotoacoustic Spectrometer for Sensitive Trace-Gas



Analysis in the Mid-Infrared Wavelength Region (5.1-8.0and 2.8-4.1 μm)’, Appl. Phys. B, 75, 335–342 (2002).

65. J.D. Lambert, Vibrational and Rotational Relaxation inGases, Clarendon Press, Oxford, 1977.

66. M. van Herpen, A.K.Y. Ngai, S.E. Bisson, J.H.P. Hack-stein, E.J. Woltering, F.J.M. Harren, ‘Optical ParametricOscillator-Based Photoacoustic Detection of CO2 at4.23 μm Allows Real-Time Monitoring of the Respirationof Small Insects’, Appl. Phys. B, 82, 665–669 (2006).

67. D.D. Arslanov, M.P.P. Castro, N.A. Creemers, A.H. Neer-incx, M. Spunei, J. Mandon, S.M. Cristescu, P. Merkus,F.J.M. Harren, ‘Optical Parametric Oscillator-BasedPhotoacoustic Detection of Hydrogen Cyanide forBiomedical Applications’, J. Biomed. Opt., 18, 107002(2013).

68. A. Olafsson, M. Hammerich, J. Bulow, J. Henningsen,‘Photoacoustic Detection of NH3 in Power-Plant Emis-sion with a CO2-Laser’, Appl. Phys. B, 49, 91–97 (1989).

69. M. Moeckli, M.W. Sigrist, ‘Field Monitoring of Traffic-Released Pollutants with Mobile Photoacoustic System’,Prog. Nat. Sci., 6, S622–S625 (1996).

70. D. Marinov, M.W. Sigrist, ‘Monitoring of Road-TrafficEmissions with a Mobile Photoacoustic System’,Photochem. Photobiol. Sci., 2, 774–778 (2003).

71. M.W. Sigrist, R. Bartlome, D. Marinov, J.M. Rey, D.E.Vogler, H. Waechter, ‘Trace gas Monitoring with InfraredLaser-Based Detection Schemes’, Appl. Phys. B, 90,289–300 (2008).

72. V. Spagnolo, P. Patimisco, S. Borri, G. Scamarcio,B.E. Bernacki, J. Kriesel, ‘Part-per-Trillion Level SF6

Detection Using a Quartz Enhanced PhotoacousticSpectroscopy-Based Sensor with Single-Mode Fiber-Coupled Quantum Cascade Laser Excitation’, Opt. Lett.,37, 4461–4463 (2012).

73. B. Zoltan, P. Andrea, S. Gabor, ‘Photoacoustic Instru-ments for Practical Applications: Present, Potentials, andFuture Challenges’, Appl. Spectrosc. Rev., 46, 1–37 (2011).

74. M.S. de Cumis, S. Viciani, S. Borri, P. Patimisco,A. Sampaolo, G. Scamarcio, P. De Natale, F. D’Amato,V. Spagnolo, ‘Widely-Tunable Mid-Infrared Fiber-Coupled Quartz-Enhanced Photoacoustic Sensorfor Environmental Monitoring’, Opt. Express, 22,28222–28231 (2014).

75. M. Pushkarsky, A. Tsekoun, I.G. Dunayevskiy, R. Go,C.K.N. Patel, ‘Sub-Parts-per-Billion Level Detectionof NO2 Using Room-Temperature Quantum CascadeLasers’, Proc. Natl. Acad. Sci. U. S. A., 103, 10846–10849(2006).

76. LSE-Monitors, http://www.lsemonitors.nl

77. J.A. De Gouw, S.T.L. Hekkert, J. Mellqvist, C. Warneke,E.L. Atlas, F.C. Fehsenfeld, A. Fried, G.J. Frost, F.J.M.Harren, J.S. Holloway, B. Lefer, R. Lueb, J.F. Meagher,D.D. Parrish, M. Patel, L. Pope, D. Richter, C. Rivera, T.B.

Ryerson, J. Samuelsson, J. Walega, R.A. Washenfelder,P. Weibring, X. Zhu, ‘Airborne Measurements of Ethenefrom Industrial Sources Using Laser Photo-AcousticSpectroscopy’, Environ. Sci. Technol., 43, 2437–2442(2009).

78. C.K.N. Patel, ‘Laser Photoacoustic Spectroscopy HelpsFight Terrorism: High Sensitivity Detection of ChemicalWarfare Agent and Explosives’, Eur. Phys. J. Spec. Top.,153, 1–18 (2008).

79. M.B. Pushkarsky, M.E. Webber, T. Macdonald, C.K.N.Patel, ‘High-Sensitivity, High-Selectivity Detection ofChemical Warfare Agents’, Appl. Phys. Lett., 88, 044103(2006).

80. R. Centeno, D. Marchenko, J. Mandon, S.M. Cristescu,G. Wulterkens, F.J.M. Harren, ‘High Power, WidelyTunable, Mode-Hop Free, Continuous Wave ExternalCavity Quantum Cascade Laser for Multi-Species TraceGas Detection’, Appl. Phys. Lett., 105, 261907 (2014).

81. F.B. Abeles, P.W. Morgan, M.E.J. Saltveit, Ethylene inPlant Biology, Academic Press, San Diego, CA, 1992.

82. S.M. Cristescu, J. Mandon, D. Arslanov, J. De Pessemier,C. Hermans, F.J.M. Harren, ‘Current Methods forDetecting Ethylene in Plants’, Ann. Bot., 111, 347–360(2013).

83. S.M. Cristescu, D. De Martinis, S.T. Hekkert, D.H. Parker,F.J.M. Harren, ‘Ethylene Production by Botrytis CinereaIn Vitro and in Tomatoes’, Appl. Environ. Microbiol., 68,5342–5350 (2002).

84. R. Schroeder, S.M. Cristescu, F.J.M. Harren, M. Hilker,‘Reduction of Ethylene Emission from Scots Pine Elicitedby Insect Egg Secretion’, J. Exp. Bot., 58, 1835–1842(2007).

85. L.A.C.J. Voesenek, M. Banga, R.H. Thier, C.M. Mudde,F.J.M. Harren, G.W.M. Barendse, C.W.P.M. Blom,‘Submergence-Induced Ethylene Synthesis, Entrapment,and Growth in 2 Plant-Species with Contrasting FloodingResistances’, Plant Physiol., 103, 783–791 (1993).

86. J. Tuomainen, C. Betz, J. Kangasjarvi, D. Ernst, Z.H.Yin, C. Langebartels, H. Sandermann, ‘Ozone Inductionof Ethylene Emission in Tomato Plants: Regulation byDifferential Accumulation of Transcripts for the Biosyn-thetic Enzymes’, Plant J., 12, 1151–1162 (1997).

87. S.C. Thain, F. Vandenbussche, L.J.J. Laarhoven, M.J.Dowson-Day, Z.Y. Wang, E.M. Tobin, F.J.M. Harren,A.J. Millar, D. Van Der Straeten, ‘Circadian Rhythms ofEthylene Emission in Arabidopsis’, Plant Physiol., 136,3751–3761 (2004).

88. B. Orihuel-Iranzo, M. Miranda, L. Zacarias, M.T.Lafuente, ‘Temperature and Ultra Low Oxygen Effectsand Involvement of Ethylene in Chilling Injury of ’RojoBrillante’ Persimmon Fruit’, Food Sci. Technol. Int., 16,159–167 (2010).



89. C. Hermans, M. Vuylsteke, F. Coppens, A. Craciun,D. Inze, N. Verbruggen, ‘Early Transcriptomic ChangesInduced by Magnesium Deficiency in ArabidopsisThaliana Reveal the Alteration of Circadian Clock GeneExpression in Roots and the Triggering of AbscisicAcid-Responsive Genes’, New Phytol., 187, 119–131(2010).

90. T. Pranamornkith, A. East, J. Heyes, ‘Influence of Exoge-nous Ethylene During Refrigerated Storage on Stora-bility and Quality of Actinidia Chinensis (cv. Hort16A)’,Postharvest Biol. Technol., 64, 1–8 (2012).

91. P.C. Ram, B.B. Singh, A.K. Singh, P. Ram, P.N. Singh, H.P.Singh, I. Boamfa, F. Harren, E. Santosa, M.B. Jackson,T.L. Setter, J. Reuss, L.J. Wade, V.P. Singh, R.K. Singh,‘Submergence Tolerance in Rainfed Lowland Rice: Phys-iological Basis and Prospects for Cultivar ImprovementThrough Marker-Aided Breeding’, Field Crop Res., 76,131–152 (2002).

92. E. Boamfa, A. Veres, P. Ram, M. Jackson, J. Reuss,F. Harren, ‘Kinetics of Ethanol and Acetaldehyde ReleaseSuggest a Role for Acetaldehyde Production in Toleranceof Rice Seedlings to Micro-Aerobic Conditions’, Ann.Bot., 96, 727–736 (2005).

93. J. Gustavsson, C. Cederberg, U. Sonesson, R. van Otter-dijk, A. Meybeck, Global Food Losses and Food Waste-Extent, Causes and Prevention, Food and AgricultureOrganization of the United Nations, Rome, 2011.

94. R.H. Veltman, J.A. Verschoor, J.H.R. van Dugteren,‘Dynamic Control System (DCS) for Apples (MalusDomestica Borkh. cv ’Elstar’): Optimal Quality ThroughStorage Based on Product Response’, Postharvest Biol.Technol., 27, 79–86 (2003).

95. R.H. Veltman, M.G. Sanders, S.T. Persijn, H.W. Peppe-lenbos, J. Oosterhaven, ‘Decreased Ascorbic Acid Levelsand Brown Core Development in Pears (Pyrus communisL-cv. Conference)’, Physiol. Plant., 107, 39–45 (1999).

96. H. Zuckermann, F.J.M. Harren, J. Reuss, D.H. Parker,‘Dynamics of Acetaldehyde Production During Anoxiaand Post-Anoxia in Red Bell Pepper Studied by Photoa-coustic Techniques’, Plant Physiol., 113, 925–932 (1997).

97. R.A. Herbert, ‘Nitrogen Cycling in Coastal MarineEcosystems’, FEMS Microbiol. Rev., 23, 563–590 (1999).

98. W.D.P. Stewart, G.P. Fitzgerald, R.H. Burris, ‘In SituStudies on N2 Fixation Using Acetylene Reduction Tech-nique’, Proc. Natl. Acad. Sci. U. S. A., 58, 2071–2078(1967).

99. M. Staal, S.T. Lintel-Hekkert, F. Harren, L. Stal, ‘Nitroge-nase Activity in Cyanobacteria Measured by the Acety-lene Reduction Assay: A Comparison Between BatchIncubation and On-Line Monitoring’, Environ. Micro-biol., 3, 343–351 (2001).

100. D.J. Wuebbles, K. Hayhoe, ‘Atmospheric Methane andGlobal Change’, Earth Sci. Rev., 57, 177–210 (2002).

101. J.H.P. Hackstein, C.K. Stumm, ‘Methane Production inTerrestrial Arthropods’, Proc. Natl. Acad. Sci. U. S. A., 91,5441–5445 (1994).

102. C.R. White, T.M. Blackburn, J.S. Terblanche, E. Marais,M. Gibernau, S.L. Chown, ‘Evolutionary Responses ofDiscontinuous Gas Exchange in Insects’, Proc. Natl.Acad. Sci. U. S. A., 104, 8357–8361 (2007).

103. S.M. Cristescu, J. Mandon, F.J.M. Harren, P. Merilainen,M. Hogman, ‘Methods of NO Detection in ExhaledBreath’, J. Breath Res., 7, 017104 (2013).

104. B.D. Costello, A. Amann, H. Al-Kateb, C. Flynn,W. Filipiak, T. Khalid, D. Osborne, N.M. Ratcliffe, ‘AReview of the Volatiles from the Healthy Human Body’,J. Breath Res., 8, 29 (2014).

105. A. Amann, B. Costello, W. Miekisch, J. Schubert,B. Buszewski, J. Pleil, N. Ratcliffe, T. Risby, ‘The HumanVolatilome: Volatile Organic Compounds (VOCs) inExhaled Breath, Skin Emanations, Urine, Feces andSaliva’, J. Breath Res., 8, 17 (2014).

106. D. Pleil, M.A. Stiegel, T.H. Risby, ‘Clinical BreathAnalysis: Discriminating Between Human EndogenousCompounds and Exogenous (Environmental) ChemicalConfounders’, J. Breath Res., 7, 017107 (2013).

107. M. Philips, ‘Breath Tests in Medicine’, Scientific Amer-ican, 74–79, 1992.

108. L. Pauling, A.B. Robinson, R. Teranishi, P. Cary, ‘Quanti-tative Analysis of Urine Vapor and Breath by Gas–LiquidPartition Chromatography’, Proc. Natl. Acad. Sci. U. S. A.,68, 2374–2376 (1971).

109. A. Amann, W. Miekisch, J. Schubert, B. Buszewski,T. Ligor, T. Jezierski, J. Pleil, T. Risby, ‘Analysis ofExhaled Breath for Disease Detection’, in AnnualReview of Analytical Chemistry, eds R.G. Cooks, J.E.Pemberton, Annual Reviews, Palo Alto, 455–482, Vol. 7,2014.

110. B. Henderson, A. Khodabakhsh, M. Metsala, I. Ventril-lard, F.M. Schmidt, D. Romanini, G.A.D. Ritchie, S.T.Hekkert, R. Briot, T. Risby, N. Marczin, F.J.M. Harren,S.M. Cristescu, ‘Laser Spectroscopy for Breath Analysis:Towards Clinical Implementation’, Appl. Phys. B, 124, 21(2018).

111. C.M.F. Kneepkens, G. Lepage, C.C. Roy, ‘The Potentialof the Hydrocarbon Breath Test as a Measure of LipidPeroxidation’, Free Radic. Biol. Med., 17, 127–160 (1994).

112. F. Harren, R. Berkelmans, K. Kuiper, S. te Lintel Hekkert,P. Scheepers, R. Dekhuijzen, P. Hollander, D. Parker, ‘On-Line Laser Photoacoustic Detection of Ethene in ExhaledAir as Biomarker of Ultraviolet Radiation Damage of theHuman Skin’, Appl. Phys. Lett., 74, 1761–1763 (1999).

113. B.W.M. Moeskops, M.M.L. Steeghs, K. van Swam, S.M.Cristescu, P.T.J. Scheepers, F.J.M. Harren, ‘Real-Time



Trace gas Sensing of Ethylene, Propanal and Acetalde-hyde from Human Skin In Vivo’, Physiol. Meas., 27,1187–1196 (2006).

114. S.M. Cristescu, R. Kiss, S.T. Hekkert, M. Dalby, F.J.M.Harren, T.H. Risby, N. Marczin, B.S.I. Harefield, ‘Real-Time Monitoring of Endogenous Lipid Peroxidationby Exhaled Ethylene in Patients Undergoing CardiacSurgery’, Am. J. Phys. Lung Cell. Mol. Phys., 307,L509–L515 (2014).

115. R. Romano, S.M. Cristescu, T.H. Risby, N. Marczin, ‘LipidPeroxidation in Cardiac Surgery: Towards Consensus onBiomonitoring, Diagnostic Tools and Therapeutic Imple-mentation’, J. Breath Res., 12, 027109 (2018).

116. B. Ryall, J.C. Davies, R. Wilson, A. Shoemark, H.D.Williams, ‘Pseudomonas Aeruginosa, Cyanide Accu-mulation and Lung Function in CF and Non-CF

Bronchiectasis Patients’, Eur. Respir. J., 32, 740–747(2008).

117. A.H. Neerincx, J. Mandon, J. van Ingen, D.D. Arslanov,J.W. Mouton, F.J. Harren, P.J. Merkus, S.M. Cristescu,‘Real-Time Monitoring of Hydrogen Cyanide (HCN) andAmmonia (NH3) Emitted by Pseudomonas aeruginosa’,J. Breath Res., 9, 027102 (2015).

118. A.H. Neerincx, Y.A.M. Linders, L. Vermeulen, R.A.Belderbos, J. Mandon, E. van Mastrigt, M.W. Pijnen-burg, J. van Ingen, J.W. Mouton, L.A.J. Kluijtmans,R.A. Wevers, F.J.M. Harren, S.M. Cristescu, P. Merkus,‘Hydrogen Cyanide Emission in the Lung by Staphylo-coccus aureus’, Eur. Respir. J., 48, 577–579 (2016).


Documents

Photoacoustic Spectroscopy in Trace Gas Monitoring · 2019. 4. 3. · Photoacoustic Spectroscopy in Trace Gas Monitoring Frans J.M. Harren and Simona M. Cristescu RadboudUniversityNijmegen,Nijmegen,