&EPA
www.epa.gov/research
science in ACTION
INNOVATIVE RESEARCH FOR A SUSTAINABLE FUTURE
A Systematic Evaluation of Dissolved Metals Loss during Water Sample Filtration
Regional Applied Research Effort - Addressing Challenges through Science and Innovation
Project Purpose
This research study examined how water quality collection and filtration approaches,
including commonly used capsule and disc syringe filters, may cause losses in the
amounts of soluble lead and copper found in a sample. A variety of commercially
available filter materials with a pore size of 0.45 micrometers (u.m) were tested,
including polyvinylidene fluoride, polytetrafluoroethylene, nylon, polypropylene, and
mixed cellulose acetate. The effects of important water quality parameters including
pH, alkalinity and phosphate on sorption losses were examined. The impact of
filtration approach, including flow rate and sequential sampling, was also explored.
Background
EPA method 16691 provides specifications for the collection and filtration of water
samples prior to the analysis of dissolved and particulate trace elements and metals in
surface waters. Filtration apparatus specifications include the use of disposable,
tortuous path, capsule, and disc filters with an effective pore size of 0.45 u.m and a
diameter of 15 millimeter (mm) or larger.
Similar filtration approaches are commonly used to separate soluble and particulate constituents in drinking water samples,
bench and pilot-scale water treatment studies, and fundamental drinking water research studies. Syringe filters are typically
used in such cases given the ease of utilization. The Texas Commission on Environmental Quality (TCEQ) noted differences
in dissolved metal levels in water samples when different filter types were employed for particulate filtration in the field.
Specifically, discrepancies were found in dissolved metal fractions between those samples filtered using a low-volume
syringe filter and those filtered through an in-line capsulated cartridge filter. Similar inconsistencies have been noted by
EPA researchers performing filtration separations during fundamental drinking water research investigations. In the case of
TCEQ, many reported considerably higher lead and aluminum values for water samples filtered using syringe filters than
those reported when utilizing in-line capsule filters, often within similar water types found in similar geographic areas. This
action can confound a state's ability to confirm the relative quality of its data and the subsequent impairment status of its
waters. The findings have clear implications on the validity of dissolved metals measurements reported by a state's 303(d)2
lists, and the reliability of research data reported to have used similar filtration procedures.
Research Approach
Twelve different 25 mm diameter, 0.45 u.m filter types were tested (Table
1) and divided into two categories: those with no prefilter (8) and those
with a prefilter (4). The prefilters were either a glass microfiber prefilter
(PFG) or a polypropylene prefilter (PFP). According to the manufacturers,
the prefilters remove larger particles and are designed for high particle
loaded samples. Lead sorption filter tests were initially performed using
water containing 50 mg C/L dissolved inorganic carbon and 40 u.g/L lead at
pH 7. The lead concentration is representative of water that exceeds the
drinking water regulatory lead action level of 15 u.g/L. Average lead
concentrations of 15 ml sequential filtered water samples up to a total
volume of 90 ml were calculated from two or three separate tests and are
shown in Table 2.
Office of Research and Development
EPA/600/F-ll/029b | Updated August 2014
-------
Research Goals
1. Based on EPA method 16691, determine if the type of filter material used to separate soluble from particulate fractions
creates significant losses of dissolved metals (lead and copper) in water during filtration due to adsorption losses.
2. Evaluate the impact of water chemistry and other factors on sorption losses during water filtration.
3. Identify the best filtration material and protocol for separating soluble and particulate metals in water samples.
Research Results
The tendency for filters to adsorb dissolved lead were categorized as
(1) good (little to no adsorption or total lead lost), (2) moderate (some
adsorption and/or total lead lost), or (3) poor (large amounts of
adsorption and/or total lead lost). Lead sorption losses were also
presented as the 'total' lead (%Pb mass) loss from the water to the
filter over the entire 90 ml filtered volume. The results of this study
clearly show that soluble lead sorption losses to filters can be
significant and sometimes extreme. Based on this work, any syringe
filter with a prefilter should be avoided, especially those with a PFG.
(1) Of the filters evaluated, only the PP and MCE filters were
categorized as good performers and should be initially considered for
use. Under the conditions of this test, the results from using these two
filters suggests that there is no need to waste the initial volume
filtered. In practice, however, users should consider performing an
evaluation to define protocol that is most acceptable for their case.
Table 1. Manufacturer and types of 0.45 urn filters tested.
Manufacturer Filter Membrane Tvoe
A
A
A
A
B
B
B
B
B
B
B
B
B
Liquid Chromatography (LCR) Hydrophilic
Polytetrafluoroethylene (PTFE)
Ion Chromatography (1C) Hydrophilic PTFE
Nylon
Mixed Cellulose Esters (MCE)
Nylon
Polyvinylidene Fluoride (PVDF)
Polypropylene (PP)
Polyethersulfone (PES)
PFG Nylon
PFG Charged Nylon
PFG Glass Microfiber(GMF)
PFP Nylon
PFP PES
(2) The filter materials categorized as moderate performers were PES, LCR PTFE, 1C PTFE, PVDF, Nylon A, Nylon B, and PFP
PES. However, lead losses associated with these filters were generally within the first 30 to 45 ml of water filtered and
could possibly be minimized in some cases if an initial volume is wasted in practice.
(3) The filter materials categorized as poor performers were the PFG Nylon, PFG Charged Nylon, PFP Nylon, and PFG GMF.
Table 2. Summary of All Testing Results. Values are total percentage of lead lost to filter surface over 90 ml of filtered water.
Lead (24 ug/L] only
Replicate 1 Replicate 2 Replica-e 3 Average 95MC...
LCR PTFE
1C PTFE
MCE 2. 68 3.08 0.50 2.09 1.57
BNylon 19.95 9.94 14.45 14.78 5.67
PVDF
PP 1.26 3.94 3.02 2.74 1.54
PES 5.77 11.81 10.57 9.38 3.61
PFG Nylon -----
ANylon
PFG Charged Nylon - - -
PFP Nylon
PFP PES
PFG GMF -----
Percent LeodSorbed
Lead (40 ug/L) only
Replicaiel Replicaie2 Replicates Average
10.79
16.76
5.8O
15.27
21.84
5.45
6.20
99.46
21.44
99.67
49.00
15.69
99.40
12.48
18.10
7.58
21.24
17.44
838
7.49
99.79
23.26
99.86
52.83
15.55
99.36
9.84
10.76
1.44
13.40
5.58
99.52
30.09
99.60
43.89
4.22
99.21
11.04
15.20
4.94
16.64
19.64
6.91
6.42
99.59
24,93
99.71
48.57
11.82
99.32
95% C.I.
1.51
4.42
3.58
4.63
4.31
2.87
1.10
0.20
5.16
O.15
5.08
7.45
0.11
Lead (40 ug/L) and Calcium (100 mg/L )
Replicate 1 Replicate 2 Replicate 3 Average 95%C.I.
.
1.82 5.03 5.29 4.05
0.33 2.80 0.89 1.34
5.60 5.52 4.60 5.24
0.46 1.16 -0.43 0.40
4.65 3.70 3.36 3.91
48.06 80.85 86.64 71.85
_
.
.
.
^^H
2.19
1.46
0.63
0.90
0.76
23.54
-
Contacts
Principal Investigators: Darren Lytle, (513) 569-7432, lytle.darren@epa.gov | Keith Kelty, (513) 569-7414, kelty.keith@epa.gov|
Maily Priam, (513) 569-7212, pham.maily@epa.gov | Michael Schaub (214) 665-7314, schaub.mike@epa.gov
This work was performed under EPA's Regional Applied Research Effort (RARE) program with EPA Region 6.
1EPA Method 1669, Sampling Ambient Water for Trace Metals at EPA Water Quality Criteria Levels; U.S. Environmental Protection Agency, 1996.
2Clean Water Act Section 303(d); 33 U.S.C. §1251 et seq.; U.S. Environmental Protection Agency, 1972.
Office of Research and Development
EPA/600/F-ll/029b | Updated August 2014
------- |