Chemical Composition of Indoor and Outdoor Atmospheric Particles at Emperor Qin ' s Terra-cotta Museum , Xi ’ an , China

Indoor particles and microclimate were measured in summer (August 2004) and winter (January 2005) periods inside and outside Emperor Qin's Terra-Cotta Museum in Xi’an, China. Indoor temperature ranged from 21.9°C to 32.4°C in summer and from 0°C to 5.3°C in winter. Relative humidity varied from 56% to 80% in summer and from 48% to 78% in winter. The number concentrations of particles were lower (0.3–1.0 μm) in summer, and were higher (1.0–7.0 μm) in winter. The average indoor PM2.5 and TSP concentrations were 108.4 ± 30.3 μg/m and 172.4 ± 46.5 μg/m in summer and were 242.3 ± 189.0 μg/m and 312.5 ± 112.8 μg/m in winter, respectively. Sulfate, organic matter, and geological material dominated indoor PM2.5, followed by ammonium, nitrate, and elemental carbon. Several milligram of sulfate particles can deposited in the museum per square meter each year based on the dry deposition estimate. High concentrations of acidic particles suspended inside the museum and their depositions have high risk for the erosion of the terra-cotta figures.

Corresponding author.Tel.: 86-29-8832-6488; Fax: 86-29-8832-0456 E-mail address: cao@loess.llqg.ac.cnEmperor Qin's Terra-cotta Museum is one of China's best-known museums located 30 km east of Xi'an (34°44'N, 109°49'E), attracting more than 2 million visitors per year.It contains more than 7,000 pottery soldiers, horses, chariots, and weapons manufactured in the Qin Dynasty , of which ~1400 figures have been excavated, restored, and put on display.Since they were put on display in 1979, changes in their appearance have been observed, including changes in color, pitting, and reeling of the surfaces.These changes could be caused, in part, by high air pollutant concentrations inside the museum enclosure that might result from the infiltration of high outdoor air pollutants (Cao et al., 2005(Cao et al., , 2009;;Shen et al., 2009).This is a comprehensive report of indoor and outdoor particulate matter (PM) concentrations and chemical compositions inside and outside of museum enclosure.

Indoor and Outdoor Sampling
Three enclosures designated Pit 1, Pit 2, and Pit 3 were built over the excavated sites from 1974 to 1994.Pit 1 (14,269 m 2 ) was built first and contains 1,087 terra-cotta warriors.Pit 1 is similar to a large aircraft hanger with no environmental controls and ample ventilation with outside air.Samples were taken inside and outside of Pit 1, as shown in Fig. 1.Indoor samples were located at ~1.0 m above ground level in the statue restoration in the back of the museum (Fig. 2).The outdoor samples were located in rooftop of a two-story building near the Pit 1 (Fig. 1).

Mass Analysis
Quartz-fiber filters were analyzed gravimetrically for mass concentrations on a Sartorius MC5 electronic microbalance with ± 1 g sensitivity (Sartorius, Göttingen, Germany) after 24-hour equilibration at temperature between 20 and 23°C and RH between 35 and 45%.Each  filter was weighed at least three times before and after samplings, and the net mass was obtained by subtracting the difference between the averaged pre-and postsampling weights.

Elemental Analysis
Filters were analyzed for 15 elements (S, K, Ti, V, Mn, Fe, Ni, Cu, Zn, As, Se, Br, Sr, Zr, and Pb) by PIXE (Proton-induced X-ray emission spectroscopy) using 2.5 MeV protons with a 10 nA beam current through a 1.7 2 MV accelerator, and thin-film calibration standards (MicroMatter Co., Oregon) (Zhang et al., 2002).Blank filter background spectra were subtracted prior to peak integration.Eight standard reference material samples from the National Bureau of Chemical Exploration Analysis, China showed reasonable precision (< 10%) and accuracy (< 15%).

Organic Carbon and Elemental Carbon Analyses
0.5 cm 2 punch from each quartz filter was analyzed for organic carbon (OC) and elemental carbon (EC) by the IMPROVE thermal/optical protocol (Chow et al., 1993) using a Desert Research Institute (DRI, Reno, NV, USA) Model 2001 thermal/optical carbon analyzer (Atmoslytic Inc., Calabasas, CA, USA).Quality assurance/quality control (QA/QC) procedures were described by Cao et al. (2003).

Microclimate and CO 2
During summer, the difference between maximum and minimum daily temperatures was 3.2°C on August 15, 5.1°C on August 11, and 8.0°C on August 16 (Fig. 3).The daily average indoor temperature was 26.2°C, ranging from 21.9 to 32.4°C.The maximum temperature of 32°C was recorded from 13:40 to 15:40 on August 11.Average daily indoor temperature ranged from 0 to 5.3°C during winter.With the highest temperature of 5.3°C occurred from 13:00 to 15:00 and low temperatures occurred between 1:00 and 3:00.
Indoor RH was similar during summer and winter (Fig. 3).RH typically varied between 56 and 80%, with an average of 70.8% during summer, and between 48 and 78% in winter, with an average of 64%, consistent with the result of Zhang et al. (1997).Even though the museum is located in a semi-arid region of China, with an annual precipitation of 550 mm, the average indoor RH still reaches 70% during summer and 60% during winter.These levels are 15-20% higher than those measured at the Sainsbury Centre for Visual Arts, UK (Brimblecombe et al., 1999), and Correr museum, Italy (Camuffo et al., 1999).
Excess indoor CO 2 indicates exhaled breath from museum visitors, which range from 2322 to 8393 visitors/day.As shown in Fig. 3, CO 2 varied from 453 ppm to 741 ppm with an average of 534 ppm in summer and from 383 to 500 ppm with an average of 430 ppm in winter.Assuming an outdoor level of 380 ppm, near the minimum Fig. 3. 15-min average of temperature, RH, and indoor CO 2 variations for three days in summer and winter. of 383 ppm for indoor measurement, the average CO 2 from visitors was 155 ppm in summer and 51 ppm in winter.The summer/winter CO 2 ratio of 3.0 was similar to the summer/winter visitor number ratio of 3.6.

Size-separated Number Concentrations of Indoor Particles
Fig. 4 displays the size-segregated number concentrations of indoor particles observed on August 11, 15, and 16 in summer and January 8, 10, and 12 in winter.There were 7.1 × 10 5 particles/cm 3 in summer and 4.2 × 10 5 particles/cm 3 in winter for 0.3 m particles, 5.6 × 10 5 particles/cm 3 in summer and 4.1 × 10 5 particles/cm 3 in winter for 0.7 m particles, respectively.For the particles with 1.0-0.7 m, there were 2.2 × 10 5 , 1.2 × 10 4 , 3.8 × 10 3 , 5.8 × 10 3 particles/cm 3 in summer, and 2.6 × 10 5 , 1.8 × 10 4 , 5.6 × 10 3 , 7.4 × 10 3 , particles/cm 3 in winter, respectively, for 1.0, 2.0, 3.0, and 7.0 m particles.The ratios of particle number concentrations in winter relative to summer is 0.6, 0.7, 1.2, 1.5, 1.4, 1.3, for particle radius of 0.3, 0.7, 1.0, 2.0, 3.0, and 7.0 m, respectively, which indicated that the small particles in the range of 0.3-0.7 m were higher in summer, and large particles in the range of 1.0-7.0m were higher in winter.All the particles showed a distinct diurnal variation with a maximum during the daytime, and minimum at night, indicating a correlation with the trend of increasing tourist flow as reflected by the CO 2 levels (Figs. 3 and 4).

PM Chemical Composition
Average indoor PM 2.5 concentrations were 108.4 ± 30.3 g/m 3 in summer and 242.3 ± 189.0 g/m 3 in winter (Table 1).Average indoor TSP concentrations were 172.4 ± 46.5 g/m 3 in summer and 312.5 ± 112.8 g/m 3 in winter, while the average outdoor TSP concentrations were 226.0 ± 104.2 g/m 3 in summer and 367.6 ± 141.5 g/m 3 in winter.This result shows that the indoor PM 2.5 accounted for majority of total particles, with 62.9% in summer and 77.5% in winter, respectively.The indoor TSP is smaller than the outdoor TSP, accounting for 76.3% (summer) and 85.0% (winter) of outdoor TSP, respectively.The higher concentrations of winter outdoor TSP implied that winter outdoor particles have more impact on indoor particle concentrations than summer outdoor particles.The average mass concentrations of PM in winter were 1.6-2.2times compared with those in summer for PM 2.5 and TSP (Table 1).All daily PM 2.5 concentrations exceed 35 g/m 3 , the U.S. 24-hour National Ambient Air Quality Standards (NAAQS) limit.The average indoor TSP (172.4 g/m 3 ) in 2004 was less than that (430 g/m 3 ) measured in the summer of 1993, but the TSP in winter of 2004 (312.5 g/m 3 ) was similar to the measure in the winter of 1993 (340 g/m 3 ) (Zhang, 1997).

Elemental Composition
The highest concentration of the 15 elements was sulfur, for the cases of both indoors and outdoors, suggesting the heavy influence of anthropogenic sources.The average S concentration reached 10.4 to 12.9 g/m 3 for indoors and 11.6 g/m 3 for outdoors in summer and 17.1 to 20.6 g/m 3 for indoors and 23.5 g/m 3 for outdoors in winter.Average S concentrations in winter were 1.6-2.0times compared with those in summer for PM 2.5 and TSP.The indoor TSP S concentrations were one order of magnitude higher than the S concentrations measured in four European museums (Camuffo et al., 2001).The element K concentrations in winter were about 3 times higher than those in summer for PM 2.5 and TSP, pointed to the influence of biomass burning from domestic cooking and heating in winter around the museum.Other element concentrations in winter were similar or 1-2 times higher than those in summer.
Enrichment factors of each element are summarized in Table 2 relative to crustal rock using Fe (Taylor and Mclennan, 1985) as a reference element.EF = (X/Fe) air /(X/Fe) crust (1 where EF is the enrichment factor of element X, (X/Fe) air is the concentration ratio of X to Fe in the PM samples, and (X/Fe) crust is the average concentration ratio of X to Fe in crustal rock.
If EF-value is approaching unity, suspended dust is  probably the predominant source for element X.Table 2 shows that EF crust values for K, Ti, Mn, and Sr in summer are all close to unity with maximum values less than 5, indicating that these elements were dominated by crustal elements.But in winter, the EF of K and Mn in indoor PM 2.5 were larger than 5, consistent with the increasing contribution of anthropogenic sources like biomass and coal combustions.Compared to the dust-derived elements, the EF values for S, Ni, Cu, Zn, As, Se, Zr, and Pb, are considerably larger than 5, illustrating the influence of anthropogenic sources such as industrial activities (Zn, Cu, S), coal-combustion (Se, As, S), oil combustion (Ni, S) and motor vehicle exhaust (Pb, Zn).The elements of Se, S, and As are the most enriched elements.The highest EFvalue, ranging from 10 3 to 10 5 , was found for Se, S, and As, suggesting the domination of coal-burning emissions (Zhang et al., 2002).A major coal-fired power plant (i.e., Baqiao Thermo Power Plant) is located ~15 km east of the museum and emits 1.6 tons of fly ash per day, which could be a major contributor of the high concentrations of Se, S, and As.Average EFs for Pb and Zn in the range of 10 2 to 10 3 are related to motor vehicle exhausts.All the EFvalues of anthropogenic elements associated with the indoor PM 2.5 are large than those with the indoor TSP for both summer and winter (Table 2), implying the dominant presence of these elements in fine particles.
Water-soluble Ions SO 4 2-contributions were the highest of the ionic species with average concentrations of 36.6/43.1 and 58.9/71.1 g/m 3 , in summer and winter, respectively, for indoor PM 2.5 /TSP (Table 1).About 83-85% TSP SO 4 2-resided in PM 2.5 fraction for in summer and winter.A high SO 4 2fraction in fine mode is consistent with the observation of Gysels et al. (2002) in the Royal Museum of Fine Arts, Belgium.All the ionic concentrations in winter were higher than those in summer for both indoor and outdoor except TSP Ca 2+ , impling the serious pollution in winter.Compared with the observations of Santis (1992), summer and winter SO 4 2-, NO 3 -, and NH 4 + inside Xi'an museum were approximately 4-6 fold, 7-20 fold, and 8-20 fold, respectively, higher than those measured at Galleria degli Uffizi in Florence, Italy.
Eqs. ( 2) and ( 3) are used to calculate the charge balance between cation and anion.The correlations between cation and anion are summarized in Table 3. (3) The cation and anion concentrations measured for indoor PM 2.5 /TSP and outdoor TSP have correlation coefficients (R 2 ) greater than 0.9, suggesting the same origin of TSP in indoor and outdoor.The slopes of indoor PM 2.5 and TSP are larger than 1.3 in summer and close to 1.1 in winter.Since most inorganic species have been measured, the deficits of cation species indoors is consistent with the presence of a hydrogen ion (H + ), implying that the particles are acidic (Kerminen et al., 2001).Acidic particles inside the museum can potentially damage antiquities.The slope 0.93 of outdoor TSP in summer is close to 1.0, consistent with cation species being neutralized in the outdoor atmosphere.High concentrations of Ca 2+ in coarse mode contribute to the neutralization of excess sulfuric and nitric acidic particles in summer (Table 1).largest component with 24% of the mass in winter.Higher percentage of SO 4 2-in summer was associated with the formation of SO 4 2-under high temperature and high RH (Fig. 3) as well as high photochemical activities (such as high OH radical concentrations).The percentage of OM was almost same in summer (27.7%) and winter (27.5%), and it was the second largest component in summer and became to the largest component winter.The percentage of geological material likes the share distribution of OM, i.e., it accounted for 12.5% in summer and 13% in winter with the third largest component for two seasons.NO 3 accounted for 7% in summer and increased to 11% in winter, which may be due to more particulate NO 3 -resided in fine particles under lower temperature in winter.The percentages of NH 4 + (9%) and EC (3.2-3.9%)kept stable in summer and winter.
Compared with the each individual contribution to the total mass of indoor PM 2.5 , the contribution of geological matter was considerably higher in indoor TSP than in indoor PM 2.5 in both summer and winter.This result is due to the fact that the contribution of coarse particles (dust) was increased as the result of the resuspension of visitors' activities in the museum.The percentage of geological matter in winter (19.6%) was less than that in summer (29%), which is owing to the fewer visitors in winter (see section 3.1).
The sum of SO 4 2-, NO 3 -, and NH 4 + accounted for 47 and 44% in summer and winter, respectively, for indoor PM 2.5 , and 37 and 44% in summer and winter, respectively, for indoor TSP.Thus, the high percentage (37-47%) of secondary inorganic aerosol suspended in museum air may exert potential hazard to works of art posed by the deposition of airborne particles.Gysels et al. (2002) showed that organic compounds, S-rich and Fe-rich particles are considered as the most harmful factor to the works of art in a museum.
EC accounted for about 3% in PM 2.5 and TSP for two seasons, which also have negative impact on culture relics because it is a catalyst to accelerate the chemical reaction of acid species and a soiling and blacken contributor to the aesthetic values of terra-cotta figures.Elemental carbon had been found to yield perceptible soiling in the three southern California museums after accumulated on vertical surfaces (Nazaroff et al., 1990).In outdoor TSP, geological material accounted for the largest components both in summer and winter.Compared with indoor PM 2.5, the percentage of SO 4 2-increased from 15% in outdoor TSP to 31% in indoor PM 2.5 in summer, while it had similar level in winter.Since the Pit 1 is always open in summer and winter, the outdoor particles can penetrate into the museum easily.The large difference between indoor and outdoor of SO 4 2-in summer can be attributed to the formation of SO 4 2in the museum rather than the transport from outside of SO 4 2-.There has no clear difference for the percentage of NO 3 -between indoor PM 2.5 , TSP and outdoor TSP in both summer and winter (Fig. 5).However, there are some different between summer and winter.For example, NO 3 accounted for 6-7% of the mass in summer and 10-11% of the mass in winter.This suggested that the increasing formation of NO 3 -in winter was ascribed to the regional atmospheric chemistry rather than the indoor chemistry.Similar changes can be observed for NH 4 + with SO 4 2-, which can be ascribed to the accelerated conversion from indoor NH 3 to particulate NH 4 + in summer.Elevated NH 3 in summer may be associated with the use of fertilizers in the farmland around the museum as well as the contributions from local sanitation waste.Additional indoor ammonia can be attributed to the large emission of NH 3 from the high traffic of visitors (about 8,000 tourists per day in summer) and from the unvented indoor restrooms under high temperature.The percentage of NH 4 + in winter was similar in indoor PM 2.5 , TSP, and outdoor TSP, which can be ascribed to the regional transport.The indoor formation of NH 4 + was reduced by the lower indoor NH 3 emissions from the frozen farmland, the deceased visitors, and the lower atmospheric temperature in winter.

I/O Ratios of Chemical Species
The indoor/outdoor (I/O) ratios for measured chemical species in TSP are illustrated in Fig. 6 to evaluate the difference between indoor and outdoor concentrations.The I/O ratios of mass, geological matter, organic matter, EC, and NO 3 -were less than or close to 1.0, which indicated the indoor sources for these species were not significant.NH 4 + and SO 4 2-have high I/O ratios in summer.The I/O ratios for NH 4 + ranged from 0.8 to 1.8 with an average of 1.2, and the I/O ratios for SO 4 2-varied from 0.6 to 1.8 with an average of 1.1 in summer, which demonstrated they were affected by indoor pollutant sources.As mentioned above, the formations of NH 4 + and SO 4 2-inside museum were important in summer.The average I/O ratios of NH 4 + and SO 4 2-decreased to 1.0 and 0.95, which implied that there was almost no indoor formation for these two species in cold season.we didn't measure the deposition velocities of these species, we used the values from other studies.Ligocki et al. (1990) obtained the deposition velocities of SO 4 2-, NO 3 -, and NH 4 + were 38-290 × 10 -6 , 110-705 × 10 -6 , and 4.6-29×10 -6 m/s, respectively, in summer, and 74-1060 × 10 -6 m/s, 90-260 × 10 -6 m/s, and no value, respectively, in winter, based on five museums studies in Southern California.The dry deposition flux (F) for a species can be calculated as the product of its concentration (Ca) and deposition velocity (Vd):

Deposited Estimates of SO
Based on the measured TSP concentrations of SO 4 2-, NO 3 -, and NH 4 + in summer and winter, the calculated the deposition fluxes of SO 4 2-, NO 3 -, and NH 4 + were 52-394, 41-279, and 1.6-10 mg/m 2 /yr, respectively,in summer, and 166-2377, 95-274, 4-24 mg/m 2 /yr (using the deposition velocity in summer for calculation), respectively, in winter.From the above calculations, the deposition of SO 4 2particles was up to several mg per square meter in the museum each year, which can induce chemical reaction in the surface of art objects.Hu et al. (2006) observed the erosion evidence from the deposition of sulfate-rich particles in the surface of terra-cotta figures.

CONCLUSIONS
Through the investigation of air quality in the Emperor Qin's Terra-cotta Museum, Xi'an, China we can obtained the following findings.Fine particle and TSP levels were considerably high in both summer and winter.Sulfate, nitrate, carbon and geological matter were important components of particulate matter.Most of the indoor particles can be resulted from outdoor activities.Thus, high concentrations of acidic particles suspended inside the museum and their depositions have high risk for the erosion of the terra-cotta figures, especially under high variation of temperate and RH.In order to protect these culture relics, hazard removal devices should be considered to install for reducing the gas precursors like SO 2 and NH 3 in the future.

Fig. 1 .
Fig. 1.Locations of the sampling sites inside and outside of the Emperor Qin's Terra-cotta Museum in Xi'an, China.

Fig. 4 .
Fig. 4. Number size distributions of indoor particles for three days in summer and winter.

+
Fig. 6.I/O ratios for TSP mass and its major components (empty refer to summer, shade ref to winter).

Table 1 .
Statistical summary of 24-hour PM 2.5 and TSP chemical composition for samples acquired over summer and winter periods, indoors and outdoors at the Emperor Qin's Terra-Cotta Warriors and Horses Museum in Xi'an, China.

Table 2 .
Enrichment factor (relative to Fe) of selected element.