Seasonal characteristics of chemical compositions of the atmospheric aerosols collected in urban seaside area at Tokaimura, eastern central Japan

Abstract

[1] To obtain the seasonal characteristics and the size distributions of chemical compositions of the atmospheric aerosols in urban seaside area of eastern central Japan, size-separated aerosol samples were collected at Tokaimura (36.27°N, 140.36°E) using an Andersen type sampler during July 2002 to July 2003. A maximum mass concentration of aerosols (about 50 μg/m³) during April and May and a minimum one (about 13 μg/m³) in January and a size distribution with two peaks at 0.43–1.5 μm and 3–8 μm were observed. The size-separated aerosols were divided into water-soluble and water-insoluble fractions, and Na⁺, Cl⁻, NH₄⁺, NO₃⁻, SO₄²⁻, Mg²⁺, Ca²⁺, and Pb²⁺ in water-soluble fraction and Na, Al, Ca, Mg, Fe, Pb, Th, and U in insoluble fractions were analyzed. Then, the seasonal variations and the size distribution in mass concentration of each chemical composition and the chemistry of sea-salt particles were also studied in detail. Our results showed that (1) considerable amount of soil-derived particles was supplied from dust storm incident in China, (2) there was a loss of chlorine in sea-salt particles and this loss was compensated with the replacement by SO₄²⁻ and NO₃⁻, (3) the size distribution of Th and U showed a bimodal curve with two peaks at 0.43 μm and bigger than 3 μm and these two elements in the atmosphere seem to have two sources, and (4) aerosols contained both soluble and insoluble Pb and the soluble one became much lower in the colder seasons. The concentrations of soluble Pb showed a close relationship with anions such as Cl⁻ and SO₄²⁻.

1. Introduction

[2] Atmospheric aerosols (included dust) play an important role in several questions related to phenomena ranging from local to global scale, such as climate change, human health and geochemical mass cycles [Houghton et al., 2001; Iwasaka et al., 1988; Tegen and Fung, 1995; VäKevä et al., 2000]. Therefore understanding of detailed atmospheric aerosols chemistry is essential when investigating the cycle and radiative impacts of various aerosols on the global climate system. In many of the environmental questions related to aerosols, the information about size distributions and chemical compositions is of central interest. The size of the particles controls strongly the dynamics of the aerosols population, and the measurements of chemical composition of aerosols are desirable for understanding the transport of chemical constituents and the effects on human health and environment of aerosols. In Japan, the size distributions and the chemical characteristics of atmospheric aerosols are considered to be varying remarkably with season and location, since aerosols in some areas may be affected by dust storm incidences in China [Tanaka et al., 1990; Yabuki et al., 2002; Zhang et al., 1993]. The aim of this study is to investigate the seasonal variation of aerosols concentration, the size distribution of aerosols, the seasonal characteristics and the size distribution of each chemical composition of atmospheric aerosols, and the chemistry of sea-salt particles in Tokaimura, an urban seaside city on the east coast of the central Japan, during July 2002 to July 2003. We also hope to provide our data as a database on the aerosols in an urban seaside area of Japan, to help us investigate the mechanisms controlling the size distribution and the chemical compositions of aerosols, and further understand the impact of aerosols on global climate system.

2. Sampling Site and Samples

[3] Tokaimura (36.27°N, 140.36°E) is an urban city located on the Pacific side of the central Japan (Figure 1), which is approximately 120 km away from Tokyo. Many facilities related to atomic energy are situated in this city.

Figure 1

Sampling site, Tokaimura, an urban seaside city located on the Pacific side of central Japan.

[4] An Andersen-type air sampler (AN-200, Sibata Scientific Technology Ltd., Japan) was used to collect aerosol samples in this study. The sampler was set at an altitude of about 10 m, on the roof of a building in Japan Atomic Energy Research Institute (JAERI), in Tokaimura, Ibaraki, Japan. As shown in Figure 2, the Andersen air sampler consists of eight stages and one backup stage, and aerosols of different sizes could be collected on different stages separately depending on their size. Aerosol samples were recovered once every month during July 2002 to July 2003, with each length of sampling time being 25 days and a flow rate being 28.3 L/min. The filters used in the experiments was a Polyflon filter (Advantec Corporation, PF050) on stages 1 to 8 and a quartz filter (Tokyo Dylec Corporation, 2500QAT-UP) with a collection efficiency of 99.99% for 0.3 μm sized particles on the backup stage.

Figure 2

A schematic diagram of Andersen sampler AN200.

3. Analytical Method

[5] A quarter of each filter on which aerosols were collected was used for the measurement of the water-soluble and insoluble components of aerosols. First, the filter was put into a Teflon vessel which can be sealed with a screw cap, and 10 ml of Milli-Q water was added into the vessel. The Teflon vessel was sonicated for 30 min. The water was, then, separated from particles by filtration using a 0.1 μm membrane filter, and the filtrate was stored into a clean polyethylene bottle and was used for the determination of water-soluble ions. The residue together with the filter was completely decomposed using 3 mL of HNO₃-HF-HClO₄ (1:1:1) acid mixture at 160°C in the Teflon vessel sealed with a screw cap. The acid solution in the vessel was evaporated to dryness by heating it gently and the residue was dissolved again using 15 ml of 5% HNO₃ solution, and the solution was used for the determination of water-insoluble ions. All acid reagents used in this study were of ultrapure grade, purchased from Kanto Chemicals Co. Ltd., and water used was Milli-Q water.

[6] The concentrations of Na, Mg, Al, Ca, and Fe in both the water-soluble and insoluble fractions of aerosols were determined using inductively coupled plasma atomic emission spectrophotometry (ICP-AES, Hitachi P-4000, Japan) using standard addition technique. The concentrations of Pb, Th, and U were determined using an inductively coupled plasma mass spectrophotometer (ICP-MS, Yokogawa-HP-4500, Japan). The concentrations of NH₄⁺, Cl⁻, NO₃⁻, and SO₄²⁻ ions in the water-soluble fraction were determined with ion chromatography (Yokogawa IC-7000, Japan). The filter blanks were obtained following the same procedures using unused filters.

4. Results and Discussion

4.1. Seasonal Characteristics of the Concentration and Size Distribution of Aerosol Collected in Tokaimura

[7] The mass concentrations of different-sized atmospheric aerosols collected in Tokaimura and their chemical components are summarized in Tables 1–3. The seasonal variation of the aerosol concentrations and their size distribution are shown in Figure 3.

Figure 3

Mass concentrations of atmospheric aerosols in Tokaimura during July 2002 to July 2003. (a) Seasonal variation of mass concentrations of total aerosols in Tokaimura. (b) Size distribution of aerosols in Tokaimura in each month.

Table 1. Mass Concentrations of Atmospheric Aerosols Collected by an Andersen Sampler in Tokaimura

Size, μm	Aerosol Concentrations, μg/m^3
	July 2002	Aug. 2002	Sept. 2002	Oct. 2002	Nov. 2002	Dec. 2002	Jan. 2003	Feb. 2003	March 2003	April 2003	May 2003	June 2003	July 2003
>11	4.47	4.66	4.69	2.14	3.59	2.40	0.86	2.30	3.90	4.69	8.76	2.93	2.90
7.0–11	3.07	3.64	3.92	3.20	2.64	2.22	1.25	1.47	1.85	3.99	4.84	1.73	3.69
4.7–7.0	4.80	4.37	6.84	4.18	3.51	2.80	1.37	3.26	2.90	6.75	6.90	5.36	5.05
3.3–4.7	4.57	4.72	6.63	4.34	3.52	2.85	1.61	3.39	2.59	6.51	5.81	4.68	5.45
2.1–3.3	3.94	2.77	3.98	3.21	2.80	2.32	0.67	2.68	1.90	4.37	3.48	3.27	3.49
1.1–2.1	2.87	2.32	2.60	2.86	1.85	2.64	1.47	2.41	1.78	2.31	3.64	4.23	2.94
0.65–1.1	3.48	3.37	3.14	4.83	2.94	4.35	2.87	4.29	3.14	4.16	6.50	6.44	2.83
0.43–0.65	3.37	3.61	2.88	4.26	3.16	4.50	1.89	4.08	2.98	4.32	5.15	4.66	2.48
<0.43	3.45	2.99	3.43	3.51	3.11	4.16	1.56	3.25	2.21	3.61	4.30	3.58	2.41
>2.1	20.85	20.17	26.05	17.07	16.06	12.60	5.76	13.11	13.14	26.31	29.78	17.97	20.59
<2.1	13.16	12.29	12.05	15.46	11.06	15.64	7.79	14.03	10.11	14.40	19.60	18.92	10.66
Total	34.01	32.45	38.11	32.53	27.12	28.24	13.54	27.14	23.25	40.71	49.38	36.89	31.25

Table 2. Concentrations of Ions in the Soluble Fraction of Aerosols Collected by Andersen Sampler in Tokaimura^a

Size, μm	2002-8 Sample								2002-10 Sample
	Cl^−	NO3^−	SO4^2−	NH4^+	Na^+	Mg^2+	Ca^2+	Pb^2+	Cl^−	NO3^−	SO4^2−	NH4^+	Na^+	Mg^2+	Ca^2+	Pb^2+
>11	545.2	163.7	141.2	-	279.2	38.8	19.6	nm	376.5	103.6	87.3	-	197.2	26.0	52.7	-
7.0–11	641.5	230.1	205.7	-	423.9	47.6	32.2	nm	683.3	187.6	171.4	-	367.1	39.1	62.5	-
4.7–7.0	832.4	367.6	250.0	-	517.1	65.4	33.4	nm	767.9	367.2	236.6	-	442.6	50.1	77.0	-
3.3–4.7	794.4	539.8	356.0	-	521.0	76.3	43.0	nm	510.4	440.6	198.8	-	342.6	41.1	71.0	0.014
2.1–3.3	534.4	535.9	766.1	117.0	476.6	76.8	37.7	nm	273.4	404.8	249.0	26.4	249.6	32.1	63.2	0.034
1.1–2.1	19	13.5	1336	299.9	157.5	30.7	30.1	nm	12	64.3	818.8	160.3	83.8	14.3	28.3	0.376
0.65–1.1	-	-	2496	736.6	64.9	11.2	11.9	nm	-	69.3	1862	540.7	22.3	5.5	11.4	0.828
0.43–0.65	-	-	1526	523.8	-	-	12.4	nm	-	57.4	1243	437.9	-	-	4.7	0.901
<0.43	-	-	1002	352.3	-	-	-	nm	-	30.2	581.1	263.3	-	-	-	0.717
Total	3367	1851	8079	2030	2440	347	220		2623	1725	5448	1429	1705	208	371	2.9
Size, μm	2003-1 Sample								2003-3 Sample
	Cl^−	NO3^−	SO4^2−	NH4^+	Na^+	Mg^2+	Ca^2+	Pb^2+	Cl^−	NO3^−	SO4^2−	NH4^+	Na^+	Mg^2+	Ca^2+	Pb^2+
>11	64.4	21.6	22.4	-	34.5	4.1	18.5	-	185.9	61.1	67.3	-	98.7	11.8	31.5	-
7.0–11	98.5	41.1	44.8	-	53.7	6.0	26.0	-	301.2	82.5	97.6	-	160.8	16.8	34.5	-
4.7–7.0	150.3	78.6	63.1	-	79.9	9.3	32.4	-	538.3	233.4	189.5	-	286.3	33.8	62.7	0.004
3.3–4.7	139.5	152.4	94.2	-	82.6	10.8	38.0	-	360.4	283.5	159.3	-	221.1	26.0	54.1	0.009
2.1–3.3	33.8	91.8	80.4	14.9	41.7	5.1	18.1	0.010	191.5	362.3	243.2	27.7	162.6	23.1	52.4	0.062
1.1–2.1	4	79.8	227.6	111.3	23.2	3.3	11.1	0.048	26	149.2	391.3	114.2	55.3	8.4	17.3	0.144
0.65–1.1	-	253.5	700.8	363.7	11.3	1.8	6.1	0.152	17	219.3	1169	411.8	23.2	3.8	7.6	0.381
0.43–0.65	-	193.4	400.7	248.0	-	-	-	0.121	-	175.0	977.6	413.2	-	-	3.6	0.339
<0.43	-	138.7	410.5	263.4	-	-	-	0.271	-	76.0	583.2	306.8	-	-	2.1	0.349
Total	491	1051	2045	1012	327	40.4	150	0.6	1621	1642	3878	1274	1008	124	266	1.3
Size, μm	2003-5 Sample								2003-7 Sample
	Cl^−	NO3^−	SO4^2−	NH4^+	Na^+	Mg^2+	Ca^2+	Pb^2+	Cl^−	NO3^−	SO4^2−	NH4^+	Na^+	Mg^2+	Ca^2+	Pb^2+
>11	610.3	137.0	148.7	-	332.6	36.6	52.8	0.037	404.5	102.7	140.6	-	236.4	27.2	28.2	-
7.0–11	877.0	169.4	200.5	-	507.6	47.7	50.7	0.017	497.2	120.8	175.4	-	297.6	29.9	26.8	-
4.7–7.0	1620.0	564.6	433.9	38.1	1014	98.5	104.2	0.052	831.0	279.6	391.8	-	498.8	52.0	48.3	-
3.3–4.7	1182.7	939.9	553.6	56.8	1018	96.7	121.0	0.124	518.6	361.9	347.1	-	396.0	39.6	39.5	-
2.1–3.3	301.0	533.1	576.1	128.1	446.1	42.6	60.1	0.199	193.4	284.9	575.9	61.0	259.3	29.4	34.8	0.166
1.1–2.1	53	176.3	771.9	303.7	132.4	13.3	18.1	0.436	20	48.1	998.1	223.8	110.3	13.9	18.4	0.349
0.65–1.1	17	147.0	1646	584.0	36.0	4.0	5.5	0.759	-	3.3	1533	466.6	23.9	4.8	7.8	0.771
0.43–0.65	-	254.5	3208	1216.6	-	-	8.8	1.864	-	-	1046	392.1	-	-	3.4	0.893
<0.43	-	30.0	748.0	380.6	-	-	2.3	0.887	-	14.0	584.5	234.2	-	-	-	0.460
Total	4661	2952	8287	2698	3487	339	423	4.4	2465	1215	5793	1378	1823	197	249	2.6

• a Concentration values are in ng/m³; nm is not measured in this study. A dash indicates the concentration was lower than detection limit and could not be determined.

Table 3. Concentrations of Ions in the Insoluble Fraction of Aerosols Collected by Andersen Sampler in Tokaimura^a

Size, μm	2002-8 Sample								2002-10 Sample
	Na	Mg	Al	Ca	Fe	Pb	Th	U	Na	Mg	Al	Ca	Fe	Pb	Th	U
>11	20.9	16.1	35.7	11.6	56.2	nm	nm	nm	10.7	5.6	15.5	13.9	42.4	0.520	0.055	0.002
7.0–11	14.1	14.5	93.1	13.4	82.4	nm	nm	nm	13.1	9.0	46.8	9.0	37.9	0.336	0.049	0.002
4.7–7.0	19.9	14.8	90.4	10.3	87.0	nm	nm	nm	17.6	7.9	27.0	9.3	44.9	0.272	0.003	0.001
3.3–4.7	17.5	13.4	73.1	9.3	83.4	nm	nm	nm	16.0	11.5	48.1	8.3	49.5	0.282	0.008	0.002
2.1–3.3	10.8	10.3	49.0	5.1	65.9	nm	nm	nm	10.7	8.2	30.9	5.9	45.7	0.440	0.002	0.002
1.1–2.1	5.4	1.7	6.5	-	12.8	nm	nm	nm	4.7	-	4.0	-	12.8	0.574	0.014	-
0.65–1.1	3.8	-	-	-	4.6	nm	nm	nm	1.2	-	-	-	3.0	1.152	-	-
0.43–0.65	-	-	-	-	2.1	nm	nm	nm	-	-	-	-	1.9	0.815	-	-
<0.43	-	-	-	-	-	nm	nm	nm	-	-	-	-	-	0.610	0.022	0.021
Total	93	71	348	50	395				74	42.3	172	46	238	5.0	0.2	0.03
Size, μm	2003-1 Sample								2003-3 Sample
	Na	Mg	Al	Ca	Fe	Pb	Th	U	Na	Mg	Al	Ca	Fe	Pb	Th	U
>11	7.9	6.2	43.3	14.9	29.6	0.237	0.023	0.003	12.3	11.1	61.7	27.3	52.9	0.542	0.058	0.003
7.0–11	13.2	8.8	51.8	17.6	34.4	0.388	0.009	0.002	9.9	7.3	43.2	9.1	34.3	0.181	0.014	-
4.7–7.0	10.0	11.8	61.6	16.7	49.3	0.556	0.008	0.003	17.6	7.1	22.1	11.0	47.8	0.223	0.006	-
3.3–4.7	9.1	6.1	34.7	9.6	39.6	0.787	0.016	0.002	17.4	10.1	45.7	9.6	43.1	0.173	0.005	-
2.1–3.3	5.3	3.5	20.2	6.6	18.2	1.335	0.008	0.004	15.3	9.6	42.1	9.7	39.6	0.328	-	-
1.1–2.1	3.6	2.0	5.0	3.9	11.0	1.396	-	-	4.6	2.7	10.5	3.9	12.5	0.546	-	-
0.65–1.1	3.6	0.9	-	2.5	5.1	1.330	-	-	4.4	1.2	-	2.5	6.7	1.071	-	-
0.43–0.65	-	-	-	2.1	-	0.717	-	-	-	-	-	2.7	-	0.761	-	-
<0.43	-	-	-	-	-	1.085	0.016	0.024	-	-	-	-	-	0.667	0.059	0.047
Total	53	39	217	74	187	7.8	0.1	0.038	82	49	225	76	237	4.5	0.13	0.050
Size, μm	2003-5 Sample								2003-7 Sample
	Na	Mg	Al	Ca	Fe	Pb	Th	U	Na	Mg	Al	Ca	Fe	Pb	Th	U
>11	21.9	13.1	73.8	21.3	53.9	0.331	0.060	0.007	8.4	3.3	35.4	9.5	46.4	0.368	0.031	0.001
7.0–11	21.7	11.6	64.6	12.4	49.1	0.245	0.015	0.003	10.4	2.7	44.7	6.5	40.4	0.238	0.012	-
4.7–7.0	31.6	13.2	55.5	15.0	63.5	0.401	0.027	0.003	14.3	4.5	36.2	5.0	45.8	0.289	0.012	-
3.3–4.7	84.3	25.2	83.4	24.3	74.8	0.500	0.016	0.003	13.4	3.8	29.8	4.7	43.1	0.371	0.004	-
2.1–3.3	27.4	14.5	55.3	12.7	47.5	0.566	0.008	0.002	8.8	2.5	24.4	3.5	37.5	0.446	-	-
1.1–2.1	8.1	4.2	10.7	4.4	14.7	0.366	-	-	7.8	1.5	1.9	3.1	13.0	0.780	-	-
0.65–1.1	4.9	-	-	3.1	4.2	0.555	-	-	3.1	-	-	2.9	4.2	0.899	-	-
0.43–0.65	8.7	-	-	5.1	-	1.384	-	-	-	-	-	3.2	-	0.726	-	-
<0.43	-	-	-	-	-	0.431	0.011	0.013	-	-	-	-	-	0.477	0.015	0.026
Total	209	82	343	98	308	4.8	0.13	0.021	66	18	172	38	230	4.6	0.1	0.027

• a Concentration values are in ng/m³; nm is not measured in this study. A dash indicates the concentration was lower than detection limit and could not be determined.

[8] Some characteristic features are observed in the seasonal variation and the size distribution. The mass concentrations of aerosols during April to Oct. were higher than those during Nov. to Mar., and highest in Apr or May (Figure 3a), as were observed in other areas in Japan [Yabuki et al., 2002; Kanayama et al., 2002]. Dividing aerosols into submicron (<2.1 μm) and supermicron (>2.1 μm) particles, the seasonal variation in the mass concentrations of supermicron particles was more conspicuous than that of submicron particles (Table 1). Compared with the results observed in Wako, another urban city located in Kanto district of Japan (see Figure 1) [Yabuki et al., 2002], the mass concentrations of atmospheric aerosols in Tokaimura in the same season were lower; those of supermicron particles are comparable, but those of submicron particles are higher in Wako by about 30% to 100% than in Tokaimura in the same season. This may be due to heavier traffic in Wako than in Tokaimura.

[9] Figure 3b shows the size distribution of aerosols in Tokaimura during July 2002 to July 2003. Two peaks, one at the range 0.43–1.5 μm and the other at 3–8 μm, were observed. A overwhelming peak in the range bigger than 4 μm was observed in April and May, and a relatively high peak at 0.43–1.5 μm was observed in summer. The size distribution pattern is similar to that of Wako [Yabuki et al., 2002], although the mass concentrations of aerosols in these two cities are different. Generally, soil-derived aerosols have a coarser size and anthropogenic secondary aerosols have a finer size [Houghton et al., 2001], and dust storm incident was reported to supply with soil-derived particles of bigger than 4 μm to the atmosphere [Kanayama et al., 2002; Yabuki et al., 2002]. Our analytical results shown in Tables 2 and 3 also indicate that most particles with the size bigger than 3 μm are soil-derived particles and particles with size smaller than 0.43 μm are secondary anthropogenic particles (carbonaceous and ammonium sulfate/nitrate particles), and much higher concentrations of Al and Fe (Table 3) were observed in April and May. The phenomenon that dust storm transported great amount of soil-derived particles to Japan have been reported by researchers [Inoue et al., 1995; Iwasaka et al., 1983; Yabuki et al., 2002]. Kosa particles are sometimes observed even in the Pacific Ocean [Blank et al., 1985; Tsunogai et al., 1985]. Our results indicate that the pacific side of Japan might be supplied with dust from dust storm incident in China.

4.2. Seasonal Variation and the Size Distribution of Chemical Compositions of Aerosols

4.2.1. Na⁺ and Cl⁻ Ions

[10] Na⁺ has been determined in both water-soluble and water-insoluble forms, but less than 5% of it was contained in the insoluble fraction (Tables 2 and 3). The size distribution of Na⁺ and Cl⁻ in the soluble fraction is similar to each other (Figure 4b), indicating that they suspended in the atmosphere as sea-salt particles. Assuming that the soluble Na⁺ comes from sea-salt particles and the insoluble one from soil-derived particles, our data in Tables 1 and 2 show that the percentage of the sea-salt particles in total aerosol were 18% (July) to 7% (January). The size distributions of Na⁺ and Cl⁻ has a peak at 4–9 μm, indicating that most of sea-salt particles suspended in the atmosphere were with a size of 4–9 μm and almost no sea-salt particles with a submicron size (<2.1 μm) in the atmosphere of Tokaimura (Figure 4b). Compared with sea-salt particles in the Antarctic (2–4 μm) [Kerminen et al., 2000], those in Tokaimura seem to have a bigger size. These should be due to the different meteorological conditions and the different sampling distance from coast between Tokaimura and Antarctic, since meteorological conditions and sampling distance from coast are important factors determine aerosol number, size and composition [VäKevä et al., 2000].

Figure 4

Chemical components in water-soluble fraction of aerosols. (a) Seasonal variations of bulk concentrations of each ion ([M]_{total soluble} = [M]_soluble). Units are μg/m³ except for the data of Pb, whose unit is ng/m³. (b) Size distribution of each component in the soluble fraction. Data plotted are average concentrations of the results shown in Table 2. Units are μg/m³ except for the data of Pb, whose unit is ng/m³.

[11] The loss of chlorine from sea-salt particles was also observed in Tokaimura. Brewer [1975] reported that the concentration of chloride (in mass units) in fresh sea-salt particles is equal to 1.82 times that of sodium. Table 4 shows the ratios of chlorine to sodium for sea-salt particles collected on stages 1 to 5 where enough amount of sea-salt particles were collected for accurate determination of Cl⁻ and Na⁺. From Table 4, the amount of Cl⁻ is always relatively lesser than that expected for sea-salt from Na⁺ and the discrepancy increased with decreasing size. This is considered to be due to so-called “chlorine loss” [Kerminen et al., 2000]. For stage 1 to 3 (with a size bigger than 4.7 μm), only several percent of chlorine loss have been observed. Whereas, the average chlorine loss was about 25% for the sea-salt particles collected on stage 4 (with a size of 3.3–4.7 μm) and about 50% for those collected on stage 5 (with a size of 2.1–3.3 μm). The replacement of particulate chloride by strong acidic compounds such as SO₄²⁻ and NO₃⁻ was suggested as a main reason responsible for the chlorine loss [Kerminen et al., 2000; Wagenbach et al., 1998]. Good stoichiometrical relationship between the chlorine loss and the excess of sulfate/nitrate ion (Figure 5), which is discussed in the next section, endorses the suggested mechanism for the chlorine loss.

Figure 5

The relationship between chlorine loss and excess of sulfate ions or excess of sulfate/nitrate ions. Data shown in Table 4 were used. The dotted line shows a 1:1 relationship.

Table 4. Cl Loss of Different Size Sea-Salt Particles^a

Stage	Size, μm	Cl/Na	Cl Loss, %	Cl Loss, μmol/m^3	SO4^2− Excess,^b μeq/m^3	Excess of SO4^2− and NO3^−,^c μeq/m^3
1	>11	1.85	0	0.0000	0.0004	0.0004
2	7.0–11	1.71	6	0.0009	0.0011	0.0008
3	4.7–7.0	1.67	8	0.0018	0.0020	0.0029
4	3.3–4.7	1.36	25	0.0042	0.0024	0.0056
5	2.1–3.3	0.93	49	0.0035	0.0026	0.0058

• a The average concentrations of all samples analyzed in this study were used in calculations.
• b SO₄²⁻ excess (μeq/m³) which calculated with respect to ammonium sulfate and calcium sulfate.
• c Excess of SO₄²⁻ and NO₃⁻ (μeq/m³) = SO₄²⁻ excess + NO₃⁻ excess, where NO₃⁻ excess was calculated with respect to magnesium nitrate.

4.2.2. SO₄²⁻ and NH₄⁺ Ions

[12] The concentrations of SO₄²⁻ and NH₄⁺ showed similar seasonal variations as shown in Figure 4a, and at the same time very similar size distributions with a peak at 0.5–1.5 μm in Figure 4b, indicating that they are in the form of ammonium sulfate ((NH₄)₂SO₄) of a submicron size. NH₄⁺ is not detected in the size range bigger than 3.3 μm, although there are some excess SO₄²⁻ have been determined in this range. Excess SO₄²⁻ is defined as excess one with respect to (NH₄)₂SO₄ and that in the size range bigger than 3.3 μm seems to correlate with the chlorine loss of sea-salt particles (Figure 5). The concentration of ammonium sulfate in summer (about 10 μg/m³) is about 3 times as that in winter (about 3 μg/m³), which is due to intensive photochemical reaction to produce SO₄²⁻ [Colbeck, 1998].

4.2.3. NO₃⁻ Ion

[13] The concentration of NO₃⁻ in summer (about 3 μg/m³) is also about 3 times higher than that in winter (1 μg/m³), like ammonium sulfate (Figure 4a). However, unlike ammonium sulfate, NO₃⁻ was detected in any particle size. The size distribution of NO₃⁻ showed a primary peak at 2–6 μm and a secondary peak at 0.5–1.1 μm, and most of NO₃⁻ is considered to be adsorbed on coarser soil-derived particles or sea-salt particles and the rest is suspended in the form of ammonium nitrate.

4.2.4. U and Th

[14] Fossil fuel, especially coal and gasoline, contains Th and U at ppb level [Gentzis and Goodarzi, 1997; Kitto, 1993]. Whether the two elements emitted from the combustion of fossil fuel suspend in the atmosphere with a particulate form or are absorbed on the secondary anthropogenic particles is an interesting question [Utsunomiya et al., 2002]. In this study, we determined trace U and Th in aerosols of different size, and found that all U and Th were contained in the insoluble fraction (Figure 6a). The size distribution of Th and U (Figure 6b) showed a bimodal curve with peaks at 0.43 μm and greater than 3 μm, and aerosols collected on stages 5 to 8 (0.43–3.3 μm) contained almost no Th and U. The amount of Th in the particles bigger than 3 μm is several times as much as that in particles smaller than 0.43 μm, whereas, amount of U in the particles smaller than 0.43 μm is much greater than that in particles bigger than 3 μm. As mentioned above, most of particles bigger than 3 μm are soil-derived particles, and almost of particles with size smaller than 0.43 μm are secondary anthropogenic ones. Therefore the results of this study indicated: most of Th in the atmosphere is originated from soil-derived particles and only a part of it might be originated from the combustion of fossil fuel, whereas most of U in the atmosphere might be originated from the combustion of fossil fuel. The result that all Th and U take a water-insoluble state indicates that Th and U originated from the combustion of fossil fuel suspend in the atmosphere rather in the form of particles. Utsunomiya et al. [2002] have reported some nanometer-sized U particles in carbonaceous particles in the atmosphere. This might explain our observation.

Figure 6

Chemical components in water-insoluble fraction of aerosols. (a) Seasonal variations of bulk concentrations of each ion ([M]_{total insoluble} = [M]_insoluble). Units are μg/m³ except for the data of Th, U, and Pb, whose units are ng/m³. The data of Pb were demagnified for 10 times to use for this plot. (b) Size distribution of each component in the insoluble fraction. Data plotted are average concentrations of the results shown in Table 3. Units are μg/m³ except for the data of Th, U, and Pb, whose units are ng/m³. The data of Pb were demagnified for 10 times to use for this plot.

4.2.5. Pb

[15] The concentration levels of Pb in the soluble fractions are comparable to those in the insoluble fraction in hotter seasons (during May to October), whereas those are much lower in the soluble fraction than in the insoluble fraction in colder seasons (during Nov. to Mar.), although the total Pb showed no remarkable seasonal variation. The soluble fraction showed a big peak at submicron size (0.43–1.5μm) and almost no Pb²⁺ in the size range bigger than 3 μm (Figure 4b), whereas the insoluble fraction contained Pb at any particle size with a broad peak at 0.43–1.5μm (Figure 6b). In order to understand why aerosols contained less soluble Pb²⁺ in colder seasons and what forms the soluble Pb²⁺ has in aerosols, the relationship between Pb²⁺ and anions in soluble fraction was plotted in Figure 7 from the data in Table 2. As shown in Figure 7, Pb²⁺ showed a close relationship with Cl⁻ and SO₄²⁻ in the soluble fraction, which implies that aerosols containing the lower concentrations of Cl⁻ and SO₄²⁻ in colder seasons may be responsible for the lower level of soluble Pb²⁺ in the colder seasons, and that soluble Pb²⁺ would be in a form of lead chloride/sulfate.

Figure 7

The relationship between Pb²⁺ and anions in the soluble fractions. Data plotted are average concentrations of the results shown in Table 2.

[16] In Tokaimura, Pb in the atmosphere is considered to be from the combustion of fuel and the soil-derived particles, since no Pb pollution from lead industry has been reported. The combustion of fossil fuel emits organic lead (e.g., Pb(CH₃)₄) and inorganic lead (e.g., PbCl₂ and PbO, etc.) into the atmosphere [Baird, 1998]. As mentioned above, the soluble Pb was detected only in submicron anthropogenic aerosols and the insoluble Pb was detected in both submicron anthropogenic and supermicron aerosols, implying the soluble Pb may be from the combustion of fuel and the insoluble Pb may be from both of the combustion of fuel and soil-derived particles.

4.2.6. Al and Fe

[17] Al and Fe are two of the most characteristic elements of terrigenous origin (soil and rocks). Al and Fe were detected only in the insoluble fraction of coarser particles (>2.1 μm) (Figure 6b). As mentioned earlier, the seasonal variations of these two elements showed a peak in May 2003 when dust storms happened most frequently in China (Figure 6a).

4.2.7. Ca and Mg

[18] Ca and Mg were detected in the soluble and insoluble fractions and most of them could be detected in coarser particles with a size bigger than 2.1 μm (Figures 4b and 6b). Most of Ca and Mg come from two sources: sea-salt and soil-derived particles. Bulk seawater containing not only Na⁺, Cl⁻ but also some Mg²⁺ and Ca²⁺. Therefore Mg²⁺ and Ca²⁺ in soluble fraction are considered to be from sea-salt particles, and the similarity of size distributions of Ca²⁺, Mg²⁺ and Na⁺ (Figure 4b) seems to support our consideration. The size distributions of Ca and Mg in the insoluble fraction are similar to that of Al. It is reasonable since all of them are considered to be from soil-derived particles. According to our results, Ca and Mg from the soil-derived particles are much less than those from sea-salt particles in the case of Tokaimura.

5. Conclusion

[19] The aerosol in Tokaimura, an urban city on the Pacific side of Japan, was also influenced by dust storm incidents in China, and much higher concentrations of coarser soil-derived particles (>3 μm) was observed during the dust storm period.

[20] Sea-salt particles in Tokaimura were in a size range of 4–9 μm, and chlorine loss was observed. The chlorine loss became larger with the decreasing size of sea-salt particles, and the chlorine loss seems to be compensated mainly by the excess of sulfate and nitrate ions.

[21] Nitrate ion was observed with a bigger peak at 2–6 μm and a smaller peak at 0.5–1.1 μm, indicating that most nitrate ions are adsorbed on supermicron particles and that just a few percent of it is as ammonium nitrate.

[22] Uranium and Thorium had two peaks in the two sides of their size distribution curves, indicating that they have two sources. Most of Th in the atmosphere was from soil-derived particles and only a part of it from the combustion of fossil fuel, whereas most of U from the latter source. Th and U originated from the combustion of fossil fuel may suspend in the atmosphere in the form of submicron particles.

[23] Lead in the soluble fraction had a peak on 0.43–1.5 μm with almost no lead in the particles bigger than 2.1 μm and the concentrations seem to correlate close to the concentrations of chloride and sulfate ions, indicating that Pb in soluble fraction originates from the combustion of fossil fuel. Lead in the insoluble fraction was detected in any particle size.

Acknowledgments

[24] This work was supported by the Japan Society for the Promotion of Science.