E ﬀ ects of Filter-Membrane Materials on Concentrations of Trace Elements in Acidic Solutions

: Filtration is commonly used to separate liquids from solid materials before the determination of trace element concentrations in aqueous solutions. Therefore, it is important to determine how much elemental content is extracted from the ﬁlter itself or lost via absorption into the ﬁlter. In this study, we investigated three types of disposable syringe ﬁlters (polyvinylidene diﬂuoride, PVDF; polytetraﬂuoroethylene, PTFE; PTFE with glass microﬁber preﬁltration media, GD / X) and their acid-cleaned versions using ammonium nitrate (NH 4 NO 3 ) and acetic acid (HOAc) solutions. The concentrations of most trace elements (Li, Al, Cr, Mn, Co, Ni, Cu, Zn, Rb, Sr, Mo, Cd, Cs, Ba, Pb, and U) were higher in ﬁltrates that had passed through acid-cleaned ﬁlters than those that had passed through uncleaned ﬁlters. In ﬁltrates from PVDF and PTFE ﬁlters, many trace elements were below the detection limit. However, regardless of the ﬁlter type and acid cleaning, Li, Cu, Zn, Mo, and Ba concentrations in ﬁltrates of NH 4 NO 3 solutions, and Al, Cr, Mn, Cu, Zn, and Ba concentrations in ﬁltrates of HOAc solutions were much higher than those of other trace elements. These di ﬀ erences were particularly noticeable in ﬁltrates from GD / X ﬁlters containing glass microﬁber layers. These data indicate that certain trace elements can be released from membrane materials by reacting with acid in the ﬁltered solution. Furthermore, the amounts of each trace element were associated with the membrane type. These ﬁndings suggest that ﬁlter type should be carefully selected to obtain the required accuracy, in consideration of the target elements and whether an acid-cleaned ﬁlter is needed. the trace element concentrations in the primary HOAc solution, as well as those in the PVDF and ﬁltrates, these results indicate that large amounts of material are extracted from GMF-layered membranes. Thus, GMF-layered ﬁlters are unsuitable for separation of acidic liquids from solid materials. and


Introduction
Filtration is a common procedure used in the field and in laboratory settings to separate dissolved ions from suspended materials. Filter paper and syringe filters are the most commonly used filtration media. However, the use of filter paper is labor-intensive and often more time-consuming than that of syringe filters. Additionally, filtration through syringe filters requires relatively low sample volumes. Although these advantages have led to the widespread use of syringe filters, the type of syringe filter should be carefully selected because some of the raw materials used to manufacture filtration membranes can leach analytes into the filtrate [1].
A syringe filter consists of a housing that is typically made of polypropylene (PP) and a filtration membrane made from a wide range of synthetic materials. Representative types of membrane materials include polyvinylidene difluoride (PVDF), polytetrafluoroethylene (PTFE), PP, polyethersulfone (PES), nylon, and glass microfiber (GMF). The chemical compatibilities of these membranes differ among acids and bases, so it is important to select a filter that is compatible with the system to avoid an effect on the experimental results. For example, PVDF, PTFE, and GMF membranes are resistant to acetic acid (HOAc), nitric acid (HNO 3 ), and hydrochloric acid (HCl), whereas PP and PES are not recommended for HNO 3 and PP has limited resistance to HCl [2].
To improve the analytical accuracy of methods that employ syringe filters, elements attached to the housing and membrane materials are often removed by acid cleaning [3][4][5]. Recently, Sakata et al. [6] suggested that filters should be pretreated via a heated acid cleaning process to decrease blank concentrations of certain elements. Using this filter pretreatment method and in consideration of the filter type, acid extractants (e.g., HNO 3 ) were passed through a variety of filter types. The resulting concentrations of trace elements in the blank solutions varied widely according to the filter material [1]. Many eluents can be used in liquid-solid separation, and it is important to determine the blank concentrations of trace elements resulting from the filtration process. Ammonium nitrate (NH 4 NO 3 ) and HOAc are common extractants used to determine dissolvable trace element contents in solid materials such as soil [7][8][9][10]. For example, in a recent study, exchangeable Sr fractions in soil were extracted with NH 4 NO 3 and HOAc to determine their concentrations and isotope ratios [11,12].
The objective of this study was to evaluate whether acid-cleaned filters are superior to uncleaned filters (i.e., new unused filters) in terms of experimental accuracy in determining trace metal concentrations when NH 4 NO 3 or HOAc are used as the extractant, and to determine the most suitable filter material for each situation. We performed extractions using three syringe filter types. For each filter type, blank concentrations of trace elements in the filtrates were compared between uncleaned and acid-cleaned syringe filters.

Methodology
All laboratory equipment used in this experiment was prewashed with 10% HCl, rinsed three times with tap water and Milli-Q ultrapure water (MQW), and completely dried in a clean area at room temperature. The diluted HCl was prepared using 35~40% HCl (electronic grade; Chemitop, Chungbuk, Korea).

Reagents
NH 4 NO 3 reagent (99.999% trace metal basis; Sigma-Aldrich, St. Louis, MO, USA) and HOAc (CH 3 COOH, >99% optimal grade; Fisher Scientific, Hampton, NH, USA) were used in this experiment. The reagents were diluted with MQW containing 1 M NH 4 NO 3 (80.04 g NH 4 NO 3 in 1 L MQW) and 1 M HOAc (57.3 mL HOAc in 1 L MQW). These concentrations are typically used to extract cations from solid materials such as soil. The diluted solutions were stored in Teflon vessels until analysis. The dilution and filtering steps were carried out on the same day.

Syringe Filters
Three types of 0.45 µm membrane syringe filters were evaluated in this experiment: PVDF, PTFE, and PTFE with GMF prefiltration media (GD/X). All filters were produced by the same manufacturer, eliminating any analytical errors due to differences among manufacturers. All filters featured a PP housing, were non-sterilized, and had a pore size of 0.45 µm. According to the chemical compatibility of the membranes and housings [13], all filters were resistant to the acids HNO 3 , HOAc, citric acid, and concentrated HCl. The GD/X filter consists of GMF 150 and GF/F layers produced from 100% borosilicate GMF on the top, with membranes formed of materials such as PVDF, nylon, or PTFE on the bottom. The GMF 150 layer retains particles sized 1-10 µm in diameter, and the GF/F layer traps particles down to 0.7 µm. The eluate from these prefiltration media flows directly onto the PVDF membrane filter. The PVDF and GD/X materials are hydrophilic, whereas the PTFE filter is hydrophobic.

Preparation of Acid-Cleaned Filters
For each filter type, three filters were acid-cleaned as follows. The filters were removed from sealed packages and soaked in a beaker containing 10% ultrapure HCl and approximately 30 mL Water 2020, 12, 3497 3 of 12 acid was passed through the filters manually once. The filters were maintained in the acid solution overnight at room temperature, without a change of acid, and then rinsed at least three times with tap water and MQW. To completely remove any remaining HCl, 50 mL MQW was manually pressed through each membrane. The filters were dried completely in a clean booth at room temperature.

Sampling and Analysis
The primary NH 4 NO 3 and HOAc solutions were passed through the uncleaned and acid-cleaned filters using 10 mL syringes. Aliquots of approximately 1 g filtrate were directly transferred into 15 mL Falcon tubes on a microbalance (ME204, Mettler Toledo, Switzerland). This procedure was performed 10 times in parallel; thus, a total of approximately 10 g solution was extracted through each membrane filter and transferred into 10 tubes in 1 g fractions. Sampling was performed in triplicate for each filter type. Trace element contents (Li, Al, Cr, Mn, Co, Ni, Cu, Zn, Rb, Sr, Mo, Cd, Cs, Ba, Pb, and U) in primary acid solutions and filtrates were analyzed using inductively coupled plasma mass spectrometry (X-Series; Thermo Elemental, Cambridge, UK) at the Korea Basic Science Institute.

Changes in Filtrate Trace Elemental Concentrations
To quantitatively calculate changes in the trace element concentrations of both solutions after passing through the filters (loss or gain), the elemental concentrations were compared to those of each primary solution using the following equation: where a and b are the concentrations of each trace element in the primary solution and filtrate, respectively.

Results and Discussion
Filtrate solutions were passed through three syringe filter types (PVDF, PTFE, and GD/X) with and without prior acid-cleaning. Table 1 lists the concentrations of 16 trace elements (Li, Al, Cr, Mn, Co, Ni, Cu, Zn, Rb, Sr, Mo, Cd, Cs, Ba, Pb, and U) in the resulting eluates. Trace elements were analyzed in the primary NH 4 NO 3 and HOAc solutions, and Table 1 also lists their initial concentrations. Several trace elements were found in the primary NH 4 NO 3 and HOAc solutions. For example, the primary NH 4 NO 3 solution contained 56.1 µg L −1 of Mo and 18.4 µg L −1 Li. Al, Cr, Mn, Cu, Zn, and Ba were also present in amounts less than 10 µg L −1 . In the NH 4 NO 3 solution, Co, Ni, Rb, Sr, Cd, Cs, Pb, and U were below the limit of detection. In the HOAc solution, Al, Cr, Mn, and Zn were detected below 5 µg L −1 . Other trace elements were not detected. Table 1. Concentrations of trace elements in ammonium nitrate and acetic acid extracted from three uncleaned and acid-cleaned types of filters (unit in µg L −1 ).

Filtrates from Uncleaned Filters
Filtrates obtained from the three filter types contained Zn (175-686 µg L −1 ) and Ba (32.3-1492 µg L −1 ) at very high concentrations compared to the primary solutions (Table 1). By contrast, trace elements such as Al, Co, Cs, and U were not detected in most samples, regardless of the filter type (Table 1). Rb and Pb were not detected in samples filtered through PVDF, and Pb was not found in PTFE filtrates. Filtrates from the GD/X filters contained Cr, Mn, Ni, Rb, Cd, and Pb at concentrations lower than 0.5 or 1 µg L −1 . Among the detected elements, Mn, Cu, Li, and Mo had lower concentrations in the filtrate than in the primary solution, and the elemental concentrations were lower on average, within the range of −79.6% to −5.7%, based on Equation (1) (Figure 1). Similarly, Berg et al. [1] found that filtrates from PTFE filters tended to contain lower concentrations of most trace elements. Thus, trace element concentrations in NH 4 NO 3 filtrates decreased after filtration due to a loss of elements to the filter material. Recently, Hedberg et al. [14] reported that trace metals, such as Zn and Cu, can be adsorbed by and interact with PVDF and PTFE membranes and that these elements can be released from the membrane during filtration. Therefore, higher concentrations of elements in filtrates compared to the primary solution can be attributed to dissolution of the filter itself. Lower elemental concentrations in the filtrate can be attributed to the absorption or adsorption of elements into or onto the filter, respectively.

Filtrates from Acid-Cleaned Filters
Filtrate concentrations obtained from acid-cleaned filters were higher than those from uncleaned filters for most of the detected elements (Table 1). Similarly, acid-cleaned filters increased the concentrations of trace element levels in the blank to a greater extent than uncleaned filters, likely due to acid-induced filter damage. In particular, Zn and Ba concentrations in filtrates from acidcleaned GD/X filters were much higher than those from uncleaned filters, up to 49-fold for Zn and 103-fold for Ba. These results differ from those of a previous study, which reported that blank levels of certain trace metals (Cd, Co, Cr, Cu, Mn, Mo, Ni, Pb, V, and Zn) in the filtrate were lowered to their limits of detection after passing through filters (PVDF membrane with PP prefiltration media) that had been acid-cleaned with HNO3 [15]. Additionally, according to the manufacturer, the membranes and housings of the three filter types are chemically compatible with concentrated HCl and HNO3, and up to 5% HOAc [2].
The high trace metal concentrations detected in filtrates from acid-cleaned filters may be explained by (1) contamination of the 10% HCl and MQW that were used to rinse the raw filter, and

Filtrates from Acid-Cleaned Filters
Filtrate concentrations obtained from acid-cleaned filters were higher than those from uncleaned filters for most of the detected elements (Table 1). Similarly, acid-cleaned filters increased the concentrations of trace element levels in the blank to a greater extent than uncleaned filters, likely due to acid-induced filter damage. In particular, Zn and Ba concentrations in filtrates from acid-cleaned GD/X filters were much higher than those from uncleaned filters, up to 49-fold for Zn and 103-fold for Ba. These results differ from those of a previous study, which reported that blank levels of certain trace metals (Cd, Co, Cr, Cu, Mn, Mo, Ni, Pb, V, and Zn) in the filtrate were lowered to their limits of detection after passing through filters (PVDF membrane with PP prefiltration media) that had been acid-cleaned with HNO 3 [15]. Additionally, according to the manufacturer, the membranes and housings of the three filter types are chemically compatible with concentrated HCl and HNO 3 , and up to 5% HOAc [2].
The high trace metal concentrations detected in filtrates from acid-cleaned filters may be explained by (1) contamination of the 10% HCl and MQW that were used to rinse the raw filter, and (2) leaching of elements from the filter materials during acid cleaning. The first explanation is unlikely because all of the laboratory equipment used in this study was acid-cleaned using the same acid solution. Additionally, Zn and Ba were found in high concentrations in both uncleaned and acid-cleaned filters. The particularly high levels of trace elements in filtrates from acid-cleaned filters suggests that the second explanation is more reasonable. In a previous study [1], filtrates from different filter membrane materials (e.g., PTFE, PVC, polysulfone, polycarbonate, and glass fiber membranes) contained wide ranges of elemental concentrations following extraction with 0.2 M HNO 3 . In particular, the amounts of Zn and Ba in glass fiber membrane filtrates ranged from 11 to 36 µg and from 15 to 44 µg, respectively. These levels were much higher than those obtained using other membrane filters. For example, PTFE filtrates contained 0.003-0.018 µg of Zn and 0.002-0.004 µg of Ba per filter. Similarly, in the present study, high Zn and Ba concentrations were only observed in acid-cleaned GD/X filters, which contained GMF. Moreover, membrane-derived contamination can be mobilized during acid-cleaning procedures [14] and leach into the filtrate.
Among the other elements (e.g., Al, Cr, Li, Mn, Cu, and Mo) detected in the primary solution, Li concentrations in filtrates from PVDF and PTFE were slightly higher than those in the primary solution (Table 1). Li concentrations were <6% higher than those of the primary solution. Similar increases in concentration were observed for Cu in filtrates from PTFE (<17%) and GD/X (<40%), and for Mn from GD/X (<23%) (Figure 1). The remaining elemental concentrations were lower than those of the primary solution. Additionally, Mo in filtrates from all filter materials decreased from -50.9% to −46.5%. These results indicate that either gain from or loss to filter materials affects some filtrate trace elemental concentrations.

Filtrates from PVDF and PTFE Filters
Regardless of pretreatment, the concentrations of 10 elements (Li, Co, Ni, Rb, Sr, Mo, Cd, Cs, Pb, and U) in most HOAc samples that had been passed through PVDF and PTFE filters were below the detection limits. Mo was detected in some PTFE filtrates, and Sr, Cd, and Cs were found in several samples (Table 1). Mn, Cu, and Ba were not detected in samples that had been filtered through uncleaned filters. Only three elements (Al, Cr, and Zn) were detected in samples that had passed through uncleaned PVDF and PTFE filters (Table 1). Due to losses of Al (−17.4%) and Zn (-19.2%) to uncleaned PVDF filter materials, their concentrations were lower in the filtrates, whereas the filter contributed to an increase in Cr (33.8%) (Figure 2). In filtrates from PTFE, gains of Cr (36.6%) and Zn (59.4%) from the filter were detected, whereas the loss of Al (-89.3%) produced slightly lower concentrations in the filtrates than in the primary HOAc solution. By contrast, several elements (Al, Cr, Mn, Cu, Zn, and Ba) were detected in filtrates from acid-cleaned PVDF and PTFE filters ( Table 1). The elemental concentrations were as high as 60-80% for Cr, 500-760% for Al, 490-650% for Mn, and 1300-1700% for Zn, compared to those in the primary HOAc solution ( Figure 2). As observed in the NH 4 NO 3 solutions, filtrates obtained from acid-cleaned filters contained markedly elevated concentrations of trace elements (Table 1).

Filtrates from GMF
The concentrations of 10 elements (Li, Co, Ni, Cu, Rb, Mo, Cd, Cs, Pb, and U) in many filtrates from uncleaned GD/X filters were below their detection limits. Only eight elements (Li, Co, Ni, Mo, Cd, Cs, Pb, and U) were below their detection limits in filtrates from acid-cleaned filters, and even fewer (Ni, Cu, Rb, Mo, and Cs) were below their detection limits in filtrates from uncleaned filters after sampling (~4 g). Close examination of three individual acid-cleaned filters revealed that Rb was detected in all filtrates from one of the filters, but was not detected after the third sampling from the second filter. Rb was not detected in filtrates from the third filter ( Figure 3). Similar results were obtained for Pb in NH 4 NO 3 filtrates from uncleaned filters (data not shown). These results indicate that elements that originate from the membrane are not uniformly extracted from every filter.
Compared to other filters, filtrates from both uncleaned and acid-cleaned GD/X filters contained significantly higher concentrations of Al, Cr, Mn, Zn, Sr, and Ba (Table 1). Filtrates from the acid-cleaned GD/X filters also contained high levels of Cu. In particular, Al, Zn, Ba, and Sr concentrations were much higher than those of PVDF and PTFE filtrates, by up to four orders of magnitude (Table 1). Compared to Al, Cr, Mn, and Zn detected in the primary HOAc solution, elemental concentrations in filtrates from both uncleaned and acid-cleaned GD/X filters were as high as 26,400-26,900% for Al, 65-160% for Cr, 790-2100% for Mn, and 35,200-136,000% for Zn, compared to those of the primary HOAc solution (Figure 2). Considering the trace element concentrations in the primary HOAc solution, as well as those in the PVDF and PTFE filtrates, these results indicate that large amounts of material are extracted from GMF-layered membranes. Thus, GMF-layered filters are unsuitable for separation of acidic liquids from solid materials.
Water 2020, 12, x 7 of 10 detected in all filtrates from one of the filters, but was not detected after the third sampling from the second filter. Rb was not detected in filtrates from the third filter ( Figure 3). Similar results were obtained for Pb in NH4NO3 filtrates from uncleaned filters (data not shown). These results indicate that elements that originate from the membrane are not uniformly extracted from every filter. Compared to other filters, filtrates from both uncleaned and acid-cleaned GD/X filters contained significantly higher concentrations of Al, Cr, Mn, Zn, Sr, and Ba (Table 1). Filtrates from the acidcleaned GD/X filters also contained high levels of Cu. In particular, Al, Zn, Ba, and Sr concentrations were much higher than those of PVDF and PTFE filtrates, by up to four orders of magnitude (Table 1). Compared to Al, Cr, Mn, and Zn detected in the primary HOAc solution, elemental concentrations in filtrates from both uncleaned and acid-cleaned GD/X filters were as high as 26,400-26,900% for Al, 65-160% for Cr, 790-2100% for Mn, and 35,200-136,000% for Zn, compared to those of the primary HOAc solution (Figure 2). Considering the trace element concentrations in the primary HOAc solution, as well as those in the PVDF and PTFE filtrates, these results indicate that large amounts of material are extracted from GMF-layered membranes. Thus, GMF-layered filters are unsuitable for separation of acidic liquids from solid materials.

Variation in Elemental Concentrations during Sequential Extraction
Regardless of acid pretreatment, elemental concentrations in both NH4NO3 and HOAc filtrates that had passed through PVDF and PTFE filters showed little change from those of the primary solution. By contrast, the levels of certain elements in uncleaned and acid-cleaned GD/X filtrates were high in the first few sample aliquots, and then gradually decreased with sequential sampling (Figure 3). Decreases in Cr, Mn, and Rb concentration are shown as representative examples; these decreases were more pronounced in filtrates from acid-cleaned GD/X filters. These results support our

Variation in Elemental Concentrations during Sequential Extraction
Regardless of acid pretreatment, elemental concentrations in both NH 4 NO 3 and HOAc filtrates that had passed through PVDF and PTFE filters showed little change from those of the primary solution. By contrast, the levels of certain elements in uncleaned and acid-cleaned GD/X filtrates were high in the first few sample aliquots, and then gradually decreased with sequential sampling (Figure 3). Decreases in Cr, Mn, and Rb concentration are shown as representative examples; these decreases were more pronounced in filtrates from acid-cleaned GD/X filters. These results support our hypothesis that certain elements are leached from the raw materials used in GD/X filters and that acid-induced damage during acid-cleaning further exacerbates this problem. Furthermore, when approximately 3-4 g samples of the acidic solutions were passed through the uncleaned GD/X filters, concentrations of trace elements in the filtrate dramatically decreased to approximately the levels of the blank, suggesting that a flushing step should be conducted to prevent the gain of trace elements from uncleaned GD/X filters when filtering an acidic solution.

Filter Selection
Based on the results of this study, filters should be carefully selected to avoid contamination from the filter itself. For example, exchangeable Sr and carbonate Sr fractions in soil have been extracted using NH 4 NO 3 and HOAc, respectively, to determine Sr isotope ratios [11,12,16,17]. After incubating the soil samples in each of the extraction solutions, the samples were filtered to separate the liquid from the solids. Considering that Sr and Rb were not detected in the primary NH 4 NO 3 and HOAc solutions, PVDF would be the best choice of filter material for these chemical and isotopic analyses.
Rb concentrations were below the detection limit with both uncleaned and acid-cleaned filters ( Figure 3, Table 1), and only small amounts of Sr (0.05-0.43 µg L −1 ) were detected in filtrates from uncleaned filters ( Figure 4, Table 1). Alternatively, an acid-cleaned PTFE filter may be used to analyze Sr concentrations and isotopic ratios, as long as the filter membrane is rinsed with a small amount of sample, although this was an inadequate strategy for uncleaned filters (Figure 4, Table 1). Most filtrates from uncleaned PTFE filters contained Rb and Sr concentrations of <0.5 µg L −1 . Glass-based membrane filters, both uncleaned and acid-cleaned, were deemed unsuitable for analyzing Sr concentrations and isotope ratios, although the concentration decreased sharply after flushing with~4 g fractions ( Figure 4); for example, NH 4 NO 3 filtrates from uncleaned and acid-cleaned GD/X filters had mean Sr concentrations of 6.91 ± 2.11 and 7.08 ± 2.48 µg L −1 for 5-10 g fractions, respectively, and HOAc filtrates included the Sr concentrations of 8.51 ± 2.69 and 8.11 ± 2.59 µg L −1 , respectively. Thus, Sr and Rb concentrations in the filtrates were much higher than those in the primary solutions (Figures 3 and 4). We strongly recommend that PVDF and PTFE filters be used when separating bioavailable Sr fractions from soil mixtures.
Based on the results of this study, filters should be carefully selected to avoid contamination from the filter itself. For example, exchangeable Sr and carbonate Sr fractions in soil have been extracted using NH4NO3 and HOAc, respectively, to determine Sr isotope ratios [11,12,16,17]. After incubating the soil samples in each of the extraction solutions, the samples were filtered to separate the liquid from the solids. Considering that Sr and Rb were not detected in the primary NH4NO3 and HOAc solutions, PVDF would be the best choice of filter material for these chemical and isotopic analyses. Rb concentrations were below the detection limit with both uncleaned and acid-cleaned filters ( Figure 3, Table 1), and only small amounts of Sr (0.05-0.43 μg L −1 ) were detected in filtrates from uncleaned filters ( Figure 4, Table 1). Alternatively, an acid-cleaned PTFE filter may be used to analyze Sr concentrations and isotopic ratios, as long as the filter membrane is rinsed with a small amount of sample, although this was an inadequate strategy for uncleaned filters (Figure 4, Table 1). Most filtrates from uncleaned PTFE filters contained Rb and Sr concentrations of <0.5 μg L −1 . Glassbased membrane filters, both uncleaned and acid-cleaned, were deemed unsuitable for analyzing Sr concentrations and isotope ratios, although the concentration decreased sharply after flushing with ~4 g fractions ( Figure 4); for example, NH4NO3 filtrates from uncleaned and acid-cleaned GD/X filters had mean Sr concentrations of 6.91 ± 2.11 and 7.08 ± 2.48 μg L −1 for 5-10 g fractions, respectively, and HOAc filtrates included the Sr concentrations of 8.51 ± 2.69 and 8.11 ± 2.59 μg L −1 , respectively. Thus, Sr and Rb concentrations in the filtrates were much higher than those in the primary solutions (Figures 3 and 4). We strongly recommend that PVDF and PTFE filters be used when separating bioavailable Sr fractions from soil mixtures.

Summary
In this study, we evaluated PVDF, PTFE, and GD/X syringe filters for use in trace element quantitation in solutions containing strong acid, NH4NO3, or HOAc. Half of the filters were

Summary
In this study, we evaluated PVDF, PTFE, and GD/X syringe filters for use in trace element quantitation in solutions containing strong acid, NH 4 NO 3 , or HOAc. Half of the filters were uncleaned and the rest were pretreated by acid cleaning with 10% HCl. Regardless of the type of filter or solution, filtrates that had passed through GD/X filters contained much higher concentrations of trace elements (Li, Al, Cr, Mn, Co, Ni, Cu, Zn, Rb, Sr, Mo, Cd, Cs, Ba, Pb, and U) than those that had been passed through other filter types. Overall, regardless of filter type, filtrates from acid-cleaned filters contained higher levels of trace elements than those from uncleaned filters. By contrast, trace element concentrations in filtrates from PVDF and PTFE filters were below their detection limits. These results suggest that PVDF-and PTFE-based filters are more suitable than GD/X filters for extracting trace elements from solid materials using acid. However, specific elements (Sr and Rb) were below their detection limits only when acid-cleaned PVDF filters were used. Therefore, GD/X filters should be carefully selected according to the purpose of the experiment. Based on sequential extractions using GD/X filters, removing the initial 3-4 g of filtrate may solve the problem of trace elements being released from membrane materials through a reaction with the acid solution.