Persulfate Wet Oxidation Method for the Determination of Total Phosphorus in Atmospheric Aerosols and Its Application for a Year-round Observation in Beijing

Measurement of the phosphorus concentration in aerosols in Beijing, which was a representative East Asian mega-city, was carried out. The optimum procedure for analyzing phosphorus in aerosols was found in this study. Recovery of phosphorus in environmental samples through the improved method was almost 100%. The concentration of phosphorus in TSP was 145±47 ng/m3, with a seasonal variation showing high concentrations in winter and low concentrations in summer. The concentrations of phosphorus in PM2.5 accounted for 35±6% of those in TSP, with no seasonal variations. The major source of phosphorus in aerosols in Beijing was soil dust, and additional sources of phosphorus in fine particles could be coal combustion and biomass burning.


INTRODUCTION
Phosphorus is an essential nutrient for any organism living in terrestrial and ocean ecosystems (Paytan and McLaughlin, 2007).The deposition of essential nutrients from the air to the ocean surface plays an important role for primary production in phosphorus-limited oligotrophic open oceans (Furutani et al., 2010).Deposition of aerosols, which are originally derived from terrestrial regions, is an important process for phosphorus transport to these oligotrophic regions (Furutani et al., 2010;Mahowald et al., 2008).The concentration of phosphorus in urban areas can be high; in particular, phosphorus in aerosols in East Asia can be important, as it is transported to the oligotro-phic North Pacific region (Furutani et al., 2010;Chen and Chen, 2008;Mahowald et al., 2008;Chen et al., 2008Chen et al., , 2006)).However, phosphorus concentration in aerosols in East Asian mega-cities has not been well characterized.In this study, concentrations of total phosphorus along with several kinds of elements in aerosols in Beijing, China, were measured.
There are several standardized methods for the determination of phosphorus in aqueous systems (JIS K0102, 2008;ISO 6878, 2004;Pai et al., 1990;Murphy and Riley, 1962).However, an analytical method for the determination of phosphorus in aerosols has not been well developed (Furutani et al., 2010).The persulfate wet oxidation method could be one possible solution to this problem.Another useful option for oxidative treatment is high-temperature dry combustion, which has been used for particulate phosphorus determination (Suzumura, 2008;Chen et al., 2006).In this study, an improved method for the determination of total phosphorus in aerosols was developed.

1 Development of an Improved Method for the Determination of Phosphorus in Aerosols
In this study, modifications to the persulfate wet oxidation methods reported previously have been applied (JIS K0102, 2008;ISO 6878, 2004;Pai et al., 1990;Murphy and Riley, 1962).The procedure developed in this study, which is known as phosphoantimonylmolybdenum blue complex (PD-MB) method, is described below.
A filter segment was placed in a PTFE container, and then 8 mL of 40 g/L potassium persulfate solution was added.The sample container was autoclaved for phosphate was subsequently performed.The solution was filtered by passing through a membrane filter (Advantec DISMIC, pore size: 0.2 μm).In order to develop the color of the solution for the colorimetric determination of phosphate, two solutions were prepared.Solution 1: 72 g/L L(+ +)-ascorbic acid (99.6%,Wako); solution 2: 12 g of ammonium heptamolybdate tetrahydrate [(NH 4 ) Wako), 240 mL of 78% sulfuric acid (95%, Wako), and 10 g of ammonium amidosulfate [NH 4 OSO 2 NH 2 ] (98.5%, Wako) in 1 L of ultrapure water.Immediately before the experiment, the color-development solution was mixed at a ratio of 1 : 5 of solution 1 to solution 2. Color-development solution (0.8 mL) was added to the sample filtrate, and then settling was allowed to occur for a certain amount of time (the color development time).The colorimetric determination of phosphate of the final solution at 880 nm and 1 cm optical path length was carried out using a spectrophotometer (Shimadzu UV-mini1240).High linearity of the calibration curve (r 2 = =0.9999)was obtained by analyzing ultrapure water and phosphate standard solutions (0.1, 0.2, 0.5, 1, 2 and 5 μg/mL) using the spectrophotometer.The detection limit, which was calculated by tripling the standard deviation of replicate absorbance measurements (n= =5) of ultrapure water (blank solution), was 0.07 μg/mL.The phosphate concentration in the final solution was adjusted to higher than 0.1 μg/mL (JIS K0102, 2008;ISO 6878, 2004).
In this study, optimization of the color development time was examined.The standard reference material for aerosol (NIES CRM#28, Mori et al., 2008), which was collected at Beijing, China, was used in the experiment.Phosphorus recovery was determined using 1-5 mg of CRM#28.Furthermore, several reference materials were also used for testing the phosphorus recovery by the procedure developed in this study.

2 Aerosol Collection
Total suspended particles (TSP) and PM 2.5 (particulate matter less than 2.5 μm) were collected using cellulose nitrate membrane filters (0.8 μm of pore size, Millipore AAWP04700) on the rooftop of a building (5 m above the ground) at Tsinghua University.The site was located 15 km northwest of the center of Beijing city, China (Okuda et al., 2013a(Okuda et al., , b, 2011(Okuda et al., , 2008)).This site could be considered a good representative of the entire area of Beijing city in terms of the concentrations of particulate matter (Okuda et al., 2004).A low volume air sampler (Tokyo Dylec Corp.) was operated at 5 L/min for 1 week to collect aerosol samples.

3 Energy-dispersive X-ray Fluorescence
Spectrometry (EDXRF) All elements (except for phosphorus) on the filter samples without pretreatment were analyzed by EDXRF using an EDXL300 spectrometer (Rigaku Corp., Japan).Quantification of each element in aerosol samples was performed using the fundamental parameter (FP) method Rigaku Profile Fitting -Spectra Quant X (RPF-SQX).Thirteen elements (Al, S, K, Ca, Ti, V, Cr, Mn, Fe, Ni, Cu, Zn, and Pb) were measured in this study.A detailed procedure for the multi-elemental analysis is described elsewhere (Okuda and Hatoya, 2013;Okuda et al., 2013a).Phosphorus concentrations could be measured by EDXRF; however, in this study they were undetectable through this method since the concentrations were almost always below the detection limit.

1 Optimization of the Color Development
Time for the Phosphorus Determination The color development time for phosphorus determination using the persulfate wet oxidation-phosphoantimonylmolybdenum blue complex method (PD-MB method) was set at 15 min (JIS K0102, 2008) or 10-30 min (ISO 6878, 2004).However, we found that the absorbance of the sample solution continued to increase even after 30 min had passed since the beginning of color development.This increase was possibly due to the considerably lowered rate of reduction of the antimony phosphomolybdate complex caused by ascorbic acid at high concentration of protons provided by potassium persulfate (Pai et al., 1990).Thus, the color development time for the PD-MB method was optimized.
The standard reference material for aerosol (NIES CRM#28), which had a reference value of total phosphorus (0.145 wt%), was subjected to the PD-MB method.The absorbance of the sample solution corresponding to elapsed time since the color-development solutions were added to the sample solution was recorded.Ambient temperature ranged from 20 to 28� C during the experiments (n= =5).Recovery of total phosphorus was calculated on the basis of the absorbance (Fig. 1).The recovery of phosphorus was significantly lower than 100% when the color development time was 15 or 30 min.The recovery was stable at approximately 95% after 60 min.Therefore, it was decided that the color development time was 60 min in this study.The recovery of phosphorus in CRM#28 using the fixed method was 95±3% (n= =5).Note that the PD-MB method may cause lower recovery for phosphorus in clay minerals (Suzumura, 2008).Despite this potential weakness, total phosphorus was successfully measured through the PD-MB method developed in this study since this method showed high recovery of phosphorus (95±3%) for CRM#28, which was urban aerosol material collected at Beijing.The high recovery of phosphorus remained consistent when the amount of CRM#28 was increased from 1.22 mg to 14.7 mg.
Higher concentrations of silicate may cause a positive artifact for phosphorus determination by the PD-MB method (ISO 6878, 2004).However, a solution of silicate and phosphate (Si: 500 μg/mL, P: 0.5 μg/mL) showed almost identical absorbance to the solution with the same concentration of phosphorus (0.5 μg/ mL) but without silicate.Therefore, interference by silicate could be ignored when the PD-MB method was used in this study.

2 Recovery of Phosphorus of Various
Reference Materials Recovery of phosphorus from various reference materials was examined using the fixed PD-MB method.In this study, seven types of reference material (Geochemical Reference Samples, provided by National Institute of Advanced Industrial Science and Technology, Japan (AIST, 2013), JB-1b (P 2 O 5 content, 0.255%), JH-1 (0.099%), JSd-1 (0.122%), JA-2 (0.146%), JB-3 (0.294%), JG-1a (0.083%), and JGb-1 (0.056%)) were analyzed using this method.All of the reference materials were igneous rocks except for JSd-1 (sedimentary rock).ATP (adenosine 5′-triphosphate disodium salt tryhydrate) was also examined as a surrogate for an organic form of phosphorus.The results showed that the recovery of phosphorus from geochemical reference samples was 102±8% (n= =7), and that of ATP was 96±2% (n= =4).Therefore, the PD-MB method could be applied to determine the total phosphorus content in various environmental samples.

3 Measurement of the Concentrations of
Phosphorus in Aerosols in Beijing, China According to previous studies, the concentration of phosphorus in aerosols could be too low to obtain analytical results by the PD-MB method using a single filter sample (Furutani et al., 2010;Luo et al., 2010;Chen et al., 2006).Hence, several samples were analyzed together in order to obtain a sufficient amount of phosphorus for detection using the PD-MB method.We divided 41 samples into 4 periods, namely, 1Q: January to March (n= =10), 2Q: April to June (n= =8), 3Q: July to September (n= =12), and 4Q: October to December (n= =11).The results are shown in Fig. 2. The concentration of phosphorus in TSP was 145± 47 ng/m 3 , with a seasonal variation showing high concentrations in winter and low concentrations in summer.A similar trend was observed at Lake Taihu, near Shanghai, located approximately 1,000 km south of Beijing (Luo et al., 2010).The concentration of phosphorus in Beijing was one order of magnitude higher than that on the North Pacific Ocean (Furutani et al., 2010), and it was several times higher than that in Taiwan (Chen et al., 2008).The concentrations of phosphorus in PM 2.5 accounted for 35±6% of those in TSP, with no seasonal variations.These results are at a level similar to those found in a previous report for East Asian aerosols (Beijing, Hong Kong, Cheju, and Sado Island, Zhang et al., 2010;Cohen et al., 2004) and that for the North Pacific (Furutani et al., 2010).It was found that less than 50% of phosphorus in Beijing existed in PM 2.5 .

4 Possible Sources of Phosphorus in
Aerosols in Beijing, China Possible sources of phosphorus in Beijing aerosols for coarse particles (TSP-PM 2.5 ) and fine particles (PM 2.5 ) are discussed separately.The enrichment factors (EFs) for each element in both fine and coarse modes were calculated.In this study, the EF of phosphorus is defined as EF= =(P/Fe) aerosol /(P/Fe) crust , where (P/Fe) aerosol is the concentration ratio of phosphorus to Fe (the reference element) in the aerosol, and (P/Fe) crust is the concentration ratio of P to Fe in continental crust (Mason and Moore, 1982).Given that crustal sources are the primary source of Fe and Fe is stable (not altered) in the atmosphere, this element was chosen as the reference element.The results of the EF calculation are shown in Fig. 3. Element concentrations measured in this study are shown in Tables 1 and 2. The EFs of elements in TSP observed in this study were similar to those reported in a previous study (Okuda et al., 2013b).
The EF of phosphorus in coarse particles is 1.5± 0.3, which is similar to the EFs of Al, Ca, and Ti, ele-ments that are generally considered to be of crustal origin.Therefore, it is reasonable that phosphorus in coarse particles originated from soil dust.This can also explain the seasonal variation in phosphorus concentration, which was high in winter and low in summer, since soil dust transport would be more active in winter because of the Asian monsoon.On the other hand, the EF of phosphorus in fine particles was 4.3 ±1.4,which was higher than that in coarse particles.This value is also different from the EFs of crustal   elements in fine particles.This EF value suggests that the phosphorus in fine particles were mainly soil dust, but additional sources should also be considered.The correlation coefficients between concentrations of phosphorus and other elements for each season were calculated.In fine particles, the highest correlation was observed between phosphorus and copper (r 2 = = 0.97).The correlations between phosphorus and zinc (r 2 = =0.69),and phosphorus and lead (r 2 = =0.83) in fine particles are also high.On the contrary, the correlations among these elements in coarse particles were low (P-Cu: r 2 = =0.07;P-Zn: r 2 = =0.02;P-Pb: r 2 = =0.01).A possible source of copper, zinc, and lead could be coal combustion (Okuda et al., 2008).A previous study suggested that coal combustion would be a significant source of phosphorus in aerosols near Shanghai (Luo et al., 2010).Therefore, one of the possible sources of phosphorus in fine particles in Beijing could be coal combustion.Another possible source of phosphorus is biomass burning, since the correlation between phosphorus and potassium, which is often considered a tracer for biomass burning, is high (r 2 = =0.85).Gas phase phosphorus such as PH 3 may also contribute to the aerosol-phase phosphorus through its uptake, nucleation and oxidation (Furutani et al., 2010;Zhu et al., 2007).
The coastal or open ocean ecosystem can be greatly affected by the atmospheric phosphorus transported from the terrestrial region (Furutani et al., 2010;Mahowald et al., 2008).However, the urban environment itself does not seem to play a significant role to add extra amount of phosphorus, since the major source of phosphorus in aerosols is still soil dust even in an urban environment, such as Beijing city.

CONCLUSIONS
Measurement of the phosphorus concentration in aerosols in Beijing, which was a representative East Asian mega-city, was carried out.The optimum procedure for analyzing phosphorus in aerosols was found in this study.Recovery of phosphorus in environmental samples through the improved method was almost 100%.The concentration of phosphorus in TSP was 145±47 ng/m 3 , with a seasonal variation showing high concentrations in winter and low concentrations in summer.The concentrations of phosphorus in PM 2.5 accounted for 35±6% of those in TSP, with no seasonal variations.The major source of phosphorus in aerosols in Beijing was soil dust, and additional sources of phosphorus in fine particles could be coal combustion and biomass burning.

Fig. 1 .Fig. 2 .
Fig. 1.Phosphorus recovery for NIES CRM#28 corresponding to time elapsed since the addition of color-development solutions to the sample solution.

Table 1 .
The concentrations of elements in aerosols in Beijing, China in 2008.
a All elements were determined by EDXRF except for P, which was determined by the phosphoantimonylmolybdenum blue complex method.SD means the standard deviation for each season (n= =4).b Elemental composition of continental crust was cited from Mason and Moore, 1982.c Coarse particles mean [TSP]-[PM 2.5 ]. d Fine particles mean [PM 2.5 ].

Table 2 .
Phosphorus and Elements in Beijing Aerosols173 The concentrations of elements in aerosols in Beijing, China in each season in 2008.