Characteristics of ozone and ozone precursors (VOCs and NOx) around a petroleum refinery in Beijing

Journal of Environmental Sciences 26(2014)

332–342

https://www.360docs.net/doc/9d2126031.html,

Journal of Environmental Sciences

Available online at

https://www.360docs.net/doc/9d2126031.html,

Characteristics of ozone and ozone precursors (VOCs and NO x )around a petroleum re?nery in Beijing,China

Wei Wei 1,2,Shuiyuan Cheng 1,?,Guohao Li 1,Gang Wang 1,Haiyang Wang 1

1.Department of Environmental Science and Engineering,Beijing University of Technology,Beijing 100124,China

2.State Environmental Protection Key Laboratory of Sources and Control of Air Pollution Complex,Tsinghua University,Beijing 100084,China

a r t i c l e i n f o

Article history:

Received 11March 2013revised 09May 2013accepted 10May 2013

Keywords:ozone

volatile organic compounds (VOCs)photochemical degradations petroleum re?nery

DOI:10.1016/S1001-0742(13)60412-X

a b s t r a c t

A ?eld measurement campaign for ozone and ozone precursors (VOCs and NO x )was conducted in summer 2011around a petroleum re?nery in the Beijing rural region.Three observation sites were arranged,one at southwest of the re?nery as the background,and two at northeast of the re?nery as the downwind receptors.Monitoring data revealed the presence of serious surface O 3pollution with the characteristics of high average daily mean and maximum concentrations (64.0and 145.4ppbV in no-rain days,respectively)and multi-peak diurnal variation.For NO x ,the average hourly concentrations of NO 2and NO were in the range of 20.5–46.1and 1.8–6.4ppbV ,respectively.For VOC measurement,a total of 51compounds were detected.Normally,TVOCs at the background site was only dozens of ppbC,while TVOCs at the downwind sites reached several hundreds of ppbC.By subtracting the VOC concentrations at background,chemical pro?les of VOC emission from the re?nery were obtained,mainly including alkanes (60.0%±4.3%),alkenes (21.1%±5.5%)and aromatics (18.9%±3.9%).Moreover,some di ?erences in chemical pro?les for the same measurement hours were observed between the downwind sites;the volume ratios of alkanes with low reactivity and those of alkenes with high reactivity respectively showed an increasing trend and a decreasing trend.Finally,based on temporal and spatial variations of VOC mixing ratios,their photochemical degradations and dispersion degradations were estimated to be 0.15–0.27and 0.42–0.62,respectively,by the photochemical age calculation method,indicating stronger photochemical reactions around the re?nery.

Introduction

Photochemical smog,produced from the photochemical reactions of NO x and volatile organic compounds (VOCs)in the presence of sunlight,has been an increasing concern.It is chemically characterized by a high level of oxidant compounds,mainly ozone (O 3),which adversely in?uence ambient air quality,human health,and agricultural pro-duction (Coleman et al.,2008;Lamorena and Lee,2008;Shao et al.,2009;Zelm et al.,2008).In eastern China,increasing surface O 3concentrations have been observed,

?Corresponding

author.E-mail address:chengsy@https://www.360docs.net/doc/9d2126031.html,

due to the increased emissions of precursor pollutants,which were caused by the vigorous economic development and rapid urbanization in recent years.Wang et al.(2009a)reported that the surface O 3at a coastal site in Hong Kong,as the background atmosphere of southern China,rose by 0.58ppbV /yr during 1994–2007;and Tang et al.(2009)found that in Beijing,the maximum and average O 3concentrations increased linearly at rates of 1.6and 1.1ppbV /yr respectively during 2001–2006.Moreover,many Chinese urban areas have frequently su ?ered from high-O 3episodes.Recently,in Beijing,the annual days over the 1-hr average Chinese national O 3standard (93ppbV)reached 40–80,and the days over the 8-hr average WHO O 3standard (50ppbV)exceeded 80(An et al.,

Journal of Environmental Sciences 26(2014)332–342

333

2008;Geng et al.,2010).Consequently,a great deal of research on the mechanism of photochemical O 3formation on the regional-and urban-scale has been conducted in China.These studies implied that photochemical O 3production was limited by the concentrations of VOCs in many Chinese eastern urban areas,and limited by the concentrations of NO x in western areas (Liu et al.,2010;Shao et al.,2009;Wang et al.,2010;Zhang et al.,2008).In addition,studies on the source apportionment of Beijing photochemical O 3by Song et al.(2008)and Wang et al.(2009b)found road vehicles and the petrochemical industry as the top 2contributors,with the contributions of 20%–30%and 15%–28%,respectively.

However,as one of the important contributors,the petrochemical industry has been the subject of few on-site measurement campaigns of its VOCs speciation emission in China,except that of Liu et al.(2008a),which tested the VOCs chemical pro?les from the petrochemical industry in April by ambient air sampling in a selected petrochemical industrial area.Here,we performed a ?eld measurement campaign around a petroleum re?nery in Beijing during the summer months,when the ground ozone concentra-tions are normally highest in Beijing.The ambient air pro?les were studied by determining a group of 56VOCs (required in US EPA PAMS)and by monitoring NO x and O 3at the same time.It would be helpful to obtain the

characteristics of ozone precursor VOCs emitted from the petroleum re?nery industry in China,to properly under-stand their roles in local photochemical ozone pollution.In addition,considering the potential threat of long-term exposure to some VOC compounds to human health,this study would also be helpful for human health assessment in the petroleum re?nery industry area.

1Field measurement

1.1Measurement protocol

The petroleum re?nery selected in this study is located in 40km southwest of Beijing city center,with annual re?ning capacity of about 10million tons.It occupies an area of about 1.5km 2,with length of about 3.0km and width of about 0.5km,and lies in a valley,closely surrounded by Yan Mountain with the height of 200–700meters in the west and northwest.Meteorological statistics show that the diurnal variation of prevalent wind direction follows the pattern of mountain-valley wind circulation in Beijing summer.Normally,northeast winds prevail before 08:00,followed by more and more winds in southern directions.South and southwest become a dominant wind

0.0%

5.0%

10.0%15.0%20.0%N

NNE

ENE E

ESE SE

SSE

S SSW

WSW

WNW

NNW 8:00

0.0%

5.0%

10.0%15.0%20.0%N

NNE

ENE E

ESE

SSE

S SSW

WSW

WNW

NNW 11:00

0.0%

5.0%

10.0%15.0%20.0%N

NNE

ENE

E ESE SE

SSE

SSW

WSW

WNW

NNW 14:00

0.0%5.0%

10.0%15.0%20.0%

NNE

ENE

ESE

SSE

SSW

WSW

W WNW

NNW 17:00

Fig.1Percentage of di ?erent wind directions at di ?erent hours in Beijing summer.

334Journal of Environmental Sciences26(2014)332–342

direction after noon.Figure1shows the percentage of di?erent wind directions at di?erent hours in the Beijing summer of2011.According to the prevalent wind direction and the limit of terrain,we set up three observation sites around the re?nery,as shown in Fig.2,site A at0.5km southwest of the re?nery as the background,site B at the north gate of the re?nery and site C at0.5km northeast of the re?nery as the downwind receptors.We thought that the complex terrain where the re?nery lies would make the southern winds(including SSW,S and SSE winds)take a bend to the east at the foot of Yan mountain,and the pollutants emitted from the re?nery would be transferred from site B to site C with the air mass.This had been con?rmed by research on the surface atmospheric di?usion model in the Yanshan petroleum industrial area(IAPCAS, 1988;Hu et al.,1999).In addition,there are no main tra?c roads in the small valley,and we believed that this comparison monitoring could represent the air quality in?uenced only by the re?nery.

This measurement was conducted from July to August 2011,the months in which the ozone concentration levels in Beijing are normally highest.During this period,month-ly samples were taken on?ve days concurrently at three sites,for three1-hr periods(8:00–9:00,12:00–13:00,and 17:00–18:00),which were used to determine the concen-trations of VOCs compounds.The height of air sampling was about1.5m at the three sites.In addition,O3and NO x(NO and NO2)were monitored by online automatic analyzers during the whole measurement months,which were set up at site B with the air inlet height of10m.

1.2Analytical methodology

We collected the gaseous samples into vacuum3L-summa canisters with?ow controller restrictor(CS1200,Entech Inc.),which could ensure a constant?ow rate during the 1-hr sampling process.Then the samples were analyzed

C A

Fig.2Three measurement sites arranged around the re?nery.in the laboratory to determine the VOC concentrations, according to EPA TO14/15methods(US EPA,1999). The samples of400mL were?rst passed through three-step cryotraps to remove water,CO,CO2,N2and O2, and to enrich the VOCs by liquid nitrogen in a cryogenic preconcentrator(Model7100,Entech Inc.).Then the con-centrated samples were thermally desorbed and carried into the GC-MS(Model7890A/5975C,Agilent Inc.).Here a HP-PONA column(50m×0.2mm ID×0.5μm)was selected in the GC.The column was initially held at an oven temperature of–20°C for1min,raised to0°C at a rate of5°C/min,then to100°C at a rate of10°C/min, then to150°C at a rate of5°C/min,and?nally raised to 260°C at a rate of12°C/min.This VOC detection system could identify most C3–C12hydrocarbons except propane. Finally,PAMS certi?ed gas(Spectra Gases,Inc.,including 56hydrocarbons)was used to quantify the concentration of target compounds.A calibration curve was calculated by using?ve concentrations(covering0–20ppbV)for each compound.Correlation coe?cients ranging from0.99to 1.00showed the good linearity between the integral areas of peaks and the concentrations of target compounds.For QA/QC,the cryogenic concentrator was baked after each analysis,and the GC column was also baked after analysis of every10samples.

A model49C O3analyzer(Thermo,Inc.)and model 42CTL NO x analyzer(Thermo,Inc.)were used in this study to measure the real-time concentrations of O3and NO x(NO and NO2)respectively at site B.The O3an-alyzer has a precision of1.0ppbV and lower detectable limit of1.0ppbV;the NO x analyzer has a precision of 0.4ppbV and lower detectable limit of0.4ppbV.Both analyzers were calibrated on-site in the beginning of the measurement campaign using an ozone calibrator(TEI Model49CPS)and a NO standard(National Centre for Standard Materials,China),respectively.In addition,zero-check was also performed every week for both analyzers. Raw concentration data of O3and NO x were recorded every5minutes during the whole measurement period. We then calculated the hourly mean concentrations from the retained high resolution data to further describe the air pollution in this paper.

1.3Estimation of photochemical e?ects on VOCs

around the re?nery

Concentrations of pollutants(VOCs,NO x)in the ambient air of the re?nery would be greatly in?uenced by the local photochemical reactions,due to the high concentrations of ground ozone observed in the re?nery industrial area (Chiang et al.,2007;Chiu et al.,2005;Nuria et al.,2008). Normally,the species with high reactivity decrease in daytime,and those with relatively low reactivity gradually increase due to accumulation.Thus,the chemical pro?les of VOCs would present temporal and spatial variations in a given area.Some studies have used the ratio of a

Journal of Environmental Sciences 26(2014)332–342

335

highly reactive species to a less reactive species,such as toluene /benzene,m,p -xylene /ethylbenzene,to describe photochemical reactions.The decrease of these ratios can be a measure of the age of the air mass (Liu et al.,2008b;Miller et al.,2012;Tiwari et al.,2010).Some studies have used the temporal and spatial variations of VOC concentrations to estimate their contributions to photochemical ozone formation,by combining a trajectory model with a photochemical mechanism,which assumed independent transport and chemistry processes and ignored their nonlinear coupling (Kleinman et al.,2003;Walker et al.,2007;Cheng et al.,2010;Ling et al.,2012).

Considering that this study mainly focused on a small research area,we made the following assumptions,(1)there is a single VOC emission source (the re?nery),(2)reactions only take place with OH radical,(3)non-di ?usive transport occurs by a unique well-de?ned travel route.Thus,the concentration of VOCs i and VOCs j at a downwind receptor (designed as site C)could be obtained from Eqs.(1)–(2),which was ?rst developed by Kleinman et al.(2003):

C i =

D (t )C i (0)exp (?k i [OH]t )

(1)C j =D (t )C j (0)exp

?k j [OH]t

(2)

where,i,j are the VOC compounds,D(t)is a dilution factor,C (0)is an initial concentration at the emission source,k is the reaction rate with OH radical,and t is the travel time from source to receptor.

Then,[OH]t (de?ned as the photochemical age by Kleinman,representing the level of photochemical re-actions)and D(t)(representing the level of atmospheric

di ?usion)could be deduced from Eqs.(1)–(2),shown as the Eq.(3)and (4):

[OH]t =

ln C i

(t )C i (0)

?ln

C j (t )C j (0)

k j ?k i

(3)

D (t )=exp ????????????k j ln C i (t )C i (0) ?k i ln C j (t )C j (0) k j ?k i ????????????

(4)

It can be seen that the [OH]t and D(t)can be given

relative to whatever sample is chosen as a reference point (t =0).So,the reference point need not be the emission point.In this study,we chose site B as the reference point.In addition,an acknowledged criterion of “ k i ?k j [OH]t <1”for the application of equations 3–4was widely used by the above studies,in order to avoid large estimation errors in the D(t)and [OH]t calculations.

2Results and discussion

2.1O 3concentrations measured

A high level of surface ozone concentration was observed near the re?nery in summer months,as shown in Fig.3.The mean values for daytime (8:00–18:00)and 24hr were calculated as 80.0ppbV and 52.4ppbV ,respectively.The lower levels of O 3occurred on day 1,5,7,10–15,19,

100 150 200 250 0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30

O z o n e c o n c e n t r a t i o n (p p b V )

Date

July

80 120 160 200

0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30

O z o n e c o n c e n t r a t i o n (p p b V )

Date

August

Fig.3

O 3concentrations observed at site B during the measurement period.

336Journal of Environmental Sciences26(2014)332–342

and31of July,and on day1,4–6,11,18–19,21,27of August,the days in which it rained,coincidentally.Thus, if one deletes these rain days,the mean O3concentrations for daytime(8:00–18:00)and24hr would increase to99.1 ppbV and63.0ppbV,respectively.In these non-rain days the re?nery area was exposed to high-level O3pollution, with the average daily maximum level of145.4ppbV and average unquali?ed hours per day of6.7hours(beyond the Chinese national environmental standard of93ppbV for 1-hr concentration).In total,there were about34days in which the93ppbV threshold value was exceeded,with the unquali?ed rate of54.8%.

However,the O3concentrations in the center and other rural areas of Beijing were relatively lower than those observed in the re?nery area by this study.Xu et al.(2009) measured the surface O3concentration at four stations in Beijing(Fengtai Station and Baolian Station as the urban stations located in the southwest of the Beijing city,Shunyi Station and Shangdianzi Station as the rural stations located in the northeast of the Beijing region) from June to September in2007,and found that the daily averaged O3concentration ranged from34ppbV to 44ppbV,with a spatial increasing trend along the route of the prevalent southwest wind direction.Duan et al. (2011)reported that the daily averaged O3concentration in Beijing center was about37–42ppbV in July and28–35ppbV in August in2010.Moreover,these monitors indicated that in Beijing summer,the daily maximum O3 concentrations were normally lower than100ppbV,and that the monthly number of days when the1-hr average O3 concentration was beyond the national standard(93ppbV) was about5–10days.So,compared with the average level of ozone concentration in the Beijing region,we believe that more serious photochemical pollution occurs in the re?nery area.In addition,Cheng et al.(2012)obtained the temporal and spatial distribution of O3concentration by aircraft sounding over the Beijing region from2007to 2010,and found that the area of this re?nery(116.1?E, 39.8?N)was exposed to a relatively higher O3pollution than the Beijing region as a whole,which also con?rmed our observation results.

In addition,the O3diurnal variation in the re?nery area mostly displayed an obvious multi-peak pattern(double peaks in most cases).However,the temporal distribution of these peaks was irregular.Figure4depicts the frequency of the maximum values occurring at di?erent hours.In general,the rates of occurrence of the O3daily maximum values in8:00–10:00,10:00–12:00,12:00–14:00,14:00–16:00and16:00–18:00,were counted to be9.5%,35.7%, 21.4%,23.8%,and9.5%respectively.This is greatly di?erent from the normal pattern of surface O3in the Beijing region,with the daily maximum value normally occurring at15:00.Such irregularity also revealed the local characteristic of ozone pollution in the re?nery area.We think that it might result from the variations of the

local

(

)

Fig.4Occurrence frequency of daily O3maximum values at di?erent hours.

surface wind?eld,which carry the precursor pollutants (VOCs)from the re?nery and produce a polluted air plume with a?uctuating moving route.The polluted air plume would undergo photochemical reactions while moving within the winds,so the accumulation in O3concentration at downwind site B could not present a regular diurnal variation mode.

2.2Temporal pattern of NOx concentrations measured NO x(NO x=NO+NO2)and NO concentrations were automatically measured by a model42CTL NO x analyzer, with the recording resolution of5min.Figure5presents the diurnal patterns for NO and NO2(subtracting NO from NO x)concentrations observed at site B,by using hourly concentrations averaged over the whole measure-ment during two months.It can be seen that the hourly mean concentrations of NO and NO2were respectively lower,in the range of1.8–6.4and20.5–46.1ppbV.Both reached the peak value at8:00–9:00,which was brought about by the transport of a polluted air current with high NO x concentrations from Beijing city center.Xu et al. (2009)has monitored the NO x daily mean concentration of about60ppbV in Beijing city center in the period of June–September of2007.As we know,in Beijing summertime, northern winds and southern winds prevail at night and daytime,respectively;and the change of prevailing wind direction occurs in the morning(8:00–10:00).Then,the concentrations deeply declined to about2.0ppbV for NO and to about20ppbV for NO2at noon,with decrease rates more than60%,and maintained this level till evening. 2.3Concentrations and chemical speciation of VOCs

measured

The gaseous samples for three periods(8:00–9:00,12:00–13:00,and17:00–18:00)were taken simultaneously at the three sites in both months,and analyzed by GC-MS. Finally,51compounds were found in this study.Then,the TVOCs(sum of total VOCs detected)concentrations by ppbC(V/V)were calculated,as shown in Fig.6,in which the points at one line meant the simulaneous observation at three sites on a monitor day.The chemical pro?les of VOCs were provided by way of the volume ratios of various VOCs species to n-hexane,because propylene,

Journal of Environmental Sciences 26(2014)332–342

337

0:01:02:03:04:05:06:07:08:09:010:011:012:013:014:015:016

:017:018:019:020:021:022:023:0V o l u m e c o n c e n t r a t i o n (p p b V )

Time

Fig.

Hourly mean NO x concentrations in daytime around the re?nery.

T V O C (p p b C )

Fig.6TVOCs concentrations at the three sites at di ?erent hours.The point at one line meant the simultaneous observations at three sites on a monitor city.

benzene and toluene indicated probable instability in this measurement,which would interfere in the percentage calculation of VOCs species.The reason for the choice of n -hexane was that it is one of the main emission species of an oil re?nery (about 10%,V /V )and has a middle-level of atmospheric reactivity.Figure 7describes the details of the chemical pro?les.

Monitoring data revealed that in the early morning (8:00–9:00),the three sites were all exposed to a high level of TVOCs,arriving at 600–2000ppbC at site A and 200–1000ppbC at sites B /C.The reason was that during Beijing summertime,the wind speed in the early morning is very low (about 1m /sec near the ground),which weakens the dispersion of VOCs emitted from the re?nery,and results in a larger ground concentration of VOCs around the re?n-ery.The higher TVOCs concentration at site A compared to site B /C could have originated from the prevalent winds from the northern direction.As mentioned above,NO x diurnal concentrations implied air mass transport from Beijing city center,so we believed that the TVOCs at 8:00–9:00were also in?uenced by the pollutants emitted from Beijing city center.For VOC species,chemical pro?le similarity was obtained for the three sites.

After noon,south becomes the dominant wind direction,and site A would become the background,for which TVOCs sharply declined to several dozens of ppbC.On the other hand,site B and site C became downwind from the re?nery and were greatly a ?ected by the re?nery’s VOCs emissions,for which TVOCs concentrations respectively arrived at 200–600ppbC and 100–300ppbC.For VOC species,a similar trend could still be found at site B and site C,which revealed that both sites were in?uenced by the same air mass polluted by the re?nery.However,there were some di ?erences between the sites.For the same time period,the ratios of C4–C5alkanes,whose atmospheric reactivities are lower than n -hexane,were about 10%higher at site C than those at site B;however,the ratios of C7–C9alkanes and most alkenes with higher reactivity became lower at site C than those at site B.The compounds with higher reactivity are more easily oxidized during the photochemical reaction process.So,we think that the decrease of TVOCs concentration from factory source to receptors could be attributed not only to the enhanced atmospheric dispersion,but also the local atmospheric photochemical reactions.Their contributions to the degradation of VOCs in the ambient air will be estimated in the following section.

Finally,we derived the chemical pro?les of VOCs

338

Journal of Environmental Sciences 26(2014)332–342

V o l u m e r a t i o o f V O C s t o n -h e x a n Site A

Site B Fig.7V olume ratios of main VOCs to n -hexane at the three sites.

emitted from the re?nery in Table 1,through subtracting VOCs concentrations at site A from those at site B /C at 12:00and at 17:00.It can be seen that the VOCs emitted from the re?nery are mainly composed of alkanes (60.0%±4.3%),alkenes (21.1%±5.5%)and aromatics (18.9%±3.9%),in which butanes,pentanes,C6linear and branched saturated hydrocarbons,propylene,1-butylene,benzene and toluene were particularly abundant,with the respective average mixing volume ratios of 16.5%,9.8%,19.7%,10.3%,5.1%,8.0%and 6.1%.This result is comparable with the study of Liu et al.(2008a)for most species,except for isobutane,cyclohexane and propylene.Then,com-bining with the maximum incremental reactivity (MIR),

the contributions of VOC species to ozone formation were calculated,as shown in Table 1.Ranking by ozone formation potentials,propylene,1-butylene,toluene,m,p -xylene,and trans -2-butylene were the top 5species.In general,alkenes played the most important role in ozone formation and accounted for 44.3%of total ozone forma-tion potential,followed by alkanes (29.6%)and aromatic hydrocarbons (26.1%).

2.4Estimate of the in?uence of photochemical reac-tions on VOC degradation As mentioned above,a great decrease in VOC concen-trations was found around the re?nery,which resulted

Journal of Environmental Sciences26(2014)332–342339

Main species VOCs in this study VOCs by Liu et al.(2008a)k(OH)×10?12MIR*(g O3/g VOC)Contribution too zone (V/V)(V/V)(cm3/(mol·sec))formation potential Isobutane(8.6±6.2)% 2.3% 2.1 1.07 2.4%

n-Butane(7.8±0.3)% 6.7% 2.4 1.08 2.2%

Isopentane(5.8±1.5)% 6.2% 3.6 1.36 2.6%

n-Pentane(4.1±1.2)% 6.6% 3.8 1.23 1.6%

2-Methylpentane(2.6±0.5)% 4.3% 5.2 1.41 1.4%

3-Methylpentane(2.9±1.0)% 3.8% 5.2 1.70 1.9%

n-Hexane(7.8±3.8)%8.1% 5.3 1.15 3.5% Methylcyclopentane(2.8±0.9)% 2.8% 5.7 2.06 2.2% Cyclohexane(3.6±3.0)%11.8%7.0 1.16 1.6%

2-Methylhexane(0.9±0.4)% 1.3% 6.9 1.100.4%

3-Methylhexane(1.0±0.5)% 1.1%7.2 1.510.7%

n-Heptane(1.4±0.9)% 3.2% 6.80.990.6%

Propylene(10.3±5.4)% 4.8%26.011.3722.4%

1-Butylene(5.0±3.8)% 1.9%33.19.4212.0%

trans-2-Butylene(1.4±1.8)%0.4%55.814.79 5.5%

cis-3-Butylene(1.0±1.1)%0.7%63.213.89 3.5%

Hexene(0.5±0.2)%0.0%37.0 5.280.9%

Benzene(8.0±3.9)%7.5% 1.20.69 2.0%

Toluene(6.1±3.7)% 4.3% 5.6 3.889.9% Ethylbenzene(1.2±0.5)%0.4%7.0 2.93 1.7%

m,p-Xylene(1.8±0.5)%0.7%20.59.528.2%

o-Xylene(0.8±0.3)%0.3%13.77.44 2.9% Isopropylbenzene(1.1±0.5)%0.0% 5.0 2.43 1.4%

Others(13.6±11.2)%21.0% 2.1 1.508.3%

*MIR denotes maximum incremental reactivity(Carter,2009).

from atmospheric dispersion and photochemical reactions.

So,here we tried to estimate their in?uences,by us-

ing the method as introduced in Eqs.(3)and(4).The

VOC species selected in this study included the most

dominant compounds(17species:n-butane,isopentane,

n-pentane,2-methylpentane,3-methylpentane,n-hexane,

methylcyclopentane,cyclohexane,2-methylhexane,3-

methylhexane,n-heptane,1-butylene,t-2-butylene,c-3-

butylene,ethylbenzene,m,p-xylene,and o-xylene).Their

mixing ratios were relatively high and stable in our gaseous

samples,and their reactivity range(1.0E-12to6.3E-11)

would distinguish C i(t)/C i(0)from C j(t)/C j(0).

Rather than considering pairs of VOCs,we used a set

of17VOC concentrations in a single age determination

by doing a linear regression of ln(C i(t)/C i)(0)versus k i.

The slope of the regression is–[OH]t,and the intercept is

ln(D(t)).Here,site B and site C were respectively assumed

as the reference point and the receptor point.One example

was chosen to illustrate this calculation method,presented

in Fig.8,whose regression parameters were D(t)=0.71,

[OH]t=2.8×1010(mol·sec)/cm3,and correlation of r2 =0.62.We obtained[OH]t and D(t)at0.5km distances downwind of the re?nery for each sampling day,and then

averaged these results in Table2.

It can be concluded that[OH]t reached1010

(mol·sec)/cm3.Because the transport route of air mass

Data D(t)[OH]t×1010R2 subset((mol·sec)/cm3)

12:000.62±0.23 3.12±1.690.52±0.34 17:000.42±0.11 1.67±0.610.40±0.22

from site B to site C might be non-straight,we could not obtain an accurate t based on the wind speeds(2–3 m/sec).However,t could still be estimated to be several hundred seconds,and[OH]would be calculated to be 107molec/cm3,whose value at the global scale and in the ozone-episode urban atmosphere(e.g.Hong Kong)is normally in the level of106mol/cm3and107mol/cm3, respectively(Jia et al.,2005;Zhang et al.,2007).This indicated a stronger photochemical reaction in the local area of the re?nery.[OH]t during12:00–13:00was about 0.8times higher than that during17:00–18:00,because the hydroxyl radical(OH)is controlled by the local abundances of VOCs,NO x,CO,O3,and as well as the intensity of solar UV,and varies greatly with time of day. However,higher variation existed,with relative standard deviation of more than50%.We think that this was caused partly by the di?erences in the photochemical reaction

340

Journal of Environmental Sciences 26(2014)332–342

Main species 12:00–13:0017:00–18:00Main species 12:00–13:0017:00–18:00Isobutane 0.060.04Propylene 0.560.35n -Butane 0.070.041-Butylene

0.640.42Isopentane 0.110.06trans -2-Butylene 0.820.61n -Pentane

0.110.06cis -3-Butylene 0.860.652-Methylpentane 0.150.08Hexene 0.680.463-Methylpentane 0.150.08Benzene 0.040.02n -Hexane

0.150.08Toluene

0.160.09Methylcyclopentane 0.160.09Ethylbenzene 0.200.11Cyclohexane 0.200.11m,p -Xylene 0.470.292-Methylhexane 0.190.11o -Xylene

0.350.203-Methylhexane 0.200.11Isopropylbenzene 0.140.08n -Heptane

0.19

0.11

Average

0.27

0.15

l n (C i (t )/C i (0))k (OH) (cm 3/(mol·sec))

Fig.8Regression plots of ln(C i (t )/C i (0))versus OH reaction rate constant for the determination of the photochemical age and dilution factor at 12:00-13:00on 8July.

process among these measurement days,partly by the di ?erences of t,which depended on wind speeds,and partly by the error of the calculation method.Then,the photochemical degradation rates of VOCs species,through exp(–k i [OH]t ),were calculated in Table https://www.360docs.net/doc/9d2126031.html,bining with the original chemical pro?les of VOCs emitted from the re?nery,the average photochemical degradation rate of VOCs at 0.5km distances downwind was estimated to be about 0.27at 12:00–13:00and 0.15at 17:00–18:00.Moreover,the calculation results of D(t)at both hours were also in accordance with the diurnal variation trend of wind speed,whose average hourly value was reported to be about 2.2m /sec at 12:00and 3.0m /sec at 17:00in summer of 2011.

3Conclusions

A ?eld measurement campaign for ozone and ozone pre-cursors (VOCs and NO x )was conducted in July–August 2011around a petroleum re?nery,located 40km southwest of Beijing city center.

Monitoring data revealed a high level of surface O 3concentration around the re?nery,with the average daily

mean value of 52.4ppbV ,although it rained in one third of the measurement days.In non-rain days,the average daily maximum level of O 3reached 145.4ppbV ,and the average hours per day beyond the 93ppbV threshold value was about 6.7hours,which was more serious than those of the Beijing region (average daily mean value of 30–40ppbV).Moreover,the O 3diurnal variation mostly displayed a multi-peak pattern,with an irregular temporal distribution of peaks.We believed that the O 3pollution around the re?nery presented an obvious local characteristic.The NO x measurement showed that NO 2was more abundant than NO in the re?nery area,with the average hourly concentrations of 20.5–46.1ppbV for NO 2and 1.8–6.4ppbV for NO.

For VOCs,a total of 51compounds were detected in this study,90%of which were composed of 23main compounds.Generally,TVOCs at background site A was about dozens of ppbC,while TVOCs at downwind sites B /C also reached several hundreds of ppbC.Then the chemical pro?le of the re?nery’s VOCs emission was calculated,through subtracting VOCs concentrations at background from those at downwind sites.Although a chemical pro?le similarity was observed at sites B and C,there were some di ?erences.Generally,compared with site B,the volume ratios of the alkanes with low reactivity increased,and those of the alkenes with high reactivity decreased at site C.This implied a degradation of VOCs by local photochemical reactions in daytime.

Finally,based on temporal and spatial variations of VOC concentrations,VOC photochemical degradations and dis-persion degradations were estimated to be 0.15–0.27and 0.42–0.62respectively,by calculating photochemical age ([OH]t ).[OH]t and [OH]were estimated to be 1010and 107(mol ·sec)/cm 3,indicating a stronger photochemical reaction in the local area of the re?nery.Although this estimate has some uncertainty due to the simpli?ed as-sumptions of the photochemical age calculation method,we can still obtain the following conclusions:(1)the VOCs emitted from the re?nery would rapidly participate in

Journal of Environmental Sciences26(2014)332–342341

photochemical reactions and be photochemically degraded on a small spatial scale,(2)the VOC emission from the re?nery would produce a high level of surface O3locally, and indirectly in?uence the air quality of Beijing city center through O3transport.

Acknowledgements

This work was supported by the National Natural Science Foundation of China(No.51108006),Ph.D.Programs Foundation of the Ministry of Education of China(No. 20111103120008),and the State Environmental Protection Key Laboratory of Sources and Control of Air Pollution Complex.

r e f e r e n c e s

An,J.L.,Wang,Y.S.,Li,X.,Sun,Y.,2008.Characteristics of atmo-spheric O3and NO x concentrations in the urban area,Beijing.Ecol.

Environ.17,1420–1424.

Carter,W.P.L.,2009.Updated maximum incremental reactivity scale and hydrocarbon bin reactivities for regulatory applications.California Air Resour.Board Contract07–339.

Cheng,H.R.,Guo,H.,Saunders,S.M.,Lam,S.H.M.,Jiang,F.,Wang, X.M.et al.,2010.Assessing photochemical ozone formation in the Pearl River Delta with a photochemical trajectory model.Atmos.

Environ.44,4199–4208.

Cheng,P.F.,Zhang,Q.,Quan,J.N.,Gao,Y.,Huang,M.Y.,2012.

Temporal and spatial distribution of ozone concentration by aircraft sounding over Beijing.Environ.Sci.33,4141–4150.

Chiang,H.L.,Tsai,J.H.,Chen,S.Y.,Lin,K.H.,Ma,S.Y.,2007.VOC concentration pro?les in an ozone non-attainment area:A case study in an urban and industrial complex metroplex in southern Taiwan.Atmos.Environ.41,1848–1860.

Chiu,K.H.,Sree,U.,Tseng,S.H.,Wu,C.H.,Lo,J.G.,2005.Di?erential optical absorption spectrometer measurement of NO2,SO2,O3, HCHO and aromatic volatile organics in ambient air of Kaohsiung petroleum re?nery in Taiwan.Atmos.Environ.39,941–955. Coleman,B.K.,Lunden,M.M.,Destaillats,H.,Nazaro?,W.W.2008.

Secondary organic aerosol from ozone-initiated reactions with terpene-rich household products.Atmos.Enviorn.42,8234–8245. Duan,Y.S.,Zhang,Y.H.,Wang,D.F.,Xu,J.,Wei,H.P.,Cui,H.

X.,2011.Spatial-temporal patterns analysis of ozone pollution in several cities of China.The Administration and Technique of Environ.Monit.23,34–39.

Gen,F.H.,Mao,X.Q.,Tie,X.X.,Gao,W.,Peng,L.,Xu,J.M.et al.,2010.Analysis of ozone characteristics and toward establishing ozone warning system in Shanghai,2006–2008.J.Tropical Mete-orol.26,584–590.

Hu,Y.T.,Zhang,Y.H.,Xie,S.D.,https://www.360docs.net/doc/9d2126031.html,ing the trajectory model to simulate the diurnal change of ozone concentration.Acta Scientiae Circumstantiae19,345–350.

IAPCAS(Institute of Atmospheric Physics of Chinese Academy of Science),1988.The research on surface atmospheric di?usion model in Yanshan Petrolieum Industrial Area.Adv.Earth Sci.1988, 83–84.

Jia,L.,Ge,M.F.,Zhuang,G.S.,Sun,Z.,Wang,D.X.,2005.Advances in the study of tropospheric OH and HO2.Chemistry68,736–743.Kleinman,L.I.,Daum,P.H.,Lee,Y.N.,Nunnermacker,L.J.,Springston, S.R.,Weinstein,L.J.et al.,2003.Photochemical age determi-nations in the Phoenix metropolitan area.J.Geophy.Res.108, 4096–4109.

Lamorena,R.B.,Lee,W.,2008.In?uence of ozone concentration and temperature on ultra-?ne particle and gaseous volatile organic com-pound formations generated during the ozone-initiated reactions with emitted terpenes from a car air freshener.J.Hazard.Mater.

158,471–477.

Ling,Z.H.,Guo,H.,Zheng,J.Y.,Louie,P.K.K.,Cheng,H.R.,Jiang,

F.et al.,2012.Establishing a conceptual model for photochemical

ozone pollution in subtropical Hong Kong.Atmos.Environ.76.

DOI:10.1016/j.atmosenv.2012.09.051.

Liu,Y.,Shao,M.,Fu,L.L.,Lu,S.H.,Zeng,L.M.,Tang,D.G.,2008a.

Source pro?les of volatile organic compounds(VOCs)measured in China:part I.Atmos.Environ.42,6247–6260.

Liu,P.W.,Yao,Y.C.,Tasi,J.H.,Hsu,Y.C.,Chang,L.P.,Chang,K.

H.,2008b.Source impacts by volatile organic compounds in an

industrial city of southern Taiwan.Sci.Total Environ.398,154–163.

Liu,X.H.,Zhang,Y.,Xing,J.,Zhang,Q.,Wang,K.,Streets,D.G.

et al.,2010.Understanding of regional air pollution over China using CMAQ,part II Process analysis and sensitivity of ozone and particulate matter to precursor emissions.Atmos.Environ.44, 3719–3727.

Miller,L.,Xu,X.H.,Grgicak-Mannion,A.,Brook,J.,Wheeler,A.,2012.

Multi-season,multi-year concentrations and correlations amongst the BTEX group of VOCs in an urbanized industrial city.Atmos.

Environ.61,305–315.

Nuria,C.,Enrique,M.,Millan,M.,2008.Analysis of tropospheric ozone concentration on a western Mediterranean site:Casteelon.Environ.

Monit.Assess.136,3–11.

Shao,M.,Zhang,Y.H.,Zeng,L.M.,Tang,X.Y.,Zhang,J.,Zhong,L.

J.et al.,2009.Ground-level ozone in the Pearl River Delta and the roles of VOC and NO x in its production.J.Environ.Manage.90, 512–518.

Song,Y.,Dai,W.,Shao,M.,Liu,Y.,Lu,S.,Kuster,W.et https://www.360docs.net/doc/9d2126031.html,-parison of receptor models for source apportionment of volatile organic compounds in Beijing China.Environ.Poll.156,174–183. Tang,G.,Li,X.,Wang,Y.,Xin,J.,Ren,X.,2009.Surface ozone trend details and interpretations in Beijing,2001–2006.Atmos.Chem.

Phys.9,8813–8823.

Tiwari,V.,Hanai,Y.,Masunaga,S.,2010.Ambient levels of volatile organic compounds in the vicinity of petrochemial industrial area of Yokohama,Japan.Air Quality,Atmos.Health3,65–75.

US EPA,https://www.360docs.net/doc/9d2126031.html,pendium Method TO-15:Determination of volatile organic compounds in air collected in specially prepared canisters and analyzed by gas chromatography/mass spectrometry. Walker,H.L.,Baker,J.,Derwent,R.G.,Donovan,R.G.,2007.The use of a photochemical trajectory model to estimate pollution levels within the west midlands conurbation,UK.Developments in Environ.Sci.6,774–783.

Wang,T.,Wei,X,L.,Ding,A,J.,Poon,C,N.,Lam,K,S.,Li,Y.S.et al., 2009a.Increasing surface ozone concentrations in the background atmosphere of Southern China,1994–2007.Atmos.Chem.Phys.9, 6216–6226.

Wang,X.S.,Li,J.L.,Zhang,Y.H.,Xie,S.D.,Tang,X.Y.,2009b.Ozone source attribution during a severe photochemical smog episode in

342Journal of Environmental Sciences26(2014)332–342

Beijing,China.Sci.China Ser.B:Chem.52,1270–1280. Wang,T.,Nie,W.,Gao,J.,Xue,L.K.,Gao,X.M.,Wang,X.F.et al., 2010.Air quality during the2008Beijing Olympics:Secondary pollutants and regional impact.Atmos.Chem.Phys.Discuss.10, 12433–12463.

Xu,J.,Zhang,X.L.,Zhao,X.J.,Xu,X.F.,Meng,W.,2009.In?uence of summer local circulation on the transportation of ozone from urban to the downwind area in Beijing.China Environ.Sci.29,1140–1146.

Zelm,R.,Huijbregts,M.A.J.,den,Hollander,H.A.,Jaarsveld,H.A.,

Sauter,F.J.,Struijs,J.et al.,2008.European characterization factors for human health damage of PM10and ozone in life cycle impact assessment.Atmos.Environ.42,441–453.

Zhang,J.,Wang,T.,Chamedies,W.L.,Cardelino,C.,Kwok,J.,Blake,

D.R.et al.,2007.Ozone production and hydrocarbon reactivity in

Hong Kong,southern China.Atmos.Chem.Phy.7,557–573. Zhang,Y.H.,Su,H.,Zhong,L.J.,Cheng,Y.F.,Zeng,L.M.,Wang, X.S.et al.,2008.Regional ozone pollution and observation-based approach for analyzing ozone-precursor relationship during the PRIDE-PRD2004campaign.Atmos.Environ.42,6203–6218.