PROGRAMS AND PLANS--Evaluation of Capsule Filters




In Reply Refer To:                              January 21, 1993
Mail Stop 412


OFFICE OF WATER QUALITY TECHNICAL MEMORANDUM 93.05

Subject: PROGRAMS AND PLANS--Evaluation of Capsule Filters

       EVALUATION OF TRACE-ELEMENT AND NUTRIENT CONTAMINATION
                 ASSOCIATED WITH CAPSULE FILTERS

                             Synopsis
An experiment was conducted to determine if capsule filters can be
precleaned to be usable in a part-per-billion (ppb) protocol for
dissolved trace elements.  Based on the experimental results, the
Gelman Supor 1/ capsule filters (Catalogue #12175) prerinsed/
conditioned with 1,000 milliliters (mL) of deionized water (DIW)
are suitable for ppb-level determinations of trace elements, and
also appear noncontaminating for nutrients at present reporting
limits.  At the ppb level for trace elements and present nutrient
reporting limits, acid prerinsing the capsule filters does not
appear to have any beneficial effect over the use of DIW.

                            Background

As indicated in Office of Water Quality (OWQ) Technical Memorandum
92.12, the OWQ has been conducting a series of studies designed to
identify equipment, supplies, and cleaning procedures suitable for
a ppb protocol for dissolved trace elements.  Traditionally, the
Water Resources Division (WRD) has processed whole-water samples
(for the subsequent determination of dissolved constituents)
through 0.45-um membrane plate filters.  The continued use of such
filters for the determination of dissolved trace elements at the
ppb level is certainly feasible, but will entail much greater care
than has been used in the past.  The additional care will be
required to limit potential sources of contamination associated
with the filtration process.  Some examples of contamination
sources include:  inadequate cleaning of the membrane filter and
filtration system, excessive or improper handling of the filter
membrane, atmospheric inputs during the filtration process, and
carryover from one sample to another due to improper between-
sample cleaning of the filter system.

1/ Any use of trade, product, or firm names is for descriptive
purposes only and does not imply endorsement by the U.S.
government.

Potential contamination during the filtration process can be
minimized if cleaning and handling requirements are kept to a
minimum, if the filtration system is used only once, and if the
filtration system is a sealed unit until used.  One way to address
these requirements is to employ disposable (used only once)
capsule filters.  This would be especially true if the capsule
filters were precleaned and prepackaged at a dedicated facility
prior to taking them to the field.  Because a number of capsule
filters are commercially available, three questions must be
addressed:

1.  Which filter(s) should be used?

2.  What concentration of trace-element and nutrient contaminants
    are associated with these filters?

3.  If necessary, could the contaminants be removed using
    relatively simple cleaning procedures?

             Selection of Capsule Filters for Testing

OWQ Technical Memorandum 80.22 recommended the use of Gelman Mini
Capsule filters (Catalogue #12123) for processing whole-water
samples containing high concentrations of suspended sediment.  The
reason for the recommendation was that these filters were
substantially harder to clog than normal plate filters (142 mm
diameter, 0.45 um pore size) because they had much higher surface
areas (about 500 cm2 as opposed to 160 cm2).  The Gelman Mini
Capsule filters are 47-mm in diameter, with a 0.45-um pore size
and contain a Versapor (acrylic copolymer on a nylon support)
membrane on a polypropylene core, in a sealed polycarbonate
housing.  More recently, Windom and others (1991) used capsule
filters to process whole-water samples to determine selected
dissolved trace-element concentrations at the part-per-trillion
(ppt) level in a number of east coast rivers.  They used Gelman
47-mm, 0.45-um capsule filters (Catalogue #12175) that contain a
Supor (polysulfone) membrane on a polypropylene core, in a sealed
polycarbonate housing.  These filters have an effective surface
area of 600 cm2.

Physically, both capsule filters are similar, albeit the Mini
Capsule filter has a slightly smaller surface area.  However, the
filter membranes are chemically dissimilar and have different
chemical compatibilities.  Traditionally, trace-element chemists
and geochemists working at low concentration levels use a weak
acid rinse to preclean (decontaminate) filters prior to use.  Some
use a weak nitric acid (Shiller and Boyle, 1987) rinse; others use
a weak hydrochloric acid (Windom and others, 1991) rinse.  The WRD
has traditionally used native water to clean and condition filters
prior to use.  However, it is very likely that any filtrations
carried out with the intent of determining trace-element
concentrations at the ppt level will require some form of acid
washing.

Currently, the OWQ is preparing a ppb-level trace-element
protocol.  Ultimately, a companion ppt protocol will be prepared
when the National Water Quality Laboratory (NWQL) develops the
appropriate analytical capability.  Therefore, the ideal filter
media chosen for divisionwide use should be amenable to either
ppb- or ppt-level protocols.  Conversations with Gelman technical
representatives indicated that the Versapor filter is incompatible
with acid of any sort, whereas the Supor filter can not resist
nitric acid at any strength, but can resist hydrochloric acid
(even at full strength).  Thus, based on the expected potential
requirement for acid precleaning in a ppt-level protocol, only the
Supor filter was evaluated for possible sources of contamination.

                         Testing Procedure

The testing and analysis of capsule filters were carried out at
the NWQL under tightly controlled conditions.  The actual
cleaning/conditioning tests were performed within a laminar-flow
hood to limit atmospheric inputs.  The test employed two
cleaning/conditioning solutions: (1) DIW, and (2) 0.6 molar
hydrochloric acid followed by DIW. Five separate capsule filters
were used to test each cleaning/conditioning solution, as follows:

1.  DIW cleaning/conditioning solutions:  The filters were used as
    they came from the supplier; there was no precleaning
    procedure employed.  Each filter was sequentially flushed with
    four 500-mL aliquots of DIW using a peristaltic pump and
    silicon tubing.

2.  0.6 molar HCl-DIW cleaning/conditioning solutions:  Each
    filter was filled to capacity with 0.6 molar HCl.  Pump tubing
    was attached to each end of the capsule filter and the tubing
    was inserted in a peristaltic pump.  The acid solution was
    circulated for about 1 minute.  At the end of that time, the
    filter was drained of all acid, and DIW was pumped through the
    filter until the pH of the DIW returned to normal (as
    determined by pH measurements).  This was done to simulate the
    post acid washing with DIW that would be required with
    capsules being prepared for actual field use.  The additional
    step is necessary because acid left on a filter during
    filtration of an actual sample could solubilize trace elements
    from the particulate matter retained on the filter.  In most
    cases, about one liter of DIW was required to return the pH to
    normal.  After this preparatory step, each filter was
    sequentially flushed with four 500-mL aliquots of DIW using a
    peristaltic pump and silicon tubing.

In summary, each of the ten test filters was flushed with four
sequential 500-mL aliquots of DIW.  These aliquots were collected
and stored in acid-rinsed polypropylene bottles and preserved with
nitric acid prior to trace element analysis.  The filter tests
themselves accounted for a total of 40 samples (5 filters x 2
cleaning/conditioning solutions [DIW, HCl/DIW] x 4 [500-mL]
aliquots) for trace-element analysis.

In addition to the trace-element tests, the capsule filters were
also evaluated for utility in processing samples for the
subsequent determination of a variety of nutrient parameters.
Separate aliquots from each of the 40 trace-element samples
described above were collected, stored, and preserved in
appropriate containers for nutrient analyses.

During the course of the tests, quality control data were
collected for both trace-element and nutrient analyses.  Fourteen
DIW blanks were collected; seven consisted of DIW poured directly
into the appropriate sample containers from a reservoir filled
prior to the initiation of testing, and seven consisted of DIW
from the same reservoir which was run through a peristaltic pump
and pump tubing, and collected in the appropriate sample
containers (this DIW did not pass through the capsule filters).
Two atmospheric blanks (open containers of DIW kept within the
laminar flow hood) were collected during the course of the entire
test.  In addition, 10 samples from the capsule evaluation (from
the second 500-mL rinse aliquot) were split to determine
analytical precision.  A total of 66 samples (40 capsule filter
rinses + 14 DIW blanks + 2 atmospheric blanks + 10 splits) were
analyzed for trace-elements and 66 samples for nutrients.

                       Results and Discussion

Table 1 presents data for the various blanks, and the sequential
flushes for the capsule filters study.  Trace-element analyses
were carried out either by ICP-MS or ICP-AES techniques.  The
specific elements and their reporting limits for the ICP-MS
determinations were:
     0.5+/-1.0 5ug/L - Al and Zn
     0.2+/-0.2 5ug/L - Sb, Ba, Be, Cd, Co, Cr, Cu, Mn, Mo, Ni, Pb, Ag,
                       and Tl

The specific elements and their reporting limits for the ICP-AES
determinations were:
     Fe (3 ug/L), Sr (0.5 ug/L), B (2 ug/L), Ca (0.02 mg/L),
     Mg (0.01 mg/L), Na (0.2 mg/L), and Si (0.01 mg/L).

Nutrient analyses were performed using standard techniques.
Reporting limits for the various nutrient determinations were:
     total phosphorus (0.01 mg/L), total nitrogen (0.05 mg/L),
     ammonium ion (0.02 mg/L), orthophosphate (0.01 mg/L),
     nitrate plus nitrite (0.05 mg/L), and nitrite (0.01 mg/L).

All results, both in the blanks and in the sequential filter
flushes, were below the respective reporting limits for:
*  ICP-MS limit of 0.2 ug/L for Sb, Ba, Be, Cd, Co, Cr, Pb, Mn,
   Mo, Ni, Ag, Tl; and 0.5 ug/L for Al and Zn;
*  ICP-AES limits of 0.5 ug/L for Sr, 0.01 mg/L for Mg, and
   0.2 mg/L for Na; and
*  Nutrient limits of 0.01 mg/L for NO2, 0.02 mg/L for NH4, and
   0.05 mg/L for NO3+NO2.

Because all analytical results for each of the above are below
reporting limits, these data are not included in table 1.

In reviewing the results in table 1, the reader must note that the
respective aliquots in the DIW-washed series (CF1-1 to CF5-4)
represent dissimilar total volumes of fluid flushed through the
filters from the respective aliquots in the acid washed series
(ACF1-1 to ACF5-4).  This stems from the extra volume represented
by the acid wash step itself plus the volume represented by the
subsequent DIW wash required to return pH to normal.  Thus, where
different concentrations of a given trace element occur for the
same aliquot number in the two sets, the difference could result
from the acid washing and/or the larger volume of fluid
represented in the aliquots in the acid wash series.

The two atmospheric blanks show no detectable concentrations of
trace elements or nutrients except for a very small level of total
phosphorus (Blank 2).  Blanks from the DIW reservoir are similarly
clean except for low levels of boron (DI-4) and total phosphorus
(DI-1 and DI-4).  The pump blanks have low levels of boron (PB-6),
total phosphorus (PB-1), and total nitrogen (PB-1, PB-3, and PB-
6).  Further, all of the pump blanks (as well as all capsule
filter blanks) contain detectable but low levels of Si.  The
presence of Si can probably be ascribed to the use of silicon pump
tubing, and occurred despite the use of new tubing that had been
heavily soaked and rinsed with 0.6 molar hydrochloric acid prior
to use. This conclusion is based on the facts that: (1) no
detectable Si occurred in either the atmospheric or the reservoir
blanks, and (2) the Si levels in the capsule filter blanks are
consistent with the levels in the pump blanks.

As cited above, the pump and/or reservoir blanks contain low
levels of total nitrogen and total phosphorus.  The concentrations
are at or very near the reporting limits for these parameters.
There is no detectable total nitrogen (>0.05 mg/L) in any of the
capsule filtrates, whereas there is an occasional detectable total
phosphorus concentration at the reporting limit (0.01 mg/L).  We
believe the sporadic low-level total nitrogen and total phosphorus
concentrations result from analytical noise and do not represent
contamination.

At the detection limits used in this study, there is little or no
contamination of trace elements and nutrients associated with the
tested capsule filters.  Further, there is little or no difference
between the acid precleaned and non-acid precleaned filters.  The
only exceptions are Ca and Cu, both of which consistently show up
in the initial 500-mL rinse for the non-acid precleaned filters
(Cu shows up in the first 500-mL rinse for the first two acid pre-
cleaned filters (ACF1-1 and ACF2-1) as well).  Where detectable
concentrations for both ele-ments occur in the first 500-mL rinse,
they drop below their respective reporting limits in the second
500-mL rinse.

                             Conclusions

Based on these results, it appears that Gelman Supor capsule
filters (Catalogue #12175) prerinsed/conditioned with 1,000 mL of
DIW are suitable for ppb-level determinations of trace elements,
and appear noncontaminating for nutrients at present reporting
limits.  At the ppb-level for trace elements and at present
nutrient reporting limits, there appears to be no benefit to acid
prerinsing.  The use of silicon pump tubing, in conjunction with
the tested capsule filters, appears to contribute minor amounts
of Si contamination.  However, the detected levels are probably
insignificant relative to most environmental samples.  If users
desire to reduce/eliminate the Si contamination associated with
the use of silicon pump tubing, Teflon tubing can be substituted,
except for a very small length required for the peristaltic pump
head.

                            References

Shiller, A.M., and Boyle, E.A., 1987, Variability of dissolved
trace metals in the Mississippi River:  Geochimica et
Cosmochimica Acta, v. 51, p. 3273-3277.

Windom, H.L., Byrd, J.T., Smith, R.G., Jr., and Huan, F., 1991,
Inadequacy of NASQAN data for assessing metal trends in the
nation's rivers:  Environmental Science and Technology, v. 25,
no. 6, p. 1137-1142.




                                David A. Rickert
                                Chief, Office of Water Quality

Attachment:  Table 1

This memorandum refers to Office of Water Quality Technical
Memorandums 80.22, 92.12, and 92.13.

Key words:  Capsule filters, contamination, nutrients,
            trace elements

Distribution:  A, B, S, FO, PO

                 [Table 1 available in hard copy only]