ARTICLE IN PRESS

Atmospheric Environment 41 (2007) 6999–7010 www.elsevier.com/locate/atmosenv

Major to ultra trace elements in rainfall collected in suburban Tokyo Tadashi Shimamuraa,b,, Masato Iwashitab, Satoe Iijimaa, Megumi Shintanib, Yuichi Takakuc a

Graduate School of Medical Sciences, Kitasato University, 1-15-1 Kitasato, Sagamihara, Kanagawa 228-8555, Japan b School of Allied Health Sciences, Kitasato University, 1-15-1 Kitasato, Sagamihara, Kanagawa 228-8555, Japan c Institute for Environmental Sciences, Rokkasho, Aomori 039-3212, Japan Received 18 December 2006; received in revised form 6 April 2007; accepted 8 May 2007

Abstract Major to ultra trace elements such as rare earth elements (REEs), platinum group elements (PGEs) in 20 rainfall events from suburban Tokyo were determined by inductively coupled plasma mass spectrometry (ICP-MS). Anion species were also determined by an ion chromatography (IC). The concentrations of PGEs were so low that only Pt was detected in some rainfall events. Enrichment factors (EFs), refer to soil and sea salt components, were calculated for the measured elements (with Al and Na as references). Be, (Na), Mg, (Al), Si, Cl, K, Fe, Rb, Sr, REEs (except La, Gd), Ta, and U were mostly originated from natural materials (soil and sea salt). For Li, B, Ca, Mn, Sr, Ba, and Cs, the contribution of natural materials was significant. EFs for Cu, Zn, As, Se, Sb, Cd, Pb, Bi, Ag, Te, Au, Pt, SO4-S and NO3-N exceeded 100 indicating non-crustal, non-sea salt origin, presumably anthropogenic; however, contribution of volcanic gases could not be excluded for As, Se, Te and Bi. Pt seemed to be uniformly distributed worldwide and a catalyst for automobile emission control may be the main source. Au also showed uniform distribution. On the other hand, EFs for Zr, Nb, Hf and Th were less than unity. Probably these elements resided in acid resistant refractory fine minerals that did not decompose with acid treatment, and did not evaporate and ionize in the ICP. An alternative explanation is that the concentration of these elements was lower in the soil of the sampling area than the average crust. In the crust normalized REE pattern plot, La, Eu and Gd showed clear positive anomalies. La and Gd could have anthropogenic components. A possible source of La and Gd is cracking catalyst for petrol refining, but this source does not fully explain the anomaly. The source of Gd may also be Gd-DTPA (Gadolinium (III) diethyltriaminepentaacetic acid) used for Magnetic Resonance Imaging (MRI) contrast agents. The Eu origin may be soil with higher concentration than the crust average. r 2007 Elsevier Ltd. All rights reserved. Keywords: Rainwater; Trace element; Rare earth element; Platinum group element; Enrichment factor

1. Introduction Corresponding author. Graduate School of Medical Sciences,

Kitasato University, 1-15-1 Kitasato, Sagamihara, Kanagawa 228-8555, Japan. E-mail addresses: [email protected] (T. Shimamura), [email protected] (Y. Takaku). 1352-2310/$ - see front matter r 2007 Elsevier Ltd. All rights reserved. doi:10.1016/j.atmosenv.2007.05.010

Atmospheric deposition of trace elements is an important contamination source for surface soil (Pirrone and Keeler, 1996; Rocha et al., 2003; Wong et al., 2003). Wet deposition is a major pathway of trace element deposition to soil via scavenging

ARTICLE IN PRESS 7000

T. Shimamura et al. / Atmospheric Environment 41 (2007) 6999–7010

processes, either in or below cloud. In urban areas of large cities, concentrations of trace elements (e.g. Ni, Cu, Zn, Cd, and Pb) in the surface soil are expected to increase rapidly (Wong et al., 2003). These elements are potentially toxic to humans and other organisms, and thus atmospheric deposition is of major significance in many ecosystems. There have been many studies of trace element concentrations in wet deposition (Poissant and Be´ron, 1994; Halstead et al., 2000; Kim et al., 2000; Luo, 2001). Reimann et al. (1997) reported 39 elements in the wet deposition of Arctic catchments in Northern Europe. They found 19 elements at high concentrations, close to a metal smelter, in comparison to other sampling points (630 times for Ni, 4 times for Se). Poissant et al. (1994) concluded that only 7 elements (Al, Ti, Fe, Rb, Y, Zr, and Cs) out of 24 could be classified ‘‘natural’’ in rainfall of Montreal Island, Canada. Extremely low concentrations of Rh, Pd, and Pt (0.0007, 0.01 and 0.01 pg g1, respectively) were found in 7000-year-old ice core samples from Greenland (Barbante et al., 1999). On the other hand, the concentrations of these elements in recent snow samples from Greenland were increasing rapidly and were two orders of magnitude higher than those in old ice core samples (Barbante et al., 2001b). Mining and smelting of these elements were likely sources from 1970 to 1990, and automobile catalytic converters after 1990. Veysseyre et al. (2001) reported concentrations of 21 elements in fresh snow from Chamonix and Maurienne valley, in the French Alps. They found that Pd, Pt, and Au were highly enriched (1.6–4.2, 0.47–1.3, 0.7–0.97 pg g1, respectively) and had very small variations with sampling altitude, suggesting that the source of these elements may be remote and anthropogenic. Airborne particulate matter (APM) collected in the vicinity of Tokyo (Tsukuba) was found to have Pt concentrations of 73–184 ng g1 (Mukai et al., 1990). These solid particles may have contained significant amounts of carbonaceous material. Urban air samples from Boston and Tokyo had respective platinum group element (PGE) concentrations (pg m3) of Pt (6.7 and 2.1), Pd (7.5 and 0.79), and Rh (1.1 and 0.47) (Rauch et al., 2005); they estimated catalyst-derived Pt depositions in roadside and regional environments to be 0.4–6.2 and 0.5–0.7 mg m2 year1, respectively. Such data for PGEs and Au in precipitation and APM are quite rare. Sholkovitz et al. (1993) reported rare earth element (REE) concentrations in precipitation (dis-

solved and particle phases) and aerosols and found large-scale fractionation among aerosols, earth upper crust, seawater, and dissolved and particle phases. They indicated complex chemical reactions between raindrops and mineral properties in the aerosols. Rainwater from Japan and the East China Sea has shown large variation in REE composition depending on location (60 to 41600 pmol kg1 total REEs; Zhang and Liu, 2004). There was an anthropogenic contribution, especially for light REEs. The enrichment of La, Ce, Nd, Pr, Gd and Sm in particulated matter witho2.5 mm diameter (PM2.5) collected at a petrochemical industrial complex in Houston, Texas was found to be primarily caused by emissions from fluidized-bed catalytic cracking operations (Kulkarni et al., 2006). Since inductively coupled plasma mass spectrometry (ICP-MS) was introduced to environmental sample analyses, the number of elements reported has greatly increased (Yamasaki and Tsumura, 1992; Dupre´ et al., 1996; Kreamer et al., 1996; Zhang and Nozaki, 1996). However, data for REEs (including Y), PGEs, Au, Nb, Te, and Ta in precipitation are still very scarce, since concentrations are usually extremely low (Sholkovitz et al., 1993; Barbante et al., 2001a; Turetta et al., 2003; Varga et al., 2003) and in some cases there are measurement interference problems. However, there is increasing use of many of these elements in industry. REEs are used for high alloy metals, catalysts, glass additives, medical reagents, semiconductor products, and cracking catalysts for petroleum refining (Molycorp Inc., 1997). PGEs are used for jewelry, medical reagents, and catalysts for automobile exhaust and chemical plants (Matthey, 2006). Thus there may be large increases in anthropogenic emission of these elements to the environment, and many are potentially toxic to humans and other organisms (Kielhorn et al., 2002) even at very low concentrations. We analyzed rainfall samples for 67 species (including REEs, some PGEs and anions) from Sagamihara City, a suburb of Tokyo. We investigated whether the detected species were natural or anthropogenic, and identified possible sources. 2. Material and methods 2.1. Sampling The sampling site (Fig. 1) was in Sagamihara City, 35 km W of central Tokyo. The population of

ARTICLE IN PRESS T. Shimamura et al. / Atmospheric Environment 41 (2007) 6999–7010

7001

Fig. 1. Location map of the sampling point, Sagamihara, Japan.

400 300 mm

the city is 4600000 and land use is mainly residential, however, 5–6 km NW and SW of the sampling site there are industrial complexes producing automobile parts, glass materials, metal plating, food and many other items. There are several volcanoes in the region; Mt. Fuji is 70 km W, Hakone Volcano 50 km W, Izu Oshima (Mt. Mihara) 90 km S, and Miyakejima (Mt. Oyama) 150 km S of the sampling site. Among these volcanoes, only Miyakejima is presently active (quiescent degassing only, very small degassing is also observed at Mt. Hakone). Volcanic ash repeatedly erupted from Mt. Fuji until 300 years ago and covered up to 10 000 km2 of the surrounding land, ash accumulated to 10 m thick around the sampling point (forming Kanto loam). Thus the surface soil of the sampling area mostly originated from volcanic ash. Wind direction at the site is generally W to NW in winter, and S to SW in summer. Rainy seasons are usually May–July and late September–October. In the rainy seasons, stationary fronts frequently occur along the Japanese island arc, then wind direction at the land surface is N to NE. Typhoons occur in late August–September. The samples were mostly col-

200 100 0 Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec

Fig. 2. Monthly precipitation at the sampling point in 2005.

lected when there was either a stationary front or the influence of a typhoon. The dry season is December–February. Fig. 2 shows precipitation in 2005 at a meteorological observation station near the sampling point (Ebina station, Japan Meteorological Agency, 2007). In order to determine major to ultra trace elements in rainwater, it was necessary to collect a large and clean sample. The rain collector consisted of four polypropylene funnels with an opening area of 710 cm2 each and four 500 ml polypropylene bottles, which were packed in a plastic container, with each bottle connected to one funnel. Approximately 2 L of rainwater sample can be collected with

ARTICLE IN PRESS 7002

T. Shimamura et al. / Atmospheric Environment 41 (2007) 6999–7010

7 mm precipitation. The funnels had an acrylic resin lid held by thin paper. Several raindrops were sufficient to wet the thin paper, which then broke, and the lid was removed by rubber springs. Since  anions were measured (NO 3 , Cl ), acid cleaning of the sampler was avoided. Instead, the funnels, lids and bottles were soaked in ultra high purity organic alkaline detergent (Tama Pure TMSC, Tama Chemical, Japan) for 3 days and rinsed with high purity water (Milli-Q, Millipore Japan, Japan) several times. The rain collector was set on the roof of a university building (20 m above the ground), located in the center of Sagamihara City. Radar observations of rain clouds were monitored on a web site and about 30 min before rain started, the rain collector was carried out from a Class 1000 clean room. The rain collector was recovered immediately after the rain stopped, thus dry deposition was virtually negligible. The rain samples were collected from September 2004 to October 2005. 2.2. Analysis After pH and electric conductivity measurements, an aliquot of collected sample was used for anion  2 3 analyses (F, Cl, Br, NO 2 , NO3 , SO4 , PO4 ) by ion chromatography (IC7000, Yokogawa Analytical Systems, Japan). Separation was performed by an anion column (Excelpak ICS-23) using 3 mmol L1 Na2CO3 as an eluate. A certified reference material of lake water (ION 915, National Water Research Institute, Canada) was intermittently analyzed to confirm accuracy of analyses;  2 Cl, total-N (NO 2 +NO3 ), and SO4 agreed well with the certified values. Analytical results for ION 915 are presented in Table 1. Unfortunately the concentrations of F, Br, and PO3 4 were generally below the detection limits (0.05, 0.01 and 0.1 mg L1, respectively), only a few data were obtained and were omitted from this report. The remaining sample was acidified by ultra high purity nitric acid (Tama Pure AA100, Tama Chemicals, Japan) to give pH 1. Sixty elements (major metallic cations such as Na to ultra trace elements) were analyzed by ICP-MS (PQ-ExCell-S, Thermo Scientific, USA; HP4500, Agilent, USA). Four different analytical modes were applied for the ICP-MS analyses and running conditions of each mode and applied elements are summarized in Supplementary Table S1. Si was determined by inductively coupled plasma atomic emission spectroscopy

(ICP-AES; Liberty, Varian, Austria) with running conditions of emission line: 251.611 nm, RF power: 1200 W, plasma gas flow: 15 L min1, carrier gas flow: 0.75 L min1, integration time: 5 s. Procedure blanks were collected intermittently between rainfall events using high purity water. Blank levels were generally less than 10% of the concentration in rain samples. However, blank levels of Li, Al, Zn, Nb, Hf, and W sometimes exceeded 10%. The blank level of Zn reached more than 60% when Zn concentration was of the order of 1 mg L1. We found that the high blank of Zn originated from clean bench (perhaps Hepa filter). Blank correction for Zn exceeded 10% if Zn concentration in the rain was less than 10 mg L1. Blank levels of REEs were generally less than 5% of actual concentrations, however, in some cases they were up to 43% (e.g. Lu on September 29th 2004). Blank levels, including instrumental background for Pt, were approximately 0.03 ng L1, 1% to comparable level of actual concentration in the rain samples. Since concentrations of PGEs and Re were extremely low, instrumental background and interference problems were significant, thus we obtained meaningful data only for Pt. The concentrations of Hf and Ta were also very low, and data was obtained only for several events. Most spectral interference by molecular ions could be avoided using a collision cell, however, it was difficult to remove oxide and hydride ions. Potential interfering species (mainly oxide and hydride species) with the isotopes used for the element determinations are summarized in Table 2. To check BaH+ interference with La+, we plotted La/Ce versus Ba/Ce. If BaH+ interference was significant, La/Ce value should systematically increase with increased Ba/Ce value. Similarly, we checked BaO+ interference with Eu+, using a Eu/Sm versus Ba/Sm plot. No systematic variation was observed for both La/ Ce–Ba/Ce and Eu/Sm–Ba/Sm plots. We determined Eu concentration with 151Eu+ and 153Eu+, excellent agreement was obtained using both isotopes. We determined SrH+, BaH+ and BaO+ production rates, using Sr and Ba standard solutions with the same running conditions as for REE determination. CeO+ and CeOH+ production rates were determined by a Ce standard solution. PrO+ production rate was assumed to be the same as CeO+ production rate. CaO+ production rate was assumed to be the same as SrO+ production rate. If interference contributions were more than 10%, interference corrections were applied. Since

ARTICLE IN PRESS T. Shimamura et al. / Atmospheric Environment 41 (2007) 6999–7010

7003

Table 1 Analytical results of some certified reference materials Species

JSAC 0301 This worka

Li Be B Na (mg L1) Mg (mg L1) Al Si (mg L1) K (mg L1) Ca (mg L1) V Cr Mn Fe Co Ni Cu Zn As Se Rb Sr Y Zr Nb Mo Ag Cd Sn Sb Te Cs Ba La Ce Pr Nd Sm Eu Gd Tb Dy Ho Er Tm Yb Lu Hf W Ta Pt Au Tl Pb Bi Th

4375 BQ 89007400 4.570.2 2.8670.14 198007900 0.9670.01 0.5770.02 11.970.6 74507300 160710 122722 47007100 12.970.6 130720 520720 220720 26076 12377 490720 160007100 271 471 BQ 390710 0.770.9 2.170.1 5.770.1 32.370.4 BQ 3.870.1 57075 0.9370.02 1.4370.06 0.3370.01 1.8370.07 0.5770.02 0.2070.005 0.6670.02 0.1570.01 0.8370.03 0.2370.01 0.8870.04 0.1570.01 1.2970.03 0.3170.01 BQ 11.170.2 BQ BQ BQ 371 4.470.1 ND 0.0670.01

SLRS-4 Certifiedb

86007300 4.470.1 2.8570.04 190007900 0.5770.02 12.070.2 150710 12577 47007300

570770 190730 240730 (130)c

380710 2.370.7

600720

(5)c

This worka 490760 ND ND 2.470.2 1.670.2 5700075000 ND 0.5970.11 6.070.3 370740 ND 34707140 108000712000 3878 7307100 18107120 11707430 760730 ND 1520750 2790073900 145710 110710 ND 210710 372 1372 ND 260710 ND 10710 1280071600 290720 360710 6571 290710 5771 971 4071 4.770.1 2371 4.270.1 1171 1.670.1 1171 ND BQ 675 BQ BQ BQ 975 89737 2.370.1 ND

ION-915 Certifiedb

772 2.470.2 1.670.1 5400074000 0.6870.02 6.270.2 320730 330720 33707180 10300075000 3376 670780 1810780 9307100 680760

2630073200

210720 1272 230740

122007600

8677

This worka

Certifiedb

ARTICLE IN PRESS T. Shimamura et al. / Atmospheric Environment 41 (2007) 6999–7010

7004 Table 1 (continued ) Species

U Cl (mg L1) 1 SO2 4 -S (mg L ) T_N (mg L1)

JSAC 0301

SLRS-4

ION-915

This worka

Certifiedb

This worka

Certifiedb

2.870.1 ND ND ND

2.970.2

55710 ND

5073 ND ND

This worka

Certifiedb

1.2170.04 3.370.2 0.32070.023

1.3970.28 3.470.64 0.34370.061

Concentration in ng L1 except indicated. ND: not determined; BQ: below quantification. a Errors indicate standard deviations. b Errors indicate 95% confidence limits. c Infromation value.

Table 2 Potential interfering species to the isotopes of analytes Isotope

Interference

59

Ca16O+ SrH+ 85 Rb16O+ 87 16 + 206 Sr O , Pb2+, 63Cu40Ar+ 89 16 + 65 Y O , Cu40Ar+ 91 Zr16O+ 93 Nb16O+ 95 Mo16O+ 138 BaH+ 135 Ba16O+ 137 Ba16O+ 141 16 + Pr O 143 Nd16O+ 149 Sm16O+ 165 Ho16O+ 179 Hf16O+ 181 Ta16O+

Co+ Y+ 101 Ru+ 103 Rh+ 105 Pd+ 107 Ag+ 109 Ag+ 111 Cd+ 139 La+ 151 Eu+ 153 Eu+ 157 Gd+ 159 Tb+ 165 Ho+ 181 Ta+ 195 + Pt 197 Au+ 89

43

88

concentration levels of Mo and Cd, Hf and Pt, (Ta+Pt) and Au were generally of the same order respectively, the oxide and/or hydride interferences of Mo, Hf and (Ta+Pt) on Cd, Pt and Au were estimated to be less than 1%, respectively.  After all, Ru, Rh, Pd, Re, Ir, F, PO3 4 and Br could not be determined since their concentrations were below quantification limits and/or spectral interference problems. To check analytical performance, certified reference materials of river water SLRS-4 (National Research Council, Canada) and JSAC-0301 (The Japan Society for Analytical Chemistry, Japan) were analyzed. The results are summarized in Table 1.

3. Results and discussion Fig. 3 shows box–whisker plots of analytical results for 20 rain samples. Each plot shows minimum, 25 percentile, median, 75 percentile and maximum. The results are also summarized in Supplementary Table S2 together with blanks, limits of quantification, and wet deposition rate for each species. The concentrations of the elements determined ranged more than 8 orders of magnitude from 4.0 mg L1 (Cl) to 0.01 ng L1 (Tm). Noticeable points are: (1) Zn, Ba and Pb are semi-major components (1–10 mg L1) after major elements like Na, Mg, Al, Si, Ca, Fe, and Cl; and (2) Be, heavy REEs, Hf, Ta and Pt have extremely low concentrations (0.1–1 ng L1). 3.1. Enrichment factors To evaluate the contribution of the non-soil (ns), non-sea salt (nss) components, that are presumably anthropogenic components, we define the enrichment factor of element X (EFx) as follows: EFx ¼ C x =½C x ðsÞ þ C x ðssÞ; where Cx is the concentration of element X in the rain, Cx(s) is the concentration of element X in the soil component and Cx(ss) is that of the sea salt component. Cx(s) and Cx(ss) are given by C x ðsÞ ¼ C Al  XðsÞ=AlðsÞ; C x ðssÞ ¼ C Na  XðssÞ=NaðssÞ; where X(s)/Al(s) is the elemental ratio (by mass) of X to Al in soil. We calculated X(s)/Al(s) values from the upper crust average (Reimann and Caritat, 1998). X(ss)/Na(ss) is elemental ratio of X to Na in

ARTICLE IN PRESS T. Shimamura et al. / Atmospheric Environment 41 (2007) 6999–7010

7005

1.E+07 Mg

1.E+06

Si

Ca Fe

1.E+05 1.E+04

1.E+02

As

Al

B

Pb

Sr

K

1.E+03

Mo

Cd

Ce

Cu

V

Gd

Nd

Bi

W Tl

Cs Dy

Eu

Cr

Rb

Li

1.E+00

Sn

Er

U

Yb Hf

Tb

Ag

Se

Ta

Te

Nb

1.E-01

La

Sb

Y Zr

Co

1.E+01

-

Cl

Ba Ni

Na

ng L-1

Zn

Mn

Pr

Be

Sm

Au

1.E-02

Ho

Tm

Th

Pt

Lu

1.E-03

Fig. 3. Box–whisker plot of the elemental concentrations in the rainwaters of 20 rainfall events. Each plot shows minimum, 25 percentile, median, 75 percentile and maximum from the bottom.

1.E+05 Sb

1.E+04

Bi

As

Cu

Pb

1.E+03 B

V

EF

Li

Co

Mn

Se

K

1.E+01 Na

1.E+00

Y Mo Nb

Ni Ca

Cr

Fe

Be

Te

Au

Gd

Ba

Pt

Cd

Ta Pr

Sn

Sm

Dy

Er

U

Yb

Th Hf

Eu

Sr

Si

1.E-01

Ag

Zn

Al Mg

W

La

1.E+02

Cs

Ce

Nd

Tb

Ho

Tm

Lu

Tl Cl

Rb Zr

1.E-02 1.E-03

Fig. 4. Box–whisker plot of enrichment factors in the rainwater of 20 rainfall events.

sea salt calculated from North Pacific Ocean average (Reimann and Caritat, 1998). Fig. 4 shows the box–whisker plot of the EFs for 20 rainfall events. The medians of EFs of Be, (Na), Mg, (Al), Si, Cl, K, Fe, Rb, Sr, REEs, (except La, Eu, Gd) Ta and U were close to unity indicating natural origins of these elements. It should be noted that although concentration ranges of these elements are greater than an order of magnitude (Fig. 3), the ranges of EF values are within a factor of 10 (with the exceptions of Rb, Y, Ta and U; for which, the sources were not obvious). The EFs of Li, Ca, Mn, Co, Y, Cs, Ba, La, Eu and Gd were also close to unity (EFo10) but systematically higher than those for previously mentioned elements. The contribution of anthropogenic sources was probably significant. For instance, recent consumption of Li has increased greatly for use in rechargeable cells. The used cells are recycled, but only Co and Ni are recovered,

recovery systems for Li are not yet established. The recovery system for Co includes burning of the cells, during which Li may be emitted into the atmosphere. EFs of La and Gd were also high and ranges wider than other REEs, suggesting a contribution from anthropogenic sources. On the other hand, Eu may have a natural origin (see later). The EFs of Zn, Se, Ag, Cd, Sb, Te, Pt, Au and NO3-N exceeded 500, and those of Cu, As, Mo, and Bi exceeded 100. Natural sources of these elements (besides soil and sea salt) are volcanoes (Hinkley et al., 1999), wild forest fires (Chankina et al., 2001; Dennis et al., 2002) and/or biogenic sources (Beauford et al., 1975; Hong et al., 1996). There is certainly a contribution of volcanic gas from Miyakejima, in the Pacific Ocean 150 km S of the sampling point and active since July 2000. In summer time, south winds would be expected to transport volcanic gas. There are no reports of trace element concentrations in volcanic gases from

ARTICLE IN PRESS 7006

T. Shimamura et al. / Atmospheric Environment 41 (2007) 6999–7010

Miyakejima, however, a volcanic emission can have significant quantities of SO2-S, As, Se, Sn, Te and Bi (Hinkley et al., 1999). In summer time SO2 level in the air of the sampling area was 1–20 nmol mol1 (Tokyo Metropolitan Government, 2006). If 10% of the SO2 was assumed to come from Miyakejima, and also that the concentration ratios of these elements to SO2 were the same as Kilauea quiescent emission (Hinkley et al., 1999), then contributions of volcanic gas from Miyakejima were estimated to be 3% for As, 10% for Se and Sn, and 100% for Te and Bi. In addition to Miyakejima, Hakone volcano is 50 km W of the sampling point, it is not as active as Miyakejima but still emits quiescent volcanic gases, which could also contribute. No wild forest fire was reported during the sampling period. Although we cannot rule out the possible contribution of natural biogenic sources, considering the huge amount of human activity in the Tokyo metropolitan area, and this is likely to be the dominant contribution for these elements. The ranges of EFs of these elements are generally quite large, since small changes in the natural contribution greatly affect EF values. In contrast with those elements mentioned above, EFs of Nb, Zr, Hf and Th were systematically lower than unity. These elements may be contained in refractory minerals such as zircon, and may not be decomposed during acidification. In the ICP, these minerals may also resist evaporation and ionization. Consequently, the concentrations of these elements may have appeared lower than their true values. However, this explanation is not fully consistent with the REE and U observations (see later). 3.2. Comparison to the existing data Poissant et al. (1994) determined 24 elements in rainfall collected in an urban area of Montreal Island, Canada. The concentrations of trace elements were generally lower in Sagamihara than average values in Montreal. The concentrations of B, Al, V, Fe, Zn, Y, Cd, Sn, Sb, and Ce were of similar levels (within a factor of 3, although Al was higher in Sagamihara), while those of Li, Cr, Mn, Ni, Cu, Sr, Zr, Mo, Pb, and Th were much lower in Sagamihara (concentration ratios of Sagamihara/ Montreal ¼ 0.09, 0.07, 0.16, 0.14, 0.2, 0.07, 0.23, 0.11, 0.24, 0.003, respectively). It is rather surprising that trace element concentrations were similar or lower in the Tokyo suburban area than Montreal, since total population and density are far higher in

Tokyo. However, air quality of Tokyo has improved much in the last decade (Minoura et al., 2006) and the sampling location was away from the center of Tokyo. Halstead et al. (2000) determined Cd, Cu, Fe, Mn, Pb, Zn, Ca, Mg, and Na in rainwater collected at Paradise, a remote site in Fiordland, New Zealand. Ca, Mg, and Na concentrations in Sagamihara were similar to Paradise on average; however, Zn, Cd, Cu, Pb, and Mn concentrations in Sagamihara were far higher (concentration ratios of Sagamihara/Paradise ¼ 530, 380, 202, 63, 22, respectively) as expected. These ratios are close to EFs for the respective elements, except Cd; indicating that Zn, Cu, Pb, and Mn concentrations in Paradise rainfall were close to natural levels. The exception is Cd, for which the EF is more than an order of magnitude higher (4200) than the concentration ratio. Therefore, Cd in Paradise rainfall consisted of a significant non-soil, non-sea salt component that is presumably anthropogenic. Veysseyre et al. (2001) determined 21 trace elements in fresh snow collected in the French Alps at different altitudes. The concentrations of most elements were two to five orders of magnitude lower than those in Sagamihara rainfall, as expected. Remarkably, Pt and Au concentrations were similar at both sites. A comparison of Rh, Pd, Pt, and Au concentrations in recently precipitated Alpine snow, Greenland snow and the present study is given in Table 3. These elements also have very high EFs at both sites (Pd and Pt 3500–5700, and Au 1600). Veysseyre et al. (2001) found these elements were uniformly distributed regardless of the sampling point altitude. Van de Velde et al. (2000) constructed depth profiles of Ag, Au, Pd, Pt, and Rh concentrations in ice and snow from Mont Blanc; apart from Ag, the

Table 3 PGEs and Au in the precipitation in the world pg g1

A

B

Rh Pd Pt Au

0.08 2.6 0.18 0.13

2.6 0.5 0.76

C 0.05 0.76 0.33

This work

0.19 1.1

A: Alps snow, recent precipitation 1980–1991 average. Van de Velde et al. Atmos. Environ. 34 (2000) 3117. B: Alps snow, recent precipitation 1998, altitude 1150–3532 m average. Veysseyre et al. Atmos. Environ. 35 (2001) 415. C: Greenland snow, recent precipitation 1991–1995, average. Barbante et al. Anal. Chem. 71 (1999) 4125.

ARTICLE IN PRESS T. Shimamura et al. / Atmospheric Environment 41 (2007) 6999–7010

concentrations of these elements in the recent snow were also similar to the present study. Annual wet deposition rate of Pt in the present study (280 ng m2 year1 in 2005, see Supplementary Table S2) was also quite similar to the estimates by Rauch et al. (2005) of 500–700 ng m2 year1 total deposition in a ‘‘regional environment’’. Contamination of these elements in the atmosphere may be worldwide and uniform concentrations levels (Rauch et al., 2005). However, the snow samples from the French Alps were collected close to main highways between France and Italy (Veysseyre et al., 2001); and Greenland snow could be affected by Russian Pt smelters (Barbante et al., 2001a, b). Thus both French Alpine and Greenland snow samples could contain higher Pt than other remote areas, which could explain the similar Pt concentration between the suburban (Sagamihara) rain and the snow in remote areas (Alps and Greenland). In any case, the most important source of Pd and Pt to the atmosphere may be the catalyst used in automobile emission control. Au is used in various industrial fields (e.g. electronics, aircraft, and electroplating), a possible source of Au to the atmosphere may be a city incinerator for waste of electronic devices, although much of the Au content is recycled. 3.3. Correlation of element concentrations in rainfall Since element concentrations in rainfall usually show logarithmic Gaussian distributions, correlation coefficients (r) were calculated for logarithms of the concentrations of each element. Al, Fe and REE (except La, Gd) were strongly correlated with each other (r40.9; Supplementary Table-S3), presumably indicating crust origin of these elements. Eu, which has a slightly higher EF, correlated very well with other REEs and the range of EF was small, indicating natural origin for Eu. The correlations between La and other REEs, and Gd and other REEs were exceptionally weak (ro0.7 and ro0.8, respectively, compared to r40.9 among the other REEs), suggesting contributions of anthropogenic sources. It should be noted that the ranges of EFs for La and Gd are quite large compared to other REEs (Fig. 4), indicating contributions of anthropogenic sources. Cr, Mn, Co, Y, Zr, Nb, Ba, Th and U were also well correlated with each other and with Al, Fe and REE (r ¼ 0.7–0.9). A crust component contribution may be significant for these elements, but other components may also be important, since

7007

EF ranges are large. On the other hand, Na correlated well only with B, Mg, Cl, and Sr (r ¼ 0.65, 0.93, 0.85, 0.68, respectively), indicating marine origin of these elements. Even K, Ca, Rb, Ba did not correlate well with Na. The contribution of sea salt may be minor for most elements in rain. 3.4. Rare earth element pattern As pointed out in Sections 3.2 and 3.3, REEs originated from soil (crust). However, if we examine our data in detail, the REE patterns of 20 rainfall events normalized to the upper crust concentrations (Taylor and McLennan, 1985) were divided into three groups (Fig. 5). The first group showed large positive La and Gd anomalies. Eu also showed a small positive anomaly (Fig. 5a). The second group showed clear but not large positive La and Eu anomalies (Fig. 5b). The third group showed rather irregular variations; and positive La, Eu and Gd anomalies were evident (Fig. 5c). Sholkovitz et al. (1993) determined REEs in precipitation in Bermuda and observed large-scale fractionation between atmospheric samples and the upper crust. They found that the shale-normalized pattern of ‘‘acid soluble/filtered’’ fraction in the wet deposition showed ‘‘convex-up’’ distribution (positive slope at light REEs, high in Sm and Gd and negative slope at heavy REEs) and a markedly negative Eu anomaly. In the present study, all three groups had convex-up shape REE patterns, they had a positive slope from Ce to Sm and a negative slope from Tb to Yb, however, in contrast with Sholkovitz et al. (1993) we observed positive La and Eu anomalies and, in many cases, a large positive Gd anomaly. A positive slope from Ce to Sm, and small positive Eu anomalies are similar to standard volcanic rocks JB-3 (Mt. Fuji basalt, Imai et al., 1995) and JA-1 (Hakone volcano andesite, Imai et al., 1995). These volcanoes are 50–70 km W of the sampling point and the soil around the sampling point is strongly influenced by volcanic ash from Mt. Fuji. However, these rocks do not show a negative slope from Tb to Yb, in fact it is almost flat (Imai et al., 1995). Sholkovitz et al. (1993) suggested selective dissolution of soil minerals during ‘‘atmospheric chemical weathering’’ and chemical treatments of the samples. They claimed that negative slope of heavy REEs might be caused by refractory minerals such as zircon, rich in heavy REEs, which do not dissolve during chemical treatment of samples. This explanation is qualitatively consistent

ARTICLE IN PRESS T. Shimamura et al. / Atmospheric Environment 41 (2007) 6999–7010

7008

1.E-04 Sept 24 2004 July 10 2005 July 30 2005 Oct. 04 2005

Sept 29 2004 July 22 2005 Aug. 23 2005

Sample/Crust

1.E-05

1.E-06

1.E-07

1.E-08 La

Ce

Pr

Nd Sm

Eu Gd

Tb Dy Ho

Er Tm Yb Lu

1.E-04 July 01 2005 July 06 2005 July 07 2005 July 09 2005 Oct. 07 2005

Sample/Crust

1.E-05

1.E-06

1.E-07

1.E-08 La

Ce

Pr

Nd Sm

Eu Gd

Tb Dy Ho

Er Tm Yb Lu

1.E-04 June 02 2005 July 25 2005 Aug 12 2005 Sept 06 2005

June 29 2005 July 26 2005 Aug 25 2005 Sept 24 2005

Sample/Crust

1.E-05

1.E-06

1.E-07

1.E-08 La

Ce

Pr

Nd Sm

Eu Gd

Tb Dy Ho

Er Tm Yb Lu

with the present study; Zr, Nb, Hf and Th, which concentrate in zircon, were depleted (EF ¼ 0.12– 0.35). These elements may not dissolve/evaporate/ ionize during sample preparation and introduction into the ICP. However, if a main source of heavy REEs is zircon, it is difficult to explain why their EFs are close to unity. Similarly, zircon may also be a main source of U, yet EF of U is 1.2. Using the crust average of the elements, the calculated contribution of zircon to Al was very small since zircon is not an alumino-silicate, and perhaps its contribution to heavy REEs and U was also minor. On the other hand, the contribution of zircon to Zr, Nb, Hf, and Th was large (Taylor and McLennan, 1985). Thus if zircon did not dissolve/evaporate/ ionize during analysis, only Zr, Nb, Hf and Th concentrations would be affected and their calculated EFs would be less than unity. The REE pattern, elemental correlations, EFs and their variation ranges indicate that La and Gd have a significant contribution from anthropogenic sources. A possible candidate for a La and Gd source is the cracking catalyst of petrol (Kulkarni et al., 2007), but this does not fully explain the anomaly. If cracking catalyst is the source, a positive Ce anomaly would also be expected, however, there were no Ce anomalies. Another source of Gd may be Gd-DTPA used for MRI contrast agents (Magnetic Resonance-Technology Information Portal, 2007; Bau and Dulski, 1996; Nozaki et al., 2000). Many rivers in Europe and Japan show a positive Gd anomaly (Bau and Dulski, 1996; Nozaki et al., 2000) and we have observed high Gd concentration in some domestic wastewater (unpublished data). We do not know how Gd-DTPA is transported into the atmosphere; however, a preliminary investigation found that dry deposition samples collected near the rain sampler position (the roof of a university building) showed a positive Gd anomaly. The soil near the sampling area may be contaminated by Gd-DTPA and be rolled up into the atmosphere. On the other hand Eu, which also shows a positive anomaly, correlates very well with other REEs and this indicates a natural origin. Possibly the soil near the sampling

Fig. 5. Rare earth element pattern normalized to upper crust concentration (Taylor and McLennan, 1985). (a) Large La and Gd anomalies. (b) Small La and Eu anomalies, positive slope between Ce and Sm and negative slope between Tb and Yb. (c) Rather irregular pattern with small anomalies on La, Eu and Gd.

ARTICLE IN PRESS T. Shimamura et al. / Atmospheric Environment 41 (2007) 6999–7010

area may also have a positive Eu anomaly. Small positive Lu anomalies were observed in many rainfall events, but it is not known if there are significant anthropogenic sources for Lu. 4. Conclusion EFs indicate that the sources of Be, Na, Mg, Al, Si, Cl, K, Fe, Rb, Sr, REEs (except La and Gd), Ta, and U in the rain may be natural. Zn, Se, Ag, Cd, Sb, Te, Pt, Au, and NO3-N were highly enriched in rain and their EFs exceeded 500, indicating anthropogenic sources were dominant. However, the contribution of volcanic gases could not be excluded. Pt and Au were highly enriched in rainfall and concentration levels were similar to those in Alpine and Greenland snows; these elements are possibly distributed uniformly worldwide. The catalyst for emission control of automobiles may be the main source of Pt. Among the REEs, La and Gd may have an anthropogenic component. The La and Gd source could be a cracking catalyst for petrol refining and an additional Gd source may be Gd-DTPA used in MRI. Acknowledgement The authors would like to thank Mr. T. Hayashi, Tohoku Nuclear Co., Ltd., who helped with sample analysis. The comments of anonymous reviewers greatly improved the manuscript. Appendix A. Supplementary materials Supplementary data associated with this article can be found in the online version at doi:10.1016/ j.atmosenv.2007.05.010. References Barbante, C., Cozzi, G., Capodaglio, G., Van de Velde, K., Ferrari, C., Veysseyre, A., Boutron, C.F., Scarponi, G., Cescon, P., 1999. Determination of Rh, Pd, and Pt in polar and alpine snow and ice by double-focusing ICPMS with microconcentric nebulization. Analytical Chemistry 71, 4125–4133. Barbante, C., Cescon, P., Cozzi, G., Planchon, F., Boutron, C.F., Wolff, E.W., Gaspari, V., Ferrari, C., 2001a. Ultrasensitive determination of heavy metals at the sub picogram per gram level in ultraclean Antarctic snow samples by inductively coupled plasma sector field mass spectrometry. Analytica Chimica Acta 450, 193–205. Barbante, C., Veysseyre, A., Ferrari, C., Van de Velde, K., Morel, C., Capodaglio, G., Cescon, P., Scarponi, G.,

7009

Boutron, C., 2001b. Greenland snow evidence of large scale atmospheric contamination for platinum, palladium, and rhodium. Environmental Science and Technology 35, 835–839. Bau, M., Dulski, P., 1996. Anthropogenic origin of positive gadolinium anomalies in river waters. Earth and Planetary Science Letter 143, 245–255. Beauford, W., Barber, J., Barringer, A.R., 1975. Heavy metal release from plants into the atmosphere. Nature 256, 35–37. Chankina, O.V., Churkina, T.V., Ivanov, A.V., Ivanov, V.A., Ivanova, G.A., Koutsenogii, K.P., Kovalskaya, G.A., 2001. Multielement composition of the aerosols of the forest fires of boreal forests upon burning of forest combustibles. Nuclear Instrument and Methods in Physics Research Section A 470, 44–447. Dennis, A., Fraster, M., Anderson, S., Allen, D., 2002. Air pollutant emissions associated with forest, grassland, and agricultural burning in Texas. Atmospheric Environment 36, 3779–3792. Dupre´, B., Gaillardet, J., Rousseau, D., Alle`gre, C.J., 1996. Major and trace element of river-borne material: the Congo Basin. Geochimica Cosmochimica Acta 60, 1301–1321. Halstead, M.J.R., Cunninghame, R.G., Hunter, K.A., 2000. Wet deposition of trace metals to a remote site in Fiordland, New Zealand. Atmospheric Environment 34, 665–676. Hinkley, T.K., Lamothe, P.J., Wilson, S.A., Finnegan, D.L., Gerlach, T.M., 1999. Metal emission from Kilauea, and a suggested revision of the estimated worldwide metal output by quiescent degassing of volcanoes. Earth and Planetary Science Letters 170, 315–325. Hong, S., Candelone, J.P., Turetta, C., Boutron, C.F., 1996. Changes in natural lead, copper, zinc and cadmium concentrations in central Greenland ice from 8250 to 149 100 years ago: their association with climatic changes and resultant variations of dominant source contributions. Earth and Planetary Science Letters 143, 233–244. Imai, N., Terashima, S., Itoh, S., Ando, A., 1995. 1994 compilation of analytical data for minor and trace elements in seventeen GSJ geochemical reference samples, ‘‘Igneous rock series’’. Geostandards Newsletter 20, 165–216. Japan Meteorological Agency, 2007. /http://www.data.kishou. go.jp/etrn/index.htmlS. Accessed March 2007 (in Japanese). Kielhorn, J., Melber, C., Keller, D., Mangelsdorf, I., 2002. Palladium—a review of exposure and effects to human health. International Journal of Hygiene and Environmental Health 205, 417–432. Kim, G., Scudlark, J.R., Church, T.M., 2000. Atmospheric wet deposition of trace elements to Chesapeake and Delaware Bays. Atmospheric Environment 34, 3437–3444. Kreamer, D.K., Hodge, V.F., Rabinowitz, I., Johannesson, K.H., Stetzenbach, K.J., 1996. Trace element geochemistry in water from selected springs in Death Valley National Park, California. Ground Water 34, 95–103. Kulkarni, P., Chellam, S., Fraser, M.P., 2006. Lanthanum and lanthanides in atmospheric fine particles and their apportionment to refinery and petrochemical operations in Houston, TX. Atmospheric Environment 40, 508–520. Kulkarni, P., Chellam, S., Mittlefehldt, D.W., 2007. Microwaveassisted extraction of rare earth elements from petroleum refining catalysts and ambient fine aerosols prior to inductively coupled plasma-mass spectrometry. Analytica Chimica Acta 581, 247–259.

ARTICLE IN PRESS 7010

T. Shimamura et al. / Atmospheric Environment 41 (2007) 6999–7010

Luo, W., 2001. Wet-deposition fluxes of soluble chemical species and the elements in soluble materials. Atmospheric Environment 35, 2963–2967. Magnetic Resonance-Technology Information Portal, 2007. Contrast Agents. /http://www.mr-tip.com/serv1.php?type= coaS. Accessed on March 13 2007. Matthey, J., 2006. Platinum 2006. Platinum Metals Review 50, 143. /http://www.platinummetalsreview.com/dynamic/article/ view/50-3-143S. Accessed October 2006. Minoura, H., Takahashi, K., Chow, J.C., Watson, J.G., 2006. Multi-year trend in fine and coarse particle mass, carbon, and ions in downtown Tokyo, Japan. Atmospheric Environment 40, 2478–2487. Molycorp Inc., 1997. A lanthanide lanthology. /http://www. molycorp.com/lanthology_a-l.pdfS. Accessed October 2006. Mukai, H., Ambe, Y., Morita, M., 1990. Flow injection inductively coupled plasma mass spectrometry of platinum in airborne particulate matter. Journal of Analytical Atomic Spectrometry 5, 75–80. Nozaki, Y., Lerche, D., Alibo, D.S., Tsutsumi, M., 2000. Dissolved indium and rare earth elements in three Japanese rivers and Tokyo Bay: Evidence for anthropogenic Gd and In. Geochimica et Cosmochimica Acta 64, 3975–3982. Pirrone, N., Keeler, G.J., 1996. The Rouge River watershed pollution by trace elements: atmospheric depositions and emission sources. Water Science Technology 33, 267–275. Poissant, L., Be´ron, P., 1994. Parameterized rainwater quality model in urban environment. Atmospheric Environment 28, 305–310. Poissant, L., Schmit, J.P., Be´ron, P., 1994. Trace inorganic elements in rainfall in the Montreal Island. Atmospheric Environment 28, 339–346. Rauch, S., Hemond, H.F., Barbante, C., Owari, M., Morrison, G.M., Peucker-Ehrenbrink, B., Wass, U., 2005. Importance of automobile exhaust catalyst emissions for the deposition of platinum, palladium, and rhodium in the northern hemisphere. Environmental Science and Technology 39, 8156–8162. Reimann, C., Caritat, P.D., 1998. Chemical Elements in the Environment. Springer, Berlin. Reimann, C., Caritat, P.D., Halleraker, J.H., Volden, T., A¨yra¨s, M., Niskavaara, H., Chekushin, V.A., Pavlov, V.A., 1997. Rainwater composition in eight arctic catchments in northern Europe (Finland, Norway and Russia). Atmospheric Environment 31, 159–170. Rocha, F.R., Fracassi da Silva, J.A., Lago, C.L., Fornaro, A., Gutz, I.G.R., 2003. Wet deposition and related atmospheric

chemistry in the Sao Paulo metropolis, Brazil: Part 1. Major inorganic ions in rainwater as evaluated by capillary electrophoresis with contactless conductivity detection. Atmospheric Environment 37, 105–115. Sholkovitz, E.R., Church, T.M., Arimoto, R., 1993. Rare earth element composition of precipitation particles and aerosols. Journal of Geophysical Research 98, 20587–20599. Taylor, S.R., McLennan, S.M., 1985. The Continental Crust: Its Composition and Evolution. Blackwell Scientific Publications, Oxford, pp. 46. Tokyo Metropolitan Government, 2006. The environment of Tokyo 2006, /http://www2.kankyo.metro.tokyo.jp/kansi/taiki/ download/data_download_top.htmS. Accessed in March 2007. Turetta, C., Cozzi, G., Varga, A., Barbante, C., Capodaglio, G., Ceson, P., 2003. Platinum group elements determination in seawater by ICP-SFMS: initial results. Journal of Physics IV France 107, 1321–1324. Van de Velde, K., Barbante, C., Cozzi, G., Moret, I., Bellomi, T., Ferrari, C., Boutron, C., 2000. Changes in the occurrence of silver, gold, platinum, palladium and rhodium in Mont Blanc ice and snow since the 18th century. Atmospheric Environment 34, 3117–3127. Varga, A., Barbante, C., Cozzi, G., Mantovan, I., Rampazzo, G., Cescon, P., 2003. Platinum group element distribution in urban particulate matter from the Lagoon of Venice. Journal of Physics IV France 107, 1337–1340. Veysseyre, A., Moutard, K., Ferrari, C., Van de Velde, K., Barbante, C., Cozzi, G., Capodaglio, G., Boutron, C., 2001. Heavy metals in fresh snow collected at different altitudes in the Chamonix and Maurienne valley, French Alps: initial results. Atmospheric Environment 35, 415–425. Wong, C.S.C., Li, X.D., Zhang, G., Qi, S.H., Peng, X.Z., 2003. Atmospheric deposition of heavy metals in the Pearl River Delta, China. Atmospheric Environment 37, 767–776. Yamasaki, S., Tsumura, A., 1992. Determination of ultra-trace levels of elements in water by high resolution ICP-MS with an ultrasonic nebulizer. Water Science and Technology 25, 205–212. Zhang, J., Liu, C.Q., 2004. Major and rare earth elements in rainwaters from Japan and East China Sea: Natural and anthropogenic sources. Chemical Geology 209, 315–326. Zhang, J., Nozaki, Y., 1996. Rare earth elements and yttrium in seawater: ICP-MS determinations in the East Caroline, Coral Sea, and South Fiji basin of the western South Pacific Ocean. Geochimica et Cosmochimica Acta 60, 4631–4644.