Influence of photolysis on multispectral photoacoustic measurement of nitrogen dioxide concentration

Abstract

Multispectral photoacoustic instruments are commonly used to measure aerosol and nitrogen dioxide (NO₂) light absorption coefficients to determine the radiation budget of the atmosphere. Here a new photoacoustic system is developed to explore the effect of photolysis on the measured signal in a multispectral photoacoustic spectrometer. In this system, a 405-nm laser is used primarily as light source for photolysis. Additionally, a well-overlapped 532-nm laser, modulated at the resonant frequency of the photoacoustic instrument, is used to probe the NO₂ concentration. As a result, the photolysis effect at 405 nm can be observed by the photoacoustic instrument through the 532-nm laser. This work determines an 11% reduction of the photoacoustic signal caused by the photolysis effect for typical conditions, which needs to be taken into account when calibrating multispectral photoacoustic spectrometers with NO₂.

Implications:

Multispectral photoacoustic instruments are commonly used to measure aerosol and nitrogen dioxide (NO₂) light absorption coefficients to determine the radiation budget of the atmosphere. A 405-nm laser is often used in these multispectral photoacoustic instruments. Although NO₂ absorbs strongly at 405 nm, it also has a strong photolysis pathway that is accessible by light of the same wavelength. Photolysis reduces the photoacoustic signal, necessitating special care when interpreting photoacoustic measurements. This paper offers a method for the multispectral photoacoustic instrument user to quantify the influence of the 405-nm NO₂ photolysis effect on the photoacoustic signal.

Introduction

The optical properties of aerosol and gases are important factors for atmospheric radiation transfer. They affect the lifetime of clouds and the global and regional radiation balance and they dominate the uncertainties in total global radiative forcing (Horvath, 1993; Moosmüller et al., 2009; Lohmann and Feichter, 2005). Specifically, nitrogen dioxide (NO₂) light absorption can play an important role in the ultraviolet (UV) and visible radiation budget of the atmosphere (Intergovernmental Panel on Climate Change [IPCC], 2007; VasilKov et al., 2009). Photoacoustic measurements can be used for the sensitive characterization of NO₂ absorption coefficients and concentrations (e.g., Terhune and Anderson, 1977; Slezak et al., 2003); in addition, NO₂ is commonly used to calibrate photoacoustic measurements of aerosol light absorption (Adams et al., 1989; Arnott et al., 2000). For all of these applications, absorption and photolysis of NO₂ need to be better understood.

The photolysis reaction can be separated into three distinct wavelength regions based on the behavior of quantum yield (molecules photon⁻¹) and photolysis mechanism as a function of wavelength (Jones and Bayes, 1973). The first region is the short wavelength region (295 nm < λ < 398 nm) where the photon energy is higher than the photodissociation energy (398 nm for reaction represented by eq 1) with a primary quantum yield of unity. The mechanism, which is described by eq 1 and eq 2, predicts a quantum yield of 2 for nitric oxide (NO) formation, which is close to the values observed (Leighton, 1961).

1

2

In the second wavelength region (398 nm < λ < 430 nm), the photon energy is less than the dissociation energy and the quantum yield falls off rapidly with wavelength. Because very few of the NO₂ molecules are vibrationally excited, most of their internal energy is rotational energy. The photolysis in this spectral region has been attributed to the utilization of rotational energy and thermal energy (transferred from molecular collision) to make up the difference between dissociation energy and photon energy (Jones and Bayes, 1973; Creel and Ross, 1976). As the wavelength is increased beyond about 430 nm, into the third wavelength region (430 nm < λ < 630 nm), internal energy and relative kinetic energy cannot offer sufficient additional energy to break the NO₂ molecule. In this region, photolysis is caused by the reaction of an electronically excited NO₂ molecule with a ground state NO₂ molecule, as shown in eq 3 (Norrrish, 1929a, 1929b), or by an alternative reaction shown in eq 4 (Ford and Jaffe, 1963).

3

4

Both reactions have a very small quantum yield for NO formation that is slowly decreasing with wavelength.

The photoacoustic effect can be used to quantify the transfer of radiative energy to heat, in many cases enabling the direct measurement of absorption coefficients (Moosmüller et al., 2009). Photoacoustic instruments have been used since 1973 to study photolysis of NO₂ by comparing extinction and photoacoustic spectra of NO₂ (Harshbarger and Robin, 1973). The original system developed by Harshbarger and Robin (Figure 1) used a 4-kW xenon lamp as the light source, a chopper to modulate the incident light, a monochromator for wavelengh scanning, a microphone, preamplifier, and lock-in amplifier to obtain the acoustic signal, and a radiometer to measure the optical power of the monochrometer output beam. As shown in Figure 2, the general features of the two spectra look quite similar between 4100 Å (410 nm) and 7000 Å (700 nm). However, below 410 nm, the spectrophone (i.e., photoacoustic) spectrum shows a sharp drop in signal strength that is not observed in the optical spectrum. This abrupt loss of part of the photoacoustic signal is due to NO₂ photolysis. Harshbarger and Robin's research did not quantify the photolysis effect, although their work demonstrated that photoacoustic instrument can be used in the measurement and understanding of photolysis.

Figure 1. Experimental setup used by Harshbarger et al. (1973).

Figure 2. Comparison of the optical extinction and spectrophone (i.e., photoacoustic) spectra of NO₂ at 1333 Pa pressure (Harshbarger et al., 1973).

Further research has been done by Lewis (2007), who quantified the photolysis effect of NO₂ in association with the calibration of a photoacoustic instrument at 405 nm, as shown in Figure 3. The ratio between the extinction and apparent photoacoustic absorption coefficients was used to quantify the photolysis effect of NO₂, since at this wavelength extinction equals absorption, because Rayleigh scattering is negligible when compared with extinction. The apparent photoacoustic absorption coefficient is significantly lower than the extinction coefficient because part of the laser energy is used to break NO₂ molecules rather than to heat the surrounding air. However, for a multispectral photoacoustic instrument, this research is not sufficient to quantify the photolysis of NO₂, as explained at the end of the following paragraph.

Figure 3. NO₂ gas calibration results at 405 nm from Lewis (2007).

Multispectral photoacoustic instruments and wide-spectral-range photoacoustic spectrometers are now available and are used to measure light absorption by aerosol and gases (Haisch et al., 2012; Ajtai et al., 2010; Lewis et al., 2008). A 405-nm laser is often included in the multispectral photoacoustic instruments. For example, a dual-wavelength photoacoustic instrument operating at 405 and 870 nm was used during the 2006 Fire Lab at Missoula Experiment to measure light scattering and absorption coefficients by smoke from the combustion of a variety of biomass fuels (Lewis et al., 2008), a three-wavelength (405, 532, and 781 nm) photoacoustic instrument was used for studying optical-chemical-microphysical relationships and for closure studies (Flowers et al., 2010), and a four-wavelength (266, 355, 532, and 1064 nm) photoacoustic instrument was used for intercomparisan with a multiwavelength aethalometer under suburban, wintry conditions (Ajtai et al., 2011). Because the NO₂ photolysis effect around 400 nm is strong (Jones and Bayes, 1973), it needs to be accounted for when calibrating multispectral photoacoustic instruments with NO₂. Although NO₂ photolysis does not prevent those multispectral photoacoustic instruments from being used for NO₂ light absorption measurements, the performance of the NO₂ light absorption measurements must be quantified and improved by characterization of the photolysis effect. Lewis’ (2007) results are not sufficient to quantify the photolysis effect for multispectral photoacoustic instruments because they were obtained at a single wavelength. In multispectral photoacoustic instruments, multiple lasers are generally operated simultaneously at different modulation frequencies and the acoustic signals are separated according modulation frequency using fast Fourier transform. As a reference, Figure 4 shows the absorption spectrum in terms of absorption cross-section of NO₂. In such instruments, coupled photolysis, which is caused by multiple laser wavelengths and the coactions of them with multiphoton absorption (Hakala et al., 1974), has effects on acoustic signals from all lasers. Although the photoacoustic signal from each laser can be calibrated individually by using Lewis’ method to correct for the photolysis effect caused by that laser itself, the coupled photolysis also needs to be taken into account.

Figure 4. Absorption cross-section spectrum of NO₂ at 294 K (Vandaele, 1998).

This paper describes a novel photoacoustic system to quantify the photolysis effect in NO₂ light absorption measurements using multispectral photoacoustic instruments.

Experimental Setup

The experimental setup shown in Figure 5 consists of two lasers, a photoacoustic resonator, a piezoelectric transducer, data acquisition electronics, and a computer. The piezoelectric transducer is used to determine the resonant frequency of the photoacoustic resonator. Beams of a 532-nm laser (123 mW) and a 405-nm laser (180 mW) are sent through the photoacoustic resonator to generate the microphone signals used as input for two lock-in amplifiers at the modulation frequency of the two laser beams. Lock-in amplifier 1 controls a chopper that modulates the 532-nm laser at the resonant frequency of the photoacoustic instrument, lock-in amplifier 2 controls modulation of the 405-nm laser. We can also manually modulate the 405-nm laser at very low modulation frequency (lower than 0.1 Hz). The laser power for each laser is measured by inserting a power meter during instrument calibration (not shown in Figure 5).

Figure 5. Schematic diagram of the photoacoustic instrument, including two lasers, chopper, photoacoustic resonator with microphone and piezoelectric transducer, two lock-in amplifiers, and personal computer (PC) (color figure available online).

Light absorption of NO₂ at 532 nm can be detected using the microphone because the 532-nm laser is chopped at the resonant frequency of the photoacoustic instrument. Light absorption of NO₂ at 405 nm does not produce a detectable photoacoustic signal because the modulation frequency of the 405-nm laser is far away from the acoustic resonance frequency. When the 405-nm laser is turned on, it causes NO₂ photolysis, which changes the absorption of the 532-nm laser beam because the two laser beams are well overlapped, as shown schematically in Figure 5. This change in absorption of the 532-nm radiation can clearly be observed in the changed microphone signal.

Calibration of the Photoacoustic Instrument at 532 nm

The system is calibrated at 532 nm by using sufficiently high concentrations of NO₂ in the resonator to obtain both the extinction (from gas transmissivity) and absorption (photoacoustic measurement) coefficients. A photodiode is placed to measure the transmissivity during the calibration. Absorption is sufficiently large that Rayleigh scattering can be ignored. The system was calibrated using Beer's Law,

5

6

where β_ext is the extinction coefficient, T the transmissivity, L the length of the photoacoustic cell, and β_abs the absorption coefficient of the NO₂ in air sample in the cell. The absorption coefficient is calibrated, as shown in Figure 6. A sample with 330 ppb NO₂ in air is pumped from a gas cylinder into the photoacoustic instrument continually with a flow rate of 1 liter per minute during the calibration. More calibration detail has been given by Arnott et al. (2000) and Lewis (2007).

Figure 6. NO₂ calibration of the photoacoustic instrument at 532 nm. For NO₂, the absorption coefficient is equal to the extinction coefficient. The measurement with the uncalibrated instrument (shown as gray solid line) is adjusted by multiplying with a constant, yielding the calibrated absorption coefficient, shown as gray dashed line.

Measurements

Monitoring 405 nm NO₂ photolysis with photoacoustic measurements at 532 nm

Once the photoacoustic system is calibrated at 532 nm, the photodiode used during the calibration is removed and the 405-nm laser is setup to be the photolysis light source, as shown in Figure 5. A sample with 330 ppm of NO₂ in air is introduced into the photoacoustic instrument continually with a flow rate of 1 liter per minute. The absorption coefficient of the sample in the photoacoustic instrument decreases due to photolysis when turning on the 405-nm laser and returns to the original value after turning off the 405-nm laser, once the sample in the resonator has been completely replaced by the continuous flow. After the 405-nm laser is turned on, it takes ∼4 sec to obtain dynamic equilibrium between photolytic destruction of NO₂ and replacement with undepleted NO₂, as shown in Figure 7. The absorption coefficient increases when the 405-nm laser is turned off and decreases when the 405-nm laser is turned on due to NO₂ photolysis at 405 nm.

Figure 7. Short-term measurement of the sample absorption coefficient at 532 nm with 405-nm laser on and off.

A more long-term measurement is shown in Figure 8, with two dashed lines indicating the absorption coefficient of the sample at two different equilibrium states. The change of the sample absorption coefficient at 532 nm can be observed clearly from the photoacoustic signal while switching on and off the 405-nm laser manually. The difference in absorption coefficient between the two equilibrium states is ∼11%.

Figure 8. Photolysis of NO₂ caused by the 405-nm laser. The line shows the change of the sample absorption coefficient at 532 nm due to the photolysis of NO₂ caused by the 405-nm laser with 123 mW laser power. The difference between absorption coefficients and NO₂ concentrations with the 405-nm laser turned on and off is ∼11%.

Multispectral photoacoustic instruments are operated at the resonance frequency of their acoustic resonator, which for our instruments is ∼1500 Hz. To determine the effect of NO₂ photolysis on the operation of a dual-wavelength photoacoustic instrument operating at 405 and 532 nm, the 405-nm laser in our system is modulated at a high frequency (i.e., 1000 Hz) and the change of absorption of the continuously flowing sample is observed with the 532-nm laser. As observed in Figure 9, the 532-nm light absorption coefficient measured with the 405-nm laser, power-modulated at 1000 Hz, is about the average of the absorption coefficients measured with the 405-nm laser turned off (i.e., the highest values of ∼80,000 Mm⁻¹ shown in Figure 9) and that with the 405-nm laser continuously operating (i.e., the lowest values of ∼73,000 Mm⁻¹ shown in Figure 9). When our setup is operated with the 405-nm laser modulation frequency larger than 20 Hz, the absorption coefficient is very stable. The reduction in 532-nm absorption and NO₂ concentration is solely dependent on the average power of the 405-nm laser.

Figure 9. Sample absorption coefficient at 532 nm with the 405-nm laser turned on and off followed by high-frequency modulation of the 405-nm laser power simulating operation conditions for multispectral photoacoustic instruments.

Monitoring 532-nm-induced NO₂ concentration change with photoacoustic measurements at 405 nm

After exchanging the roles of the 405-nm laser and the 532-nm laser, the 532-nm laser is turned on and off, while the 405-nm laser is used for photoacoustic monitoring of the NO₂ concentration. The resulting change in sample absorption coefficient between the 532-nm laser turned on and off was ∼2%, as shown in Figure 10. The 532-nm wavelength is located in the third photolysis region, as discussed in Introduction. It is commonly assumed that the quantum yield for NO production from NO₂ photolysis is lower than 1% at 532 nm and is ∼30% at 405 nm (Jones and Bayes, 1973; Creel and Ross, 1976), which wouldn't explain the strong signal seen in Figure 10.

Figure 10. Change of sample light absorption coefficient measured at 405 nm. The difference between 405-nm absorption coefficients with the 532-nm laser turned on and off is ∼2%.

Further qualitative information comes from measurements from an NO analyzer (Thermo 42C NO-NO₂-NO_X Analyzer) that was connected to the output of the photoacoustic resonator to measure NO concentrations created by photolysis of NO₂. For a sample gas containing 330 ppm NO₂ in air, the 405-nm photolysis of NO₂ is sufficiently strong that the analyzer is saturated when only the 405-nm laser is turned on, as shown in Figure 11.

Figure 11. NO concentration change measured by the NO analyzer. Photolysis of NO₂ at 405 nm was sufficiently strong that the NO analyzer was saturated at its upper limit of 20,000 ppb by the NO produced.

The same method was used to check if there is any photolysis of NO₂ at 532 nm by turning the 532 nm on and off (only the 532-nm laser is used). The NO analyzer did not show a meaningful signal of NO concentration. It is speculated that the 2% concentration difference (Figure 10) may be caused by the following:

1.	Multiphoton absorption and photolysis involving coactions of the 405- and 532-nm radiation. Multiphoton absorption and photolysis phenomena for NO2 have been previously demonstrated by others (Hakala et al., 1974).
2.	Sequential photon absorption (rather than simultaneous multiphoton absorption). Some intermediate states are sufficiently long lived, for example, a relaxation time as high as 4 µsec has been reported for the exited state 2B2 (Kalkman and Kesteren, 2008).
3.	Temperature effects influence the implicit photochemistry within the laser beams. Absorption cross-section and photolysis yields are affected by temperature (Burrows et al., 1998; Roehl, 1994). When the 532-nm laser is turned on, the temperature within the laser beam increases, thereby increasing the absorption cross-section and photolysis yield and consequently lowering the NO2 concentration in the laser beam.

Discussion

Two main sources of error need to be considered for high-sensitivity photoacoustic detection of NO₂. First, due to photolysis and photochemical reaction (Jones and Bayes, 1973), ozone can be produced in the resonant cavity of the photoacoustic instrument. The absorption of ozone at 532 nm will increase the photoacoustic signal and decrease the measured photolysis effect. Because the ozone absorption cross-section is about 50 times smaller than that of nitrogen dioxide (Orphal, 2003), this decrease is small when the ozone concentration is low. Second is the relaxation effect of the measured molecules, reported by Kalkman and Kesteren (2008) in a high-sensitivity photoacoustic detection of NO₂ at 444 nm. Their research shows that in detection of NO₂ in N₂, the relaxation does not affect the results. However, a signal decrease is observed when oxygen is mixed into the gas mixture. The relaxation time can be described by the phase shift of photoacoustic signal. The signal is directly proportional to the cos θ (Moosmüller et al., 2009), where θ is the phase shift of photoacoustic signal.

In our research, the phase shift of photoacoustic signal that represents the concentration of NO₂ is measured. An 8° phase shift at 532 nm and an 11° phase shift at 405 nm were observed. These phase shifts are relatively small. The effect on the measured concentration is less than 1.2%.

Conclusion

A new method is developed to quantify the photolysis effect in NO₂ using multispectral photoacoustic instruments and the effect on calibrating such instruments with NO₂. A photoacoustic instrument with continuous flow is used to measure the NO₂ photolysis effect using two overlapped lasers. In this experiment, which simulates the operation of multispectral photoacoustic instruments, a maximum 11% reduction of the 532-nm photoacoustic signal is due to the photolysis effect observed at 405 nm, and a 2% reduction of the 405-nm photoacoustic signal is observed due to the simultaneous presence of 532-nm radiation. For typical operating conditions of a multispectral photoacoustic instrument, half of these reductions are expected because of the high-frequency modulation and 50% duty cycle of the laser. The reduced photoacoustic signal at 405 nm due to the presence of the 532-nm beam is not fully understood but may be due to multiphoton processes involving simultaneous 405- and 532-nm radiation.

Acknowledgment

This paper is based upon work supported by NASA EPSCoR under Cooperative Agreement No. NNX10AR89A, by NASA ROSES under grant no. NNX11AB79G, and by the National Science Foundation under grant no. AGS-1040046. Helpful discussions with Dean Venables are gratefully acknowledged.