DIRECT DETERMINATION OF CARBON, HYDROGEN AND NITROGEN FOR ENVIRONMENTAL AND INDUSTRIAL FILTER APPLICATIONS
The technology for easy, direct automated elemental analysis of material retained on certain
membranes used in water filtration and air monitoring applications is now available on the Exeter
Analytical Model CE440 Elemental Analyzer. Greatest efficiency is achieved using the CE440
equipped with a multi-sample injector, but the technique is also applicable to instruments with single
sample automation. The method requires a minimum of sample preparation and only minor alterations
in standard instrument operating procedures.
Direct determination of total CHN retained on a filter is more sensitive, precise and cost-effective than
making measurements on pre- and post- filtered samples and calculating the difference. This has been
confirmed in a study conducted for the EPA on particulate concentrations in water samples from the
Chesapeake Bay Monitoring Program (1). Direct determination is preferred over the traditional “by
difference” method because of the formers reduced sources of error, rapidity of analysis and increased
sample representativeness due to the greater volume filtered.
The advantages of the Exeter Analytical CE440 versus other elemental analyzers lie primarily in its
horizontal combustion train which prevents clogging from accumulation of filter and sample ash and
its capacity to process filters up to 47 mm in diameter in one analytical cycle. A 7×5 mm nickel sleeves
(P/N 6703-0499) is used to contain a filter up to 25 mm or half of a larger filter. For the latter, two
sleeves are inserted simultaneously to produce one complete analysis.
This technique applies equally to glass or quartz microfiber filters, the choice depending on the
particular application. However, to maintain a low, consistent blank value, only the highest quality
filter without binder should be used (Whatman 934-AH or GF series of glass fiber filters, Ultra-Pure
QM-A quartz filters, or equivalent).
Low cost glass fiber filters of various pore sizes and dimensions are the most common type for
collecting particulates in water or air. Smaller filters are preferred when excessive suspended
particulates are absent; the advantage is a reduced blank signal and the entire filter can fit into one
sleeve. Although glass fiber filters can withstand temperatures up to 550oC, to minimize crumbling
and shrinkage it is best to precombust them at a maximum of 450-500oC for up to 16 hours. Nickel
sleeves are required during analysis to contain the melt and to prevent spattering during the
combustion. The static combustion feature combined with 980oC temperature of the CE440 ensures
that no material is entrapped in the melt.
Quartz fiber filters are used in specialized applications, such as flue stack monitoring, because they
withstand temperatures up to 1000oC and have increased chemical resistance. For best results, quartz
filters should be muffled overnight at 500oC prior to use. They are easy to handle and will remain
intact during analysis.
Silver membrane filters, which are completely solvent resistant, should also be compatible with this
elemental analysis technique.
A. Sampling Procedure
A protocol appropriate for the particular water filtration or air monitoring application should be
selected, taking particular care to ensure sample homogeneity and representativeness. Sampling is the
single largest determinant of data quality, as the magnitude of variability from handling and analysis is
comparatively small. Duplicate, or even triplicate sampling is recommended. Filter blanks should be
treated the same as filter samples in all respects except for actual sample collection.
For extra large filters, such as those used in high volume air monitoring, the filter paper should be
folded with exposed surface inside and a standard diameter portion excised. The resulting two aliquots
should be dried prior to loading for analysis.
A typical method for filtering aqueous solutions is as follows:
1. Place the correct size pre combusted filter pad in a vacuum filtration assembly.
2. Agitate a known volume of water (usually 100 ml) vigorously and quickly pour the sample into
the filtration assembly. Another 100 ml of de ionized water can be used to wash through the
system. If a wash is used, do the same for the filter blank.
3. Filter at 15 in Hg vacuum to dryness and break the seal.
4. The filter can be optionally vacuum desiccated overnight, folded in half (exposed surface
inside), wrapped in aluminum foil or placed individually in a covered petrie dish, labeled and
frozen at -10oC for later analysis.
5. Prior to actual analysis, non-desiccated samples or those that have collected condensation
should be placed in a drying oven at 45oC overnight.
B. Filter Preparation for Analysis
Depending on filter size, some variation of the following general approach using the Exeter
Analytical HA Filter Kit (P/N 125-00020) is recommended for packing the filter into a
7x5mm nickel sleeve:
1. Work on a non-contaminating surface (e.g. tape aluminum foil that has been muffled at 500oC
for one hour onto a work bench).
2. Using two pair of clean forceps (not included in kit), fold the 25 mm filter in half so that the
exposed surface is inside. Cut 30-47 mm diameter filters in half, fold each half
separately and proceed.
3. Continue folding the filter down from both dimensions until you have a compact package.
Compress further with the forceps.
4. Place a 7×5 mm nickel sleeve into the filter loading die, which functions as a holding device.
5. Use the 4 mm loading plunger to force the compressed filter through the Pyrex loading funnel
and into the nickel sleeve. For all glass fiber filters and other filters that have heavily loaded
with material, we recommend that a small square of muffled aluminum foil be loaded through
the funnel into the nickel sleeve before the filter is added. This helps contain possible spatter
and loose material.
6. Make sure no excess filter protrudes above the lip of the sleeve.
7. For larger filters, repeat this procedure for each half. The two halves, in separate sleeves,
compose one sample for analysis.
Once the filters are packed into the nickel sleeves they can be transferred into the 64 sample wheel.
The calibration series must be placed at the beginning of the wheel. This normally consists of a
conditioner, a blank, a conditioner and three standards. Normally, OAS grade acetanilide is used to
calibrate the instrument. For low level samples you may choose to use a standard that more closely
resembles the samples or use a smaller amount (as low as 500.0 ug) of acetanilide. Another option is to
use the ‘Expanded Read’ and ‘Expanded Reporting’ parameter choices. A standard should also be
placed at the end of the wheel and a filter blank should follow each series of ten samples.
By entering 100 for the weight of the filter samples, unitless results are printed out which represent
micrograms of C, H and N per filter. This corresponds directly to the known amount of liquid or air
that has passed through the filter. The maximum sample capacity per run is approximately 4,000 to
5,000 micrograms of carbon. Filters containing more than that amount (i.e., they appear very dark) can
be cut in half and analyzed separately and the results added. For particulate samples collected from
water, the hydrogen result is meaningless.
If you are running large filters that require double drop operation, you must put a double drop blank
(i.e. – two sleeves) immediately after the standard and use the “Automatic Ks and Blanks” option in
order for the correct blank value to be used. For samples or blanks that require the double drop option
to occur, you must preface the sample ID with a plus sign (+). Refer to section 4.A in the CE440
manual for more detail.)
Filter blank values should be only slightly higher than the normal instrument blank level; a 0 to 5 uV
increase in Nitrogen and up to 100 uV increase in Carbon are typical. Large filters will show more
variability in their blank values. Greater increases or high variability indicate poor quality filter pads or
a problem in sample/blank preparation technique.
Precision of sample duplicates is generally lower than typical organic standards due primarily to the
variability in sampling. However, greater than 15 to 20% variation may indicate an error beyond a
sampling problem. Samples containing different particle sizes should be run in triplicate. One
recommended procedure is to spike some filters with a known quantity of material and look for good
The technique described above is limited to total C, H and N. A number of alternatives to analyze for
organic, inorganic or free carbon have been proposed, but there is no standard method available to
(1) D’Elia, C.F. et al “Nitrogen and Phosphorus Determinations in Estuarine Waters: A
Comparison of Methods Used in Chesapeake Bay Monitoring”, Report prepared for
Chesapeake Bay Program Liaison Office, Region III, US EPA, Annapolis, MD. 1987.