QUALITY OF WATER BRANCH TECHNICAL MEMORANDUM NO. 85.07

                                         January 23, 1985 



QUALITY OF WATER BRANCH TECHNICAL MEMORANDUM 85.07

Subject: WATER QUALITY--Preservation of Nutrient Samples for 
            Nutrient Analysis; An Update on Two Issues

Since Quality of Water Branch Technical Memorandum 80,26 was 
issued directing the use of HgCl2/NaCl tablets plus chilling 
at 4 degrees C for preservation of water samples for nutrient 
determination, there has been periodic debate over the 
effectiveness of this treatment relative to others.  Recently 
also, evidence has surfaced that the preservative tablets may 
be randomly contaminated with ammonia (NH3-N), either as a 
trace impurity in the original reagents or by sorption of 
gaseous NH3 from the laboratory or field-office atmosphere.  
The purposes of this memorandum are to discuss results of 
studies recently completed with regard to these matters, 
advise the field and project offices on steps to take at once 
to reduce the chance of NH3 contamination, and describe 
additional studies planned or underway.

The attached memorandum from Marvin Fishman and LeRoy 
Schroder dated October 18, 1984, describes the results of 
tests conducted in the Denver Central Laboratory for 
effectiveness of five popular preservation treatments over a
l6-day period.  The findings and implications are clear; the 
combination of the HgCL2/NaCl tablet with chilling at 4 
degrees C is the best of the treatments tested.  We intend to 
retain this treatment as the standard practice within Water 
Resources Division (WRD) until a better one is demonstrated.

The second memorandum from Marvin Fishman, also attached, 
describes studies conducted in both Central Laboratories to 
determine the levels of NH3-N contamination in preservative 
tablets in storage at those facilities.  Analyses were 
performed by automated wet-chemical techniques in both labs.
Atlanta finds no significant contamination down to the 
detection limit (0.002 mg/L) whereas Denver finds 
contamination around a level of 0.01 mg/L.  There is also an 
indication in the Denver lab that NH3-N might be sorbed
from the atmosphere into samples standing open awaiting 
analysis (witness the slight but steady increase in results 
of replicate determinations).  These results will be checked 
shortly by exchanging tablets between laboratories for 
reanalysis.  We will alert you of the results when these
studies are complete.

A major unanswered question is the degree to which the 
preservative tablets have become contaminated during storage 
in field and project offices.  We will get a fast assessment 
of this question by a random sampling of preservative tablets 
in at least four District Offices.  Marvin Fishman has been 
asked to send bottles of ammonia-free water to Districts 
selected at random.  The Districts will add a tablet selected 
at random to all but one bottle, leaving that one as a travel 
blank, and return the samples for analysis at no charge to 
the District.  Appropriate action will be taken based on an 
evaluation of the data.

Despite the uncertainties regarding degree of NH3-N 
contamination in the tablets, it seems prudent to take some 
steps at once to reduce the chances of contamination and to 
prevent erroneous analyses of water samples.  Field and 
project offices are advised to take the following steps 
immediately:

1.  Assess the possibility that NH3 gases are or have been 
present in the area where the tablets are stored--is the area 
cleaned periodically with ammonia-containing solutions, has 
concentrated ammonia been left open to the atmosphere outside 
a hood, do personnel smoke in the area, is the area affected 
by automotive exhaust?  If exposure is suspected, discard all
remaining tablets and reorder from the Central Laboratory.

2.  Isolate all remaining tablets from the ambient 
atmosphere.  Protection from all ammonia sources is 
essential!  Tablets must be kept in a sealed container.  The 
ammunition can recommended for storing metallic mercury is
satisfactory.

3.  Until further notice, the preservative tablets should be 
used only for samples in which NH3-N determination is not 
requested or where expected concentrations are above 
0.02 mg/L and the regular analysis is requested (LC 0301: 
detection limit 0.01 mg/L).  If pristine-area, precipitation, 
or otherwise low ionic-strength samples are being analyzed 
and LC 0830 (low-level colorimetric) is requested, special 
arrangements should be made with the receiving laboratory to 
obtain priority handling (at no additional charge), and 
samples should be shipped chilled without preservatives by an
overnight express carrier.  Please be discrete; none of us 
want the labs to be overloaded with priority requests.  As a 
reminder, the preservative tablets cannot be used when LC 
0520 (Ion chromotographic anion determination) is requested.

4.  Each District is encouraged to conduct its own tests for 
NH3-N contamination at cost to the District.  Ammonia-free 
water may be prepared by adjusting deionized or distilled 
water to pH 8 with NaOH or KOH and boiling for about 
15 minutes.  Submit at least one blank and three samples
with tablets (selected at random) added.  Alert the 
laboratory that the bottles are for test purposes and 
indicate which are blanks and which are tests because the 
laboratory standards need to be changed if HgCl2 is absent.  
Please send a copy of the data to the Quality of Water 
Branch.  Stocks of tablets at the laboratories are sufficient 
to carry us through fiscal year 1985.  The process of 
reordering will be accelerated and a different packaging 
arrangement will be sought.




                        David A. Rickert
                        Acting Chief, Quality of Water Branch



A, B, FO, PO

This memorandum temporarily modifies instructions given in 
Quality of Water Branch Technical Memorandum 80.26.

Key words: Water quality, sampling, preservation


====================================================================

1 Attachment:

Memorandum dated Nov. 5, 1984 with 3 enclosures (dated Oct. 24, 1984,
Oct. 25, 1984, and Oct. 5, 1984).  Tables 1 to 5 of the last enclosure
are presented separately as graphical documents.

====================================================================


									MEMORANDUM

DATE:			November 5,1984

IN REPLY REFER TO:	Marvin J. Fishman, WRD, National Water Quality Laboratory, Arvada, CO
  ATTN OF:

SUBJECT: 		WATER ANALYSIS--Possible contamination of the HgC12/NaCl pills by ammonia
			and correspondence

TO:			Analytical Services Coordinator, WRD, Reston, VA


	Enclosed are data and correspondence from both laboratories for the mercuric
chloride contamination study. They are self-explanatory. For the majority of
samples there is no problem. The pills are not contaminated with ammonia. Based
on the Denver data, there could be problems when determining ammonia on pristene
or precipitation samples.

	You can also see from the Denver data (Samples 1-25) that the ammonia
concentration tends to creep upward from the first to the second, and final runs.
This again shows the need for a separate laboratory for pristene and precipitation
samples.

Please call if you have any questions.


						Marvin J. Fishman


Enclosures (3)

cc w/encl:	Chief, Quality of Water Branch
		B. Malo
		Chief, Central Laboratory, Denver
		Chief, Central Laboratory, Doraville
		Vance Kennedy





									MEMORANDUM

 
DATE:		October 24, 1984

REPLY TO
  ATTN OF:	Merle W. Shockey, Chemist, Central Laboratory, Arvada, CO

SUBJECT:
		HgC12 Tablet Contamination Study

TO:		Marvin J. Fishman 
thru: 		Chief, Central Laboratory, Arvada, CO 	
thru: 		Chief, Inorganic Chemistry, Central Laboratory, Arvada, CO 


Attached is a tabulation of the results for the HgCl2 tablet ammonium ion
contamination study you requested. The results indicate possible
contamination at the .003 to .015 mg/L levels. This is below the normal
detection limit of our regular method. It will, however, pose problems on
our low level method which attempts to reach .003 mg/L. All data is
presented in replicate for precision purposes.

		



					Merle W. Shockey		
					Chemist		


Attachment		

MWS/rw





 Table 1.--Deionized water with reagent grade HgC12 (52 ¥g/L) and HaCl (450
mg/L) added.

		1st Analysis	2nd Analysis	3rd Analysis	4th Analysis
A		-0.003		-0.004		-0.003		0.002
B		-0.004		-0.004		-0.004		0.002
C		-0.004		-0.003		-0.004		0.002
D		-0.004		-0.003		-0.004		0.000
E		-0.001		-0.001		-0.001		0.002





Table II.--DeionRzed water with one randomly selected HgC12 tablet (from
cards) added per 250 L bottle.

		1st Analysis	2nd Analysis	3rd Analysis

1		0.003		0.006		0.011
2		0.011		0.009		0.015
3		0.011		0.015		0.017
4		0.006		0.006		0.009
5		0.003		0.005		0.006
6		0.011		0.011		0.014
7		0.006		0.006		0.008
8		0.005		0.008		0.011
9		0.003		0.008		0.006
10		0.005		0.006		0.012
11		0.006		0.009		0.012
12		0.006		0.008		0.011
13		0.006 		0.012		0.011
14		0.009		0.011		0.014
15		0.008		0.009		0.011
16		0.003		0.005		0.009
17		0.006		0.008		0.012
18		0.011		0.009		0.012
19		0.014		0.015		0.018
20		0.009		0.011		0.011
21		0.009		0.011		0.011
22		0.011		0.011		0.015
23		0.005		0.008		0.009
24		0.014		0.012		0.017
25		0.011		0.015		0.015





Table III.--Deionized water with one randomly selected tablet taken from
polyethylene bottle.


		1st Analysis

26			0.014
27			0.011
28			0.012
29			0.005
30			0.012





								
									MEMORANDUM
DATE;		October 25, 1984 			
REPLY TO
  ATTN OF:	Randy Brown, Chief, Section II
 THRU:		Laboratory Chief, NWQL-Atlanta, Doraville, GA

SUBJECT: 	WATER ANALYSIS - Reply to Memo of Oct 5~ 1984

TO:		M. J. Fishman, NRD, MS 407, Denver,
		NWQL-Denver, CO

Type I water was used to make up all samples. New bottles of both NaCl
and HgC12 were brought to make up the solutions. The pills were taken at
random from the manufactured cards. The solutions were placed in
250-mL white plastic bottles.

	No.		Type of Solution

	25	Type I water + one HgC12/NaCl pill
	 5	Type I water
	 5	Type I water + 13 mg HgC12/172 mg NaCl
	 1	Type I water + 130 mg HgC12/ 1.72 g NaCl
	 1	Type I water + 2 HgClz/NaCl pills
	 1	Type I water + 4 HgC12/NaCl pills
	 1	Type I water + 8 HgC12/NaCl pills
	 1	Type I water + 10 HgC12/NaCl pills

	40 Total





All of the 40 solutions were then analyzed on the IQAS for NH3-N.

	Type of Solution		No.	1st		2nd
	Up
	Type I water + one		25	<.001		<.001
	HgC12/NaCl pill			
						<.001		<.001

						<.001		<.001

						<.001		<.001
					
						<.001		 .004

						<.001		<.001
						
						 .006		 .007
		
						<.001		<.001

						<.001		(.001

						<.001		<.001

						<.001		<.001
					
						(.001		<.001

						(.001		<.001
	
						<.001		<.001

						<.001		<.001

						<.001		<.001

						<.001		<.001

						<.001		 .004
						
						<.001		<.001

						<.001		<.001

						<.001		<.001

						<.001		<.001

						<.001		<.001

						<.001		<.001

						 .017		 .016

						<.001		<.001


	Type I water + .013 mg
					 5	<.001		<.001
	HgC12/.172 mg NaCl
						<.001		<.001
						
						<.001		<.001
		
						<.001		<.001
			
						<.001		<.001
		_
	Type I water + .130 g HgC12/
	1.72g NaCl			1	<.001		<.001

			
	Type I water			5	<.001		<.001
		
						<.001		<.001

						<.001		<.001

						<.001		<.001

						<.001		<.001


	Type I water + 2 HgC12 Pills	1	<.001		<.001

	 "	"      4   "	 "	1	.016		<.001

	 " 	"      8   "	 "	1	.008		 .025

	 "  	"     10    "    "	1	<.001		<.001 





	Quality Control

		     CURVE				EPA #1
	 Calc.	  1st	 2nd 			Calc.	1st	2nd

	 .000	 .001	.000 			.19	.201	.194

	 .125	 .135	.125

	 .250	 .247	.247

	 .500	 .501	.506

	1.000	1.002	.996


After the data was analyzed in the Atlanta Laboratory, it was concluded
that there seems to be no problem of ammonia contamination in the
mercuric chloride pellets.




						Randy Brown
						Supervisory Chemist

MRB:dds





 		
									MEMORANDU



DATE:		October 5, 1984

REPLY TO
  ATTN OF:	Marvin J. Fishman, WRD, National Water Quality Laboratory, Arvada, CO

SUBJECT:	WATER ANALYSTS - Possible contamination of the HgC12/NaCl pills by ammonia

TO:		Central Laboratory Chief, Doraville, GA
		Central Laboratory Chief, Denver, CO

			At the request of Herman Feltz and our discussions, please perform the
		following experiment as soon as possible.

		1. Withdraw into a large enough clean container sufficient ammonia-free
		   water to complete the tasks described below.

		2. Wash plastic bottles, and all other glassware thoroughly before preparing
		   standards and samples.

		3. Prepare a set of low level NH3-N standards using either reagent grade or
		   ultrex NaCl and HgC12. Use a new unopened bottle of each.

		4. Prepare 5 water blanks in 250-mL plastic bottles using the same NaCI and
		   HgC12 as in step 3 above.

		S. Prepare 2S water blanks in 2S0-mL plastic bottles and add a HgC12/NaCl
		   pill to each. Randomly select these pills from the remaining stock.

		6. Determine NH3-N on~ thSe above prepared solutions in triplicate. The
		   analysis of these samples should be made immediately after preparation.

		7. Send the results to Marvin J. Fishman.

		If these instructions are not clear, please contact me.





							Marvin J. Fishman c
		
		cc: Analytical Services Coordinator





			

				
									MEMORANDUM
	

DATE:		October 18, 1984

RELY TO
  ATTN OF:	Marvin J. Fishman and Leroy Schroder, WRD, National Water Quality Laboratory
	
SUBJECT:	WATER ANALYSIS - Preservation of Nutrient Samples

TO:		Chief, Quality of Water Branch, WRD, MS 412, Reston, VA

			In 1980 the Quality of Water Branch issued a Technical Memorandum (No.	
	80.26) stating that on October 1, 1980, water samples submitted to the National
	Water Quality Laboratories for determination of nitrogen and phosphorus compounds
	must contain 40 mg/L of mercuric ion and must be chilled to 4¡C. This directive
	was based on a briefing paper prepared by Turk and Bradford. The authors made an
	exhaustive search of the literature on techniques being used to preserve samples for
	nutrient analysis. Their studies indicated that chilling alone, the only preservation
	technique used up to 1980 by the USGS was not reliable. These studies also stated
	that there are no completely reliable techniques for preserving water samples for 3
	to 5 days. The conclusion of their review was that chilling plus a biocide,
	specifically mercuric ion, appears to be the most effective means to preserve
	nutrient samples.
	
		There have been many investigators both in USGS and other organizations who
	have felt that mercuric ion is not needed and that chilling will suffice. A study
	conducted by Fishman in 1973 showed that chilling alone was sufficient for
	preserving samples up to a period of nine days. This study was done on a very
	limited number of sample types. Also, it was not determined if these samples were
	biologically asctive.
	
		Recently Vance Kennedy in a memo to the Analytical Services Coordinator
	stated that the mercuric chloride pills may be contaminated with ammonium ion.
	He also stated that one can possibly get away with contaminated pills on high
	nutrient samples, but certainly not on pristene samples or precipitation samples.

		Because of the doubts by many investigators on needed use of mercuric ion and
	possible contamination of the present stock of pills, Leroy Schroder and myself felt
	that another study should be undertaken to determine, if mercuric ion was needed to
	preserve samples. No additional preservation techniques were reported in the
	literature within the past few years. Most laboratories are following EPA protocol.
	For ammonia-N, organic-ON, nitrite plus nitrate-N, hydrolyzable and total P. EPA
	recommends chilling to 4 C and adding sulfuric acid to pH less th3n 2. For nitrate
	or nitrite alone and orthophosphate, they recommend chilling to 4 cr. The holding
	time recommended is 24 hours.
	
		With the cooperation of the Denver Central laboratory a study was set up.
	The preservation treatments agreed upon were 1) chilling, 2) mercuric chloride plus
	chilling, 3) sulfuric acid plus chilling, 4) chloroform at room temperature, and 5)
	chloroform plus chilling. Chloroform was chosen as a possible preservative because
	some investigators have had success with it (T(leene and Galloway).

		The turn-around time for nutrient sample analysis is specified as 3 to 5 days from
	time of collection. This time frame is probably unrealistic and it was decided to run this
	study for 16 days even though 16 days is too long to delay analysis. Verbal discussion with
	laboratory personnel indicates 10 days from collection to analysis is occurring.

	The study was set up as follows:

	Samples collected

		1.	Platte River near 88th Street in Commerce City, CO
		2.	Clear Creek in Prospect Park in Arvada, CO
		3.	Arvada Lake in Prospect Park in Arvada, CO
		4. 	Tap water NWQL, Arvada, CO
		5.	Deionized water NWQL, Arvada, CO

		Samples 1, 2, and 3 were collected in 5-gallon plastic bottles on September 9, 1984.
	Samples 4 and 5 were collected in 5-gallon bottles on September 10, 1984.

	Sample handling

		1.	Filtered sample Nos. l, 2, and 3 (natural samples) through 10 M-nominal
			pore-size filter on September 9, 1984.

		2.	Transferred 250 mL of each sample (Nos. 1, 2, and 3) to plastic amber bottles
			and added a mercuric chloride tablet to each and placed them along with the
			5-gallon filtered samples in a refrigerator (4¡C).

		3.	Sample Nos. l, 2, and 3 (250 mL samples) were analyzed for nutrients on
			September 10, 1984.

		4.	Based on the above analyses on September 10, 1984, Sample No. 3 as well as
			Sample Nos. 4 and 5 were spiked with nitrogen and phosphorus compounds.
			The amounts added are shown in footnotes of tables 3, 4, and 5.

		5.	Each of the five samples were then subdivided into sixty-two 250 mL-amber
			bottles.

		6.	Twelve bottles of each sample were labeled "chilled" and placed in a
			refrigerator.

		7.	A mercuric chloride tablet was added to 12 bottles of each sample, labeled
			"HgC12 plus chilled", and placed in a refrigerator.

		8.	Concentrated sulfuric acid (1.0 mL) was added to 12 bottles of each sample,
			labeled "H2SO4 plus chilled", and placed in a refrigerator.

		9.	Chloroform (0.5 mL) was added to 24 bottles of each sample. Twelve of the
			bottles were labeled, "chloroform room temperature", and stored at ambient
			temperature. The other 12 bottles were labeled "chloroform plus chilled", and
			placed in a refrigerator.

		10.	Two bottles of each sample were labeled, "room temperature", and stored at
			ambient temperature.

		11.	A complete set of samples in triplicate were submitted to the Denver Central
			laboratory on the following days:
			
			September 11 (Day 1)
			September 14 (Day 4)
			September 18 (Day 8)
			September 26 (Day 16)

		12.	The samples stored at ambient temperature containing no preservative were
			submitted for analysis on September 26 (day 16).

	Analytes determined

	1. 	Organic nitrogen plus ammonia nitrogen
	2.	Ammonia nitrogen
	3.	Nitrite plus nitrate nitrogen
	4.	Nitrite nitrogen
	5.	Orthosphosphate as phosphorus
	6.	Total phosphorus


Results


	Tables 1-5 show the data obtained from the preservation study. As previously
mentioned samples were analyzed on days 1, 4, 8, and 16 after collection. The data for
the Platte River, Clear Creek, and Arvada Lake samples reported for day 0 were actually
obtained one day after collection; however, these samples were filtered and a portion
fixed with mercuric chloride and stored at 4¡C the day of collection. All values, except
for the unpreserved samples, are means of triplicate analyses. The standard deviations
calculated for each set of replicates are within the precision of the methods except for
chloroform at ambient temperature. The results obtained for each preservation technique
is discussed below.

Table 1
Table 2
Table 3
Table 4
Table 5

	Mercuric chloride (4¡C) - The data supports the present preservation technique.
There are no significant differences in data for any of the samples from day 0 through day
16 for any of the six constituents, although orthophosphate has increased slightly on day
16 in the Arvada Lake, Arvada tap water and deionized water samples.

	Chilled (4¡C) - It appears from the data that the six constituents in the five samples
were stable for at least eight days. The data for the first eight days support the 9-day
period data by Fishman in 1973. At what point changes occurred prior to day 16 is not
known. The changes in concentration, which were obtained on day 16, were not the same
in each sample. For the Platte River sample, the ammonia concentration increased
significantly and nitrite disappeared from the sample. The nitrite concentration in the
Clear Creek sample was beginning to decrease. In the Arvada Lake sample the ammonia
and orthophosphate concentrations increased significantly. No changes occurred in the
tap water sample. This may be due to the presence of chlorine which may cause the
sample not to be biologically active. The orthosphosphate in the deionized samples
started to increase.

	Sulfuric acid (4¡C) - The ammonia concentration in the Platte River and Clear
Creek samples are higher than any of the other preservation techniques. This can be seen
from day 1. It appears that orthophosphate is increasing on day 8 and there are definite
increases by day 16 for the Arvada Lake, Arvada tap water and deionized water samples.
Also our present methodology for NO2 and NO3 has an interference for highly acidic
samples.


	Chloroform (ambient and 4¡C) - The individual data points for ammonia, nitrite plus
nitrate, nitrite, and orthophosphate were erratic for the Platte River, Clear Creek and
Arvada lake samples. Data for the chilled samples looked good for eight days. On day 16
ammonia decreased significantly in the Platte River sample and there appears to be a
slight increase in nitrite which is probably not significant. Ammonia levels also decreased
in the Arvada Lake sample as did nitrite.

	Unpreserved (ambient temp.) - Data was collected only on day 16. For the Platte
River, Clear Creek, and Arvada Lake samples, the data obtained for organic nitrogen plus
ammonia, ammonia, nitrite plus nitrate, and nitrite are not similar to any of the data
obtained by the preservation techniques. In the deionized water sample the
orthophosphate increased significantly. From these results, we can confidently state that
the Platte River, Clear Creek, and Arvada Lake samples were biologically active and
preservaton of samples is necessary.

Conclusions

	This preservation study has demonstrated that preservation is required and that
mercuric chloride is an excellent preservative, and should be continued for the bulk of
samples submitted to the Central Laboratories. However, as stated in QW Technical
Memorandum No. 80-26, samples without mercuric chloride can be handled by the
laboratories if special arrangements are made. If samples are to be handled using chilling
alone, the samples need to be analyzed as quickly as possible after collection.

	If you have any questions concerning the data, etc., please contact either Leroy
Schroder or myself. We are also planning to publish this data in the "white literature" in
the near future.






					Marvin J. Fishman	Leroy SCHroder


cc:	Analytical Service Coordinator
	Bernie Malo
	Linda Friedman
	Regional Research Hydrologist (CR)
	Vance Kennedy