Open Access
Add to BibSonomyAbstract
Nitric oxide (NO) and carbon monoxide (CO) in Shanghai urban atmosphere have been measured during the EXPO 2010 by an optical trace gas monitoring system based on room-temperature pulsed quantum cascade lasers (QCL). The results showed obvious diurnal variation in their concentrations. A great correlation was found by analyzing the diurnal variation, indicating common emission sources for both NO and CO. However analysis of the data from street canyon measurements showed that on a vehicle-by-vehicle basis NO emissions correlated only weakly with CO emissions. A possible explanation to this issue has been given.
©2011 Optical Society of America
1. Introduction
Urban atmospheric pollution is currently a common issue confronted by humanity in the world. Especially in recent years, with the development and expansion of cities, the rapid increase in the number of cars being driven, this environmental problem has become more significant [1,2]. One type of urban pollution in China is characterized by nitrogen oxides (NOx), which can cause potentially further photochemical pollution (secondary pollution). The atmospheric concentration of NOx in many cities is beyond the level as specified in the national vehicle emission standards [3], while carbon monoxide (CO) concentration level is within the limit of this standard. On-road motor vehicles are considered as one of the main sources of NOx and CO in urban areas, and a related report [4] showed that more than 95% of the NOx released by light-duty gasoline vehicles was released in the form of NO. It is known that, both NO and CO are very harmful to human health and closely related to the destruction of atmospheric ozonosphere and the formation of acid rain, which could greatly influence the global climate. Thus, it is necessary to quantify the emission influence, particularly the concentrations of NO and CO in urban areas to enable protection of human and ecological health.
The optical detection technique of Tunable Diode Laser Absorption Spectroscopy (TDLAS) has proven to be very suitable for some applications such as remote sensing of environmental gases and pollutants in the atmosphere [5–11]. TDLAS can achieve higher measurement accuracy, lower detection limit, and larger detection dynamic range than some other type instruments, such as Fourier-transform spectrometers [12]. However, TDLAS technique has hardly been adopted to measure urban NO concentrations during the last decade. This is mainly due to a lack of suitable tunable laser sources in the mid-IR wavelength region for accessing the strong absorption band of NO, as most diode lasers operate only in the near infrared (NIR) wavelength region, where the absorption lines of NO are very weak and difficult to detect. However, the recent developments of quantum cascade lasers (QCLs) enable emission of coherent radiation in mid-IR at room temperature. This has led to an evolution and utilization of mid-IR spectroscopy in many technological and scientific fields [13], as most gas species of interest have much stronger absorption lines in mid-IR region. In particular, the advantages of high optical power output, wide tuning range and room temperature operation make QCLs especially suitable for the trace gas detection such as NO with TDLAS technique.
In this paper, a set of instrument for direct measurements of NO and CO trace gases based on room-temperature pulsed QC lasers has been introduced. The instrument combines two room-temperature pulsed QC lasers with a long-path multi-pass absorption cell to achieve the detection requirements of high sensitivity, selectivity and rapid response. This system is capable of taking highly sensitive measurements with a time resolution of 1 second, allowing time-resolved investigations of individual plumes and sources. Such instrument has been employed to monitor NO and CO emissions in Shanghai during the EXPO 2010. This paper reports variation and the range of typical urban background concentrations, along with a discussion of the possible causes to the variation of these concentrations. The data is also analyzed to examine the hypothesis that road vehicle traffic is the common source of CO and NO emissions in this local region.
2. Experimental principle and apparatus
2.1 Experimental principle
The basic formula of absorption spectroscopy is the Beer-Lambert law. It describes the relationship between transmitted and incident light intensities when the laser beam passes through a uniform gaseous medium [14]. To enhance the absorption path-length for high detection sensitivity, a multi-pass cell arrangement has been used in our experiment, as illustrated in Fig. 1 .
The intensity of a laser beam after traversing a multi-pass cell of a total path-length of l, filled with an absorption gas can be approximated as:
where I and I0 are the transmitted and incident laser intensity, respectively; R is the reflectivity of mirrors of the multi-pass cell; n is the number of reflections; X is the concentration of the gas in the cell; P is the total pressure; S0 and Φ(ν-ν0) are the linestrength and lineshape of a particular transition of the gas. The reflectivity R of mirrors forming the cell is typically higher than 99.99% in order to achieve a significant number of reflections, resulting in a very long effective absorption path-length l. By re-arrange the quantities of Eq. (1), we can obtain the following expression:where αν represents the spectral absorbance. The baseline value of ln(R n) can be determined by making an intensity ratio I0/I measurement with the multi-pass cell “absorber-free”.2.2 Experimental apparatus
Figure 2 shows the schematic diagram of our experimental apparatus. The experimental system consists of two mid-infrared QCL lasers, control electronics, a multi-pass cell, a mid-infrared detector and a data acquisition system.
The key components of this instrument consist of two distributed feedback QC lasers (QCL1 and QCL2, as shown in Fig. 2) operating at ~2190 cm−1 (or ~4.57 μm) and ~1900 cm−1 (or ~5.26 μm) which are suitable for detections of CO and NO, respectively. A laser controller manages the operation temperatures and current pulses of both QC lasers. The output beams of the QC lasers are combined together firstly by a set of steering mirrors, and then be compressed by several off-axis parabolic mirrors to reduce their beam sizes for an effective transmission through the multi-pass cell. The cell with a base length of 0.5 m is of Herriott type, resulting in a total optical pathlength of 76 m. As the mid-infrared light is not visible to human eyes, the alignment of the optical components is aided by a visible He-Ne laser beam which was adjusted to be coaxial with the QC laser beams. After traversing the multi-pass cell, the transmitted QCL light was focused onto a photovoltaic high-speed mercury cadmium telluride (MCT) detector, incorporating an integrated immersion lens and a Peltier thermoelectric cooler. The MCT detector output was sampled by a fast data-acquisition system with high precision and accuracy, and typically averaged over ~50000 data samples. The data was then transferred to a personal computer via a Universal Serial Bus (USB) interface for further analysis. Although two QC lasers have been employed in the system, only one detector was used to achieve similar response to both wavelength channels, and to minimize the system complexity. The pulsed radiation from each of the two QC lasers was generated in such a way so that they do not overlap in time. Therefore, a single detector can be used to measure them separately in a time multiplex fashion. Several additional mirrors help to fold the optical path, making the system more compact and easily adjustable. The gas cell was connected to a vacuum pump and a pressure controller, such that the pressure and flow rate in the measurement cell was maintained stable at preset values. The temperature of the multi-pass cell was also controlled to maintain a stable measurement conditions.
For all the measurements, the operation temperature of the 1900-cm−1 QCL for NO detection was set around ~29°C, while the temperature of the 2190-cm−1 QCL for CO detection was set around ~15.5°C. Both lasers were excited with 500-ns pulses of 10.3 Volt at a repetition rate of 50 kHz. The frequencies of the QCL outputs chirped by about ~1.0 cm−1 (or ~30 GHz) during the pulse operation. Signal average has been applied to improve the signal-to-noise ratio of acquired spectra. Figure 3 displayed two recorded absorption spectra of NO and CO, respectively. The corresponding total pressure in the measurement cell was maintained at 230 Torr, at a flow rate of 2L/min.
Fig. 3 Absorption spectra of NO and CO gases around 1900-cm−1 and 2190-cm−1, respectively. (a) a fit to 60 ppb NO spectrum. (b) a fit to 700 ppb CO spectrum. The measurement data are shown in red, and model fits to the data are shown in black.
The measured spectra were analyzed by fitting them with model profiles. The gas concentrations were calculated by a fitting algorithm, as shown in Fig. 3. By replacing air with pure nitrogen to the cell, the absorption baseline was measured. The detection limit was estimated based on the maximum standard deviation of three repeated baseline measurements. By analyzing sixty data points measured in one minute, the detection limit was calculated to be 3ppb Hz−1 for NO and 2 ppb Hz−1 for CO, respectively.
3 Field experiment
The field measurements were conducted in Shanghai, China, during the World EXPO 2010 (May–October). In fact, our measurements started from April 26, as the administrator of EXPO started test operation from April 20 before its official opening. The objective of this work was to measure the atmospheric level of NO and CO in an urban environment. During this period, the QCL instrument was running continuously to simultaneously monitor NO and CO concentrations, with a data measuring rate of 1 Hz. The detection limit of the system is tested to be about 3 ppb Hz−1 for NO and 2 ppb Hz−1 for CO, much lower than typical levels of NO and CO concentrations in atmosphere air.
The system was installed in a rooftop laboratory centrally located on a tall building in the Baoshan district in Shanghai, China. The concentrations of CO and NO in the surrounding atmosphere air were measured. The concentration level at this site was considered to be very close to background level. Close to the building is a wide and flat road called Youyi Road. Youyi Road is an access road to a large business facility, with relatively heavy traffic. We could hardly identify any other sources of pollution in the surrounding area, thus we make a preliminary assumption that motor vehicle exhaust are probably the major source of air pollution here. This is consistent with the general opinion that on-road motor vehicles are thought to be one of main sources of NO and CO in urban areas [4].
Air sampling was carried out at an approximate height of 35m from the roof via a pump and Teflon tube; and the atmosphere air was extracted through sampling inlet and created a constant flow through the cell. Concentration of CO and NO were recorded at a rate of 1 Hz. Data points have been further averaged over a time interval of one minute or one hour for presentation and analysis.
4 Results and discussion
4.1 Diurnal variation of NO and CO concentrations
Figure 4 shows the outcome of a continuous monitoring measurement for 24 hours. This specific measurement was conducted on April 29, 2010 and reflects a typical trend. Hourly data are used here to analyze the diurnal variation. Based on these measurements, the concentration varies typically from 17 ppb to 170 ppb for NO, and 360 ppb to 1170 ppb for CO, respectively. The strong diurnal variation in their concentrations can be seen obviously from Fig. 4. During night time, there was a continual increase in concentration level, and the high value appeared before seven o’clock in the morning, attributed to heavy-duty diesel vehicles emissions at night; after seven o’clock, the concentration showed a downward trend that the value were lower in the afternoon. Both NO and CO have similar diurnal variation. Since they all participate in atmospheric photochemical reactions, weather conditions such as solar radiation may have impact on NO and CO concentrations. After seven o’clock in the morning, as the solar radiation enhanced, the NO converted to NO2 by photochemical reaction at the presence of oxidants, this made the concentration decrease; later, the rate of that reaction began to slow down due to the weakening of solar radiation, and the concentration reassembled at night. Similar variation had been observed in CO. It participated in local photochemical reaction during the day, and accumulated at night. In addition, the atmospheric boundary layer had the same effect. It was more active during the day, at this time there was strong turbulence, and the emissions quickly spread into the atmosphere, which led to the low concentration. The boundary layer became more stable at night, thus the emissions accumulated in the ground and made the concentration increased.
4.2 Correlation between NO emissions and CO emissions
The NO emissions and CO emissions were measured simultaneously. To verify our assumption of their common local emission sources, we analyzed the relationship between NO and CO concentrations since they showed similar diurnal variations. For such analysis, as shown in Fig. 5 , we believe that effective data at short time interval will be more informative, hence minute-averaged data were used, instead of the hourly-averaged data from the same measurement as displayed in Fig. 4.
Fig. 5 Correlation between NO and CO emissions. A regression analysis gives a great correlation coefficient of R = 0.85 for CO to NO.
The analysis revealed that there was strong positive correlation between them. A regression analysis returned a large correlation coefficient of R=0.85 for CO to NO concentrations, as shown in Fig. 5. Their concentrations [CO] and [NO] exhibite the following dependence:
indicating that the NO emission level could be predicted from the measured CO level, and vice-versa. Such a strong positive correlation implies that NO and CO sources in this local region are the same. This is consistent with the fact that no other pollution sources other than exhaust emissions could be identified around the experimental site.4.3 Street canyon measurements
To make a simple investigation on the concentrations at the street level and to compare it with the concentrations at roof level, we transferred the inlet from the roof to the roadside 4 m away from the Youyi Road, at a height of 1 m above the ground. Figure 6 shows the concentration variation during this operation in an afternoon. It appeared that the concentration at the roadside was a little different from that on the roof of the building, implying that the gas distribution in this scale of space was not uniform during the experiment. However, it should be noted that the time of this measurement is very important. The reason for such a result may due to the enhanced dispersion associated with a deeper layer in the afternoon, making a difference of layers at different height less obvious. Only minutely data are displayed in Fig. 6. The occasional spikes seen in it may occur when there happened to be very busy traffic.
Measurement details of a selected time segment of Fig. 6 is presented in Fig. 7 , showing variations at short time scale by second-by-second data values. The sub-second time resolution of our instrument can resolve individual spikes in the 1-Hz data, which can be easily related to vehicles passing-by; but it quickly became apparent that this could not conclusively be achieved as concentration hops, looked similar as spikes, could be observed when the signal-to-noise ratio became too low. Low signal-to-noise ratio was usually caused by dropping of signal strength as a result of thermal creep, vibration or condensation, or the transmission baseline was insufficiently close to flat if the laser shifts slightly or changes temporal profile slightly [15]. However, large spikes of more than a few seconds in duration were considered to be related to vehicle density. As is shown in Fig. 7, the first peak of CO was produced by a single passing car. The second group of CO peaks, in which the peaks are not fully isolated from each other, was attributed to several consecutive cars. And the last large CO peak occurred when the traffic became very busy. The rapid formation of CO peaks was clearly identified, which then disappeared rapidly as a result of surface air turbulence. Such a strong correlation between CO concentration and passing vehicles is not reproduced in the associated NO data. The reason for this is that it is gasoline vehicles that accounted for the majority of vehicles driving in the city during the day. Compared to diesel vehicles, gasoline vehicles have relative higher CO emission but lower NO emission due to different fuels, engine structure, mixture formation and combustion model [16]. In addition, NO has relative lower background level. In the absence of high NO emission sources, the measured concentration is closed to a background level, and susceptible to system noise of the instrument at such a low NO concentration level.
Also we analyzed the correlation of this part of data. On a vehicle-by-vehicle basis, a regression analysis gave a very small correlation coefficient, that is R = 0.31 for CO to NO, as Fig. 8 shows, suggesting NO and CO have no significant association.
Fig. 8 Correlation between NO and CO emissions on a vehicle-by-vehicle basis, by analyzing the same data as plotted in Fig. 7.
This means one cannot predict the NO emission level of a vehicle from the CO emissions of the same vehicle. This differs from the result in Fig. 5. As we can see in Fig. 5, there was an approximate linear dependence between hourly averaged CO concentrations and NO concentrations. This kind of difference is mainly caused by different fuel engines, that is, different fuels, engine structure, mixture formation and combustion mode among different vehicles lead to the different emission characteristics of CO and NO concentrations.
A plausible explanation for these trends is that this is a measurement at micro timescales whereas the previous measurement in section 4.1 is at macro timescales. In the measurement at micro timescale, the sample inlet was set close to the roadside, thus what we measured were the separate emissions of individual vehicles in situ. In this situation, the instantaneous peak CO value could be identified clearly, but quickly disappeared, with only a few seconds duration, implying a rapid spread with the surface turbulence. Since different vehicles have different fuels, engine structure, mixture formation and combustion mode, they have different characteristics of CO and NO emissions. Even vehicles of the same model but different age of use, can have different emission characteristics [15]. Furthermore, catalytic converters of vehicles’ exhaust systems may also differ, and their efficiency may vary with the age of the vehicles [13]. They are generally less efficient when become aged. These factors make CO emissions emitted by individual vehicle correlated very weakly with NO emissions. While in previous measurement, our instrument was located in a rooftop laboratory on a tall building; our objective was to measure the atmospheric level of NO and CO in an urban environment. When there was only motor vehicles pollution present, this measurement could be taken as the total emissions for a geographical region, thus the result could be considered as low concentration accumulation after pollutions fully spread out. In this case, the effect of individual vehicle emissions became less significant. Only the quantity of vehicles passing-by became relevant. The more vehicles that pass by, the higher the concentrations are. While the number of vehicles on the street varies during the day, the ratio between gasoline and diesel vehicles stays similar. This contributes to a strong correlation between CO and NO emission. Although both CO and NO participate in atmospheric photochemical reactions, weather conditions such as solar radiation have similar impact on both of them. In addition, local air turbulence also has same effect on both of them. As a result of these factors, the diurnal variation of NO and CO concentrations as presented in Fig. 4 still remain strongly positive correlated.
5 Conclusions
In conclusion, NO and CO in Shanghai urban atmosphere have been measured by the portable optical sensing instrument based on room-temperature pulsed QC lasers during the EXPO 2010. The results showed very obvious diurnal variation in both of them. By analyzing the data at macro timescale measurement, a great correlation coefficient of R = 0.85 for CO to NO was obtained, indicating the same emission sources for both NO and CO. While analysis of the CO data from street canyon measurement with simultaneously recorded NO and CO concentration data showed that on a vehicle-by-vehicle basis NO emissions only weakly correlate with CO emissions. This obvious difference may attributed to separate emissions of individual vehicles which have different types of fuel engines, that is different fuels, engine structure, mixture formation and combustion mode among different vehicles lead to different characteristics of CO and NO concentrations. Furthermore, exhaust emissions may also depend on the types of catalytic converters used and their age of use. Therefore, the correlation at micro timescale was weak. However, when average numbers of gasoline and diesel vehicles remain similar, there exists a strong correlation between NO and CO emission at long term macro timescale.
Acknowledgments
The authors gratefully acknowledge the financial supports by the Main Direction Program of Knowledge Innovation of the Chinese Academy of Sciences (Grant No. KGCX2-YW-121), by the Strategic Priority Research Program of the Chinese Academy of Sciences (Grant No. XDA05040102), and also by the National “973” Program of China (Grant No. 2010CB234607).
References and links
1. B. Terje, “Impact of increased anthropogenic emission in Asia on troposphereic ozone and climate,” Tellus 3, 251–254 (1996).
2. L. J. Clapp and M. E. Jenkin, “Analysis of the relationship between ambient levels of O3, NO2 and NO as a function of NO in UK,” Atmos. Environ. 29, 923–946 (1995).
3. S. Zufeng, “Present status and developing countermeasure of tail exhaust gas measurement of China motor vehicle,” Auto Ind. Res. 4, 25–28 (2002).
4. J. B. Heywood, Internal Combustion Engine Fundamentals (Graw-Hill, New York, 1988).
5. D. D. Nelson, J. H. Shorter, J. B. Mcmanus, and M. S. Zahniser, “Sub-part-per-billion detection of nitric oxide in air using a thermoelectrically cooled mid-infrared quantum cascade laser spectrometer,” Appl. Phys. B 75(2-3), 343–350 (2002). [CrossRef]
6. F. D’Amato, P. Mazzinghi, and F. Castagnoli, “Methane analyzer based on TDL’s for measurements in the lower stratosphere: design and laboratory tests,” Appl. Phys. B 75(2-3), 195–202 (2002). [CrossRef]
7. G. Hancock, V. L. Kasyutich, and G. A. D. Ritchie, “Wavelength-modulation spectroscopy using a frequency-doubled current-modulated diode laser,” Appl. Phys. B 74(6), 569–575 (2002). [CrossRef]
8. P. W. Werle, P. Mazzinghi, F. D’Amato, M. De Rosa, K. Maurer, and F. Slemr, “Signal processing and calibration procedures for in situ diode-laser absorption spectroscopy,” Spectrochim. Acta A Mol. Biomol. Spectrosc. 60(8-9), 1685–1705 (2004). [CrossRef] [PubMed]
9. G. C. Sauer, T. J. Pisano, and R. D. Fitz, “Tunable diode laser absorption spectrometer measurements of ambient nitrogen dioxide, nitric acid, formaldehyde, and hydrogen peroxide in Parlier, California,” Atmos. Environ. 37(12), 1583–1591 (2003). [CrossRef]
10. P. Vogel and V. Ebert, “Near shot noise detection of oxygen in the A-band with vertical-cavity surface-emitting lasers,” Appl. Phys. B 72, 127–135 (2001).
11. T. Iseki, H. Tai, and K. Kimura, “A portable remote methane sensor using a tunable diode laser,” Meas. Sci. Technol. 11(6), 594–602 (2000). [CrossRef]
12. J. L. Jiménez, M. D. Koplow, D. D. Nelson, M. S. Zahniser, and S. E. Schmidt, “Characterization of on-road vehicle NO emissions by a TILDAS remote sensor,” J. Air Waste Manage. Assoc. 49, 463–470 (1999).
13. M. T. McCulloch, N. Langford, and G. Duxbury, “Real-time trace-level detection of carbon dioxide and ethylene in car exhaust gases,” Appl. Opt. 44(14), 2887–2894 (2005). [CrossRef] [PubMed]
14. X. Zhou, “Diode-laser absorption sensors for combustion control,” Ph.D. Dissertation (Stanford University, 2005), page 5.
15. K. G. Hay, “Gas sensing using quantum cascade lasers,” Ph.D. Dissertation (University of Strathclyde, 2010), page 97.
16. Y. Yuan, Ya ping Xiao, Li xin Fu, Motor Vehicle Pollution Control (Chemical Industry Press, Beijing, 2000).
