|
|
Fluorometric Ammonium Analysis
Modified from Holmes et al, and
various memos (30 Jun 99, 14
February 2000)
Introduction
Fluorescence is produced by
the reaction of OPA with ammonium.
Fluorometry is sensitive and simple so seems to be a good way to measure
ammonium, particularly at low
levels.
Details of methods, reagents, etc are given in Holmes et al. 1999 (CJFAS 56:1801-1808). This document supplements the manuscript
and is intended to give a quick,
user-friendly overview of the
procedure. It also details a variation of the
method not discussed in the
manuscript, which uses 10 mL sample
and 10 mL working reagent (we are
calling this variant Protocol B-1).
General Comments
Safety
Gloves should be worn when dealing with OPA. At the concentrations we use, it won’t immediately kill you, but it’s not good for you either. According to Andy Mattox (the safety officer at MBL), all chemicals used in this analysis can go down the drain. However, at Toolik we are treating them as hazardous waste.
Contamination & Problems
Always use ammonia-free water for
all reagents.
Get the water from the Nanopure unit immediately prior to use.
Make absolutely sure the DI
used to make the standards is also
used for the blank.
Never acid wash these bottles. Instead, just pre-reacted prior to first use, and rinse between uses.
Accuracy
When making the standard dilution
series, make sure the pipettes are
calibrated properly.
It is important to get very accurate
dilutions when working w/ such low
concentrations. Check calibration by weighing the
desired amount on the balance and
adjust pipet accordingly.
Apparatus & supplies
Wheaton scintillation vials (Wheaton 986704, Fisher
Catalog # 03-341-72C):
4 liter amber bottles (fisher # 028846b) 6/case
Reagents
The following is enough for about 48
liters of WR and 4 liters of Borate
Buffer.
Preparation of Stock reagents
BORATE BUFFER (BB): Borate buffer without the
sodium sulfite or the OPA is used to
evaluate the sample background
fluorescence (BF).
sodium sulfite: Next prepare the sodium sulfite solution (2 g sodium sulfite to 250 mL
DI water)
orthopthadialdehyde (OPA):
add 8 g OPA to 200 mL ethanol
(keep this solution as dark as
possible), shake vigorously until
OPA dissolves.
WORKING REAGENT (WR):
Working
reagent appears to be stable for
months, and its blank fluorescence
decreases over time, so it is best
to make WR in large batches and let
it age. We make WR batches of about 4 L in 1
gallon brown Nalgene bottles (these
bottles actually hold about 4.4 L).
To a clean 4-liter bottle (pre-react, or just rinse with DI if previously used for WR), add approximately 3 L DI. Then add 160 g sodium borate, cap, and shake vigorously until your arms are tired, then rest, then do it some more, and add 20 ml of sodium sulfite solution to the 1 gallon jug with DI and sodium borate already added. Shake the jug some more. Finally, add 200 ml of OPA solution to the 4-liter jug. Shake some more, then add DI until the bottle is nearly full – about 1 inch from the top. Shake a bit more, let age for at least a few days if possible, and then the WR is ready to use.
Sample bottle preparation
To pre-react, add 10 mL working
reagent (WR) to bottle, cap and
shake, and let sit for at least 3
hours (days or weeks is fine).
Dump WR, then rinse three times with pure DI water. Next add 10 mL WR, shake, dump, and then
load with another 10 mL WR.
Store in dark (WR is
light-sensitive).
Once loaded with WR, the bottles are
ready to go (keep WR in dark at all
times).
Sample collection
1. Rinse syringe 3 x with water to be sampled.
2. Attach filter to end of syringe (either a GF/F or .2 um
membrane filter) and filter 10-30 ml
though filter to rinse filter.
3. Dump DI from sample bottle and rinse with 30-40 ml of
filtrate.
4. Re-fill syringe and fill sample bottle with filtrate.
Blanks
Duplicates should be run for all blanks, with reagent blanks run at the beginning and end of each sample set. Three types of blanks should be run with each set of samples. 1. Trip blanks (DI blanks that are carried out to the field and
back).
These would consist of sample
bottles filled with DI which brought
into the field and then returned.
2.
Field blanks (DI blanks filtered
in the field as if they were
samples).
BACKGROUND FLUORESCENCE (BF):
All samples auto-fluoresce to some degree. This BF must be subtracted from the
observed sample fluorescence in
order to quantify ammonium
concentration. If it is found that BF doesn’t change
though the water column or down a
stream transect it may be possible
to take fewer BF measurements.
If it does change however you will
need to take a BF each time a sample
is taken. In surface waters around Toolik Lake,
ammonium concentrations tend to be
very low and background fluorescence
is relatively significant.
Therefore, it is important to
accurately quantify BF.
In our limited experience so far, BF is relatively constant in a given
water-body on a given day (for
example, Toolik Main station or
Kuparuk River transect), but BF
varies across sites (and maybe
temporally). Therefore, BF does not need to be
sampled at every station within a
given “site”, but must be sampled at
each stream. Another example: On June 23, 1999, BF was essentially constant at 11 Kuparuk
River stations, but differed
significantly in Hershey Creek, a
small tributary to the Kuparuk
River.
If BF had not been measured in
Hershey Creek and instead the
Kuparuk BF was used, the Hershey
Creek ammonium result would have
been erroneous.
To quantify BF, collect 10 mL sample in the field to
an empty scintillation vial, and
upon return to the lab, add 10 mL
borate buffer (see manuscript) and
read fluorescence.
No reaction period is necessary.
It is not necessary that bottles used for measuring BF are pre-reacted –
in fact, never add WR (with OPA) to
bottles used to measure BF.
MATRIX EFFECTS (ME):
OPA and ammonium react differently in different
waters.
In DI water, a given amount of
ammonium tends to produce more
fluorescence than it would in lake
or river or soil solution samples.
To quantify ME and correct for it,
standard additions are done to
samples and compared to DI water
standards.
For surface waters around Toolik
Lake, we have been spiking samples
with 50 ul of 50 uM ammonium stock
solution to quantify ME.
In general, ME have been around 5-25
%.
This correction is generally on the order of 0.01-0.03 uM for surface
waters around Toolik Lake, but will
be greater when ammonium
concentrations are greater.
Therefore it is important to note
that for higher ammonium values a
larger spike is required to assess
ME.
As with BF, ME appears to be
relatively constant within a given
water-body but will probably vary
across sites and maybe temporally.
In order to calculate matrix effect four measurements are required. 1. The fluorescence of a known amount of standard added to DI (spike std) this could be a standard from your standard curve. 2. A DI blank (0 std), which can also be from your standard
curve.
3. A sample spiked (sample spike) with the same amount of
standard as the spike standard
4. The fluorescence of the sample with only WR added, but
without any spike (sample obs).
The equation for calculating matrix effect is as follows:
(((spike std - 0 std)-(sample spike
- sample obs))/(spike std - 0
std!))*100
Procedure
It is important to measure the 10 mL sample accurately. Disposable 10 mL syringes may work well. Alternately 2x 5 ml delivered from a pipette could also work. Only open sample bottles for a short time, and be aware of potential sources of contamination when bottle is open. Rinse syringe thoroughly between samples.
Variant #1
In the field add 10 mL sample
to bottles pre-loaded with WR, shake
to mix, and store in dark. The reaction takes about 3 hours to
reach peak fluorescence (see
manuscript), so wait at least that
long after taking the last sample
before reading on the fluorometer.
Since the linear period of the fluorescence is from about 4-8 all samples
and standards need to be read within
that window. The speed of the reaction is temperature
dependant therefore samples and
standards should be at similar
temperatures when reagents are
added.
If this is not possible both should
be allowed to react at least 4
hours.
Variant #2
Samples are taken in the
field in pre reacted 20 ml scint
vials that have been rinsed with
fresh DI followed by three rinses in
the field with sample. Upon returning to the lab samples and standards are added to
pre reacted pre loaded scint vials.
Samples may either be read after 4
hours.
We are going to run some test to see
if these samples could be read after
reacting over night.
Standard Preparation
Variant #1 I (Max) prefer to make standards in the field. The fluorescence reaction is time
sensitive, so it is good to start
standards at roughly the same time
as samples are collected.
However, fluorescence asymptotes
after a few hours and stays there
for several hours, so there so
leeway here.
For surface water samples around
Toolik, standards ranging from 0 to
0.5 uM work well.
I recommend using a 50 uM stock
ammonium solution and a 10-100
ul Eppendorf pipette to make the
standards.
To make the standards, DI (10 mL) is added to bottles
loaded with WR (10 mL), and then
stock ammonium solution is added.
(I
have been adding DI to sample
bottles in the field, but it may be
possible to add both WR and DI to
standard bottles in the lab prior to
going to the field.
This will require testing before we
know if it works well).
Recipe for
Standards:
NOTE 1: Standard regressions have been fairly consistent, with a slope around 2.5 and y-intercept about 0.07 (Protocol B-1 using above standards). If this continues to be the case, it may be easier to make standards in the lab and not worry about field preparation. NOTE 2: We recently had a jump in blank fluorescence, from about 0.07 to about 0.15. This appears to be coming from the DI water. Since no DI water is added directly to samples (only WR is added), bad DI impacts standards but not samples. Therefore, if we can pinpoint the increased blank to ammonium in the DI (equivalent to only about 0.03 uM), we might want to adjust the intercept to the fluorescence of WR (or WR plus DI but read immediately) so that we will not underestimate the ammonium content of samples).
Variant #2 (made by nutrient RA)
Stock A Solution-1000mM NH4-N (made by nutrient RA)
Dissolve 0.06607g of dry (dried in oven overnight)
(NH4)2SO4
(m.w.= 132.14) in approximately
900ml of deionized water contained
in a 1 L volumetric flask. Dilute the solution to the mark with
deionized water and mix it well.
Transfer this solution to an amber bottle. Refrigerate when it is not in use.
Stock B Solution-20mM NH4-N (made by nutrient RA)
Prepare this solution daily. Using a volumetric pipette or a
calibrated automatic pipette, add
2.0ml of Stock A to approximately 90
mL of deionized water contained in a
100ml volumetric flask. Dilute the
solution to the mark with deionized
water and mix it well.
Working Standards (made by
nutrient RA)
Prepare these solutions daily. Use adjustable, microliter pipettes to
add the designated volumes of stock
A or B listed in the following
table. Calibrate the pipette for each required
volume.
(NOTE: The standards for each day are presented
in bold typeface.
The other concentrations are
included for reference, if needed.)
Prepare each standard by adding the required amount of stock to the
required volume volumetric flask
containing deionized water. Dilute each to the mark with deionized
water and mix well.
Keep these solutions tightly sealed.
Record the fluorescence for each standard and spike
sample in the nutrient logbook.
Determine the standard curve by plotting absorbance (y-axis) versus
standard concentration (x-axis).
We have created an Excel workbook,
entitled 2000lternh4fluor.xls, for
data input.
Method Detection Limit
The detection limit for this method should be
determined daily.
If the detection limits are
consistent for a couple of weeks,
then it will be necessary to perform
this task weekly.
Spike DI water with 2-3 times the estimated
instrument detection limit.
This should be around 0.2 to 0.3 mM. (Use the 0.2 or 0.3 mM standard as the spike.)
Run 7 spikes as samples after the
standard curve has been run.
Calculate the method detection limit (MDL) by
MDL= [t(7, 0.01)
* s]
Where t=t statistic for 7 reps
(t=3.14) with 99% confidence and
s=standard deviation of the
calculated concentration.
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
|
Please contact arc_im@mbl.eduwith questions, comments, or for technical assistance regarding this web site. | |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
