EXPERIMENT 4
ATOMIC SPECTROMETRY - DETERMINATION OF CALCIUM AND MAGNESIUM
IN SAND WITH A STATISTICAL TREATMENT OF THE MEASUREMENTS
PURPOSE
The analysis of field samples requires three operations: (1) acquiring the raw samples employing representative sampling techniques; (2) preparing the raw samples to convert them to an analyzable form; (3) analyzing the prepared samples to determine the analyte(s) of interest. The results of the analysis will be affected not only by the raw sample composition but also by the three operations performed to collect the results. This experiment models a complete analysis of sand from sampling through instrumental measurement. Two types of atomic spectrometry will be used to perform the measurements. Analysis of variance (ANOVA) will be used to quantify the effects of the three analytical operations on the data obtained.
INTRODUCTION AND THEORY
Atomic spectrometry is an analytical measurement method relying on the spectroscopic processes of excitation and emission. Atoms are capable of undergoing electronic transitions due to absorption of light or through excitation and emission of light. Since atoms have no rotational or vibrational energy, these transitions are a function of the quantized energy levels available for the atom. These will be unique to the element and produce narrow, sharp, spectral bands when measured. Since the bands are narrow multiple elements are often resolved without interference as might occur in molecular spectroscopy.
The first type of atomic spectrometry used in this experiment is atomic absorption (AA), as shown in the following block diagram:
Electromagnetic radiation from a suitable source passes through the atomic aerosol, resulting in absorption, due to excitation of electrons to higher energy levels. The radiation transmitted by the sample then passes through the monochromator and is quantified by a detector. For AA, the signal strength reaching the detector is small when a large concentration of the analyte atoms is in the atomic aerosol.
The atomic aerosol for the instrument utilized in this experiment is provided by a combination of a concentric tube nebulizer and an air-acetylene flame. Pressurized air is passed through an outer tube causing the sample solution to be drawn into a concentric inner tube (see figure below) The velocity of the rushing air then breaks the sample solution stream into a fine mist. This mist is then mixed with acetylene and burned to produce a stable flame. The flame creates the atomic aerosol. Naturally, the sample must be in solution form in order to employ this process.
The very narrow absorption lines of atoms, about 0.02 Å, require the use of a discrete source. A continuous radiation source, such as the tungsten or deuterium lamps used in molecular spectroscopy cannot be employed. Such a source would swamp the signal. The very narrow absorption band would appear as only a minute change against the broad incident radiation, if it could be detected at all. In atomic absorption, a specialized source, the hollow cathode lamp, is generally used. A hollow cathode lamp is constructed such that the cathode contains the element of interest. The lamp used in this experiment contains both calcium and magnesium. When such a lamp is energized, atoms are vaporized from the cathode and excited by this process. The atoms then relax back to the ground state, emitting light that is used as the source radiation for AA. The hollow cathode lamp will emit an atomic spectrum that matches the spectrum for the analyte. This emission spectrum provides the source radiation. This should suggest a major drawback of AA. Lamps are required for each element to be measured. Even with multi-element lamps, such as used in this experiment, this can become cumbersome, especially if many elements must be measured. It also makes it impractical to perform qualitative work with AA.
The second atomic spectrometric method employed in this experiment is atomic emission (AE). When atoms are excited by the absorption of thermal energy to a higher electronic state, they will tend to relax to a lower state. This relaxation process is often accompanied by the emission of light. The measurement of this light emission can be used an analytical tool and is the basis for AE. The following block diagram describes the basic AE technique:
AE can take on many forms. Both solid and solution samples may be employed, in contrast to AA. This experiment focuses on the solution techniques. Excitation can occur by flame, furnace or noble gas plasma. By far the most popular technique is plasma. The very high temperature available in a plasma provides a stronger signal and makes many more elements available as compared to flame and furnace techniques. The air-acetylene flame used in this experiment provides only a small amount of excitation energy and is therefore useful only for easily excited elements like sodium and calcium. It does not provide sufficient excitation energy for magnesium. To summarize, in AE the thermal energy in the atomic aerosol provides excitation energy to the atoms. The atoms then relax and emit photons that can be detected. These photons, as with the absorbed photons in AA occur in a narrow band that is essentially unique to the element of interest. Unlike AA, in AE the signal strength reaching the detector will increase as the concentration of atoms in the aerosol increases.
There exists an interesting relationship between AA and AE measurements. Since the air-acetylene flame used in this experiment provides sufficient energy to cause emission for calcium, this could adversely effect the absorption measurements for calcium. As the calcium concentration in the flame increases, the emission signal at the detector will also increase. This occurs at the same wavelength used for absoprtion. This is in conflict with the absorption measurement where it is expected that the detected signal will decrease. This error is elminated by modulating the system. The source is pulsed to produce intermittent light for the absorption measurement.The detector is coupled to the frequency of this pulse. In this way, the detector can correct for the emission signal.
The most serious source of error in atomic spectrometry is chemical interference. Interferences occur when matrix components (materials in the sample other than the analyte), cause a variation in the observed absorbance of an analyte compared with that observed for a pure standard of the same concentration. This is alleviated by matrix matching. Whenever possible, standards are prepared in the same matrix as the samples. We will be using a solution of ammonium chloride and strontium chloride to prepare the sample solutions. The standard must use the same solution. Further chemical interferences are also present when analyzing sand and soil samples for calcium. Such samples may also contain phosphate, aluminum, silicates and sulfate. These can combine with calcium to form non-volatile compounds that are not readily available in the aerosol for AA or AE. This problem can be overcome by the addition of a releasing agent such as strontium. Strontium will also combine with these compounds. If it is added in excess it will tend to keep the calcium available for AA and AE. The purpose of the ammonium chloride in the preparation of the samples is to extract the calcium and magnesium from the sand, through an ion exchange process.
EXPERIMENTAL PROCEDURE
Supplies (provided by the TA)
1. Extraction Solution (25.9 g/L NH4Cl and 2.3 g/L SrCl2 in deionized water)
2. Stock Standard Solution (10 mg/L Mg and 50 mg/L Ca in the Extraction Solution)
3. Unknown Sample Solution (Mg and Ca in the Extraction Solution)
PART A: AQUISITION OF SAMPLES
Obtain samples from the sand-box. It has been divided into a 5' x 5' grid. Before arriving in the lab to perfrom the experiment, generate 3 pairs of random numbers corresponding to coordinates (x,y) on the grid. You can use a computer, calculator, random number table, or less preferably, dice or drawing lots. You don't want any bias to result in your selection.
Select three clean dry beakers to collect the samples. Carefully mix the sand in the grid locations selected. Obtain three separate sand samples of about 2 g from each of the selected grid locations. Please do not remove more sand than needed. Since there is no way to include the rocks or pebbles in our analysis try to avoid collecting these when sampling.
The initial analytical operation of sample collection is now complete.
PART B: CHEMICAL TREATMENT OF RAW SAMPLES TO CONVERT THEM TO ANALYZABLE FORM
In order to perform atomic spectrometry on the instrument, a solution is needed. In this experiment the calcium and magnesium will be extracted from the sand. The sand will then be filtered out and the resulting solutions analyzed. Complete extraction of Ca and Mg is assumed.
1. Choose one of the three samples. Using an analytical balance, weigh three 0.5 g portions of this sample into 3 separate 50 ml Erlenmeyer flasks..
2. For the remaining two samples, weigh a single 0.5 g portion each and place each into a 50 ml Erlenmeyer flask. You should have a total of 5 flasks containing sand samples.
3. Add approximately 25 ml of the Extraction Solution to each of the sand samples in the Erlenmeyer flasks. Seal the top of the flasks with parafilm.
4. Place these on a shaker, and shake vigorously for 30 minutes.
5. After shaking, filter the solutions using filter paper and glass funnels, with several rinses of Extraction Solution, into 50 ml volumetric flasks. Be careful to avoid spillage during filtration and rinsing. Be careful not to overfill the volumetric flasks during rinsing. Dilute to the mark with Extraction Solution. Seal the flasks tightly and mix well.
The second analytical operation is now complete. The samples are now in an analyzable form.
PART C: INSTRUMENTAL ANALYSIS OF THE SAMPLES
The instrument may now be started and prepared for the analysis. The Perkin Elmer 3100 atomic absorption spectrometer will be used. In order to obtain maximum sensitivity and precision, the instrument operating parameters must be optimized. Start by adjusting the measurement wavelength, lamp alignment, flame height, and gas flows and pressures.
I. Instrument Start-up and Optimization
1. Turn on the instrument. When the message "Perkin-Elmer-Model 3100" appears, press [Param Entry]. Enter the lamp current (probably 15 mA - consult the instrument log book for the specific value). Enter the lamp current value selected in the log book.
2. On the top of the instrument are some settings for the monochromator optics. Set the slit width lever to 0.7 nm and the slit height to "high."
3. Press [Energy]. A bar chart for CTS and EN will appear on the instrument display. This is a relative measure of the strength of the signal reaching the detector. Now locate the wavelength control dial on the left side of the instrument. Set the value to the optimum wavelength setting for Mg, 285.2 nm. The CTS and bar chart display should change as you rotate the dial. Delicately adjust the wavelength control dial so that a maximum value is obtained for CTS and EN. If the bar chart goes off scale during this adjustment, press the [Gain] key. The EN value should increase when this is done. By maximizing the CTS and EN values the wavelength setting is optimized.
4. Locate the vertical and horizontal alignment screws for the lamp on the top right side of the instrument. Slightly rotate each screw starting with the horizontal adjustment and ending with vertical. Once again maximize the CTS and EN readings. After reaching a maximal reading, record the CTS and EN values in the log book. Please record the wavelength and element under analysis as well.
5. Press [Cont]. An absorbance reading will appear on the display. Locate the vertical adjustment knob for the burner. Hold a piece of white paper in the light path to the determine the level of the light beam with respect to the burner, Rotate the vertical adjustment knob as necessary to assure that the burner is lowered completely out of the light path. Press [A/Z] to autozero the instrument.
6. Slowly rotate the vertical adjustment knob to raise the burner into the light path. This will be apparent when a slightly positive absorbance reading appears on the display. Now rotate the knob to slightly lower the burner until the reading returns to zero. The light beam is now at the top of the burner, this would be at the very base of the flame if it were lit. The temperature is slightly lower at the base of the flame due to desolvation of the aerosol; the hottest, cleanest portion of the flame is slightly higher. Atomic spectrometry requires completion of the desolvation process and vaporization of the individual atoms of analyte. The burner must be lower so that the light beam will be higher in the flame. Rotate the knob an additional 3/4 turn to achieve the optimum position.
7. Turn on the air and acetylene flow under supervision of the TA. Record the air and acetylene cylinder pressures in the log book. Note: do not start the instrument if the acetylene cylinder pressure is below 100 psig, or if air pressure is below 500 psig.
8. Turn the "Oxidant" switch on the instrument to "air." Carefully, under supervision of the TA, stand away from the burner and press the red "Ignite" button to light the burner. Record a start time in the instrument log book. Check the log book for the proper supply pressures and, as necessary, adjust the flow pressures for air and acetylene, record these in the log book (according to the instrument manufacturer, air supply pressure should be at 50 — 65 psig and acetylene at 12 — 14 psig). Consult the log book for the proper flow rates and, as necessary, adjust the flows for air and acetylene, record these in the log book (according to the instrument manufacturer, air flow should be 4.0 and acetylene should be between 2.0 to 2.5).
9. Aspirate deionized water.
The instrument is now ready to run. Make sure that the instrument is aspirating deionized water whenever it is lit and a determination is not in progress. The water provides necessary cooling to the burner assembly. Failure to do this could result in damage to the instrument. Use a large beaker to hold the deionized water for aspiration. Fill it completely and check it periodically to be sure that the sampling tube is in the liquid. The instrument needs about 15 to 20 minutes of warm-up to reach optimal stability.
II. Preparartion of the Standard Solutions and Calibration of the Instrument
1. Prepare the standard solutions as determined in the pre-laboratory assignment or as directed by the TA.
2. Before beginning the calibration, the nebulizer should be checked for proper operation. Care must be exercised in adjusting the nebulizer. Poor nebulization is probably the most likely source of poor results in this experiment. Please consult the TA while performing this operation.
a. Aspirate deionized water.
b. Slowly turn the nebulizer lock ring clockwise until it is well clear of the nebulizer adjustment knob.
c. Slowly turn the nebulizer adjustment knob counter-clockwise until bubbles begin to appear at the end of the sampling tube.
d. Aspirate Standard Solution #3 (The most concentrated standard).
e. Slowly turn the nebulizer adjustment knob clockwise until the absorbance goes through a maximum and just begins to decrease. The inversion point at which absorbance changes marks the optimal setting. Slowly turn the adjustment knob counter-clockwise to obtain the maximum stable absorbance. When this is done, aspirate deionized water once again.
f. Lock the nebulizer adjustment knob in its final position by turning the nebulizer locking ring counter-clockwise while holding the nebulizer adjustment knob securely.
3. Please remember to keep aspirating deionized water. This is especially important between analysis of the standards and samples. Because of the high salt concentrations in these solutions, extended aspiration of them will lead to build-up within the nebulizer and on the burner. This in turn will lead to poor instrument operation and poor results.
4. In making absorbance measurements for AA and AE it is common to average the signal over a specified time interval. This is called integration. It is likewise routine to make measurements that are actually averages of a collection of individual integrations. Flame atomic spectroscopy, as well as plasma and furnace techniques, have inherently noisy components. The flame is not perfectly stable, nor is the nebulizer consistent over a short time frame. This noise can be very effectively diminished by the averaging technique. Press [Param Entry] . Press [Enter] to reach the integration time screen. Enter an intergration time of 3 seconds. The instrument will then request the number of replicates, enter 3. Each measurement made for the standards and the samples will now be an average of three, three second integrations.
5. Press [Data] to enter the data acquisition mode where the multiple integration data points can be collected. Zero the spectrometer using deionized water, by pressing [A/Z].
6. Start the calibration with the Extraction Solution. This is the blank. Press [Read] to make a measurement. The display will show each of the three replicates and then report a mean, ABS(MEAN), standard deviation, SD, and relative standard deviation, RSD. Record data in lab notebook.
7. After measuring the Extraction Solution, measure each of the standard solutions in the same manner. Calibration measurement is now complete for magnesium.
III. Measurement of the Samples
1. Aspirate deionized water for at least two minutes after calibration.
2. Autozero the instrument with deionized water.
3. Measure the absorbance of each sand extract solution. To determine the precision of the instrument, choose one of the singular sample solutions, and measure its absorbance three times. Record data in lab notebook.
4. Measure the absorbance of the unknown sample solution along with the prepared samples.
5. Aspirate deionized water.
6. Set the wavelength dial to 422.7 nm. This is the analytical wavelength for calcium. Press [Energy] and rotate the wavelength control slightly so that the CTS and EN readings are at a maximum.
7. Repeat steps 4 — 6 of the calibration section II above for calcium.
8. Repeat steps 1 — 5 of this section for calcium.
The section on AA is now complete. Proceed to AE.
IV. Instrumental Set-up for Atomic Emission
1. Aspirate deionized water.
2. Press [Param Entry]. Enter a lamp current of zero to shut off the lamp.
3. Press [Em] to put the instrument in emission mode.
4. Apirate Standard Solution #3. Press [Gain] to adjust the detector for emission.
5. Aspirate deionized water.
V. Calibration and Measurement for Atomic Emission
1. Repeat steps 4 — 6 of the calibration section II above for calcium.
2. Measure only the emission of the unknown sample solution for comparison to AA data. Perform this measure in triplicate so that instrumental variance can also be compared.
The measurements are now complete. The instrument must run another 15 minutes while aspirating deionized water in order to clean the nebulizer and burner head.
To shut down the instrument, consult with the TA. First, remove the sampling tube from the water and allow it to run dry for about 30 seconds. Next, shut off the acetylene at the cylinder valve. The instrument will shut down automatically when it loses acetylene pressure. Turn the gas control switch off. Turn off the instrument power switch and close the air cylinder valve. Record the stop time in the instrument log book. Record any problems or observations about the instrument in the log book.
DATA REDUCTION AND ANALYSIS
1. Construct separate calibration plots of absorbance versus concentration for Ca and Mg. Construct a third plot of emission vs. concentration for Ca, include the data for the Extraction Solution (the blank) on these plots. Perform linear regression for these plots. The plots should have data plotted as points only, with the linear regression line drawn through the points. Include the equation of the line on the plots. In the results section of the report include a brief discussion concerning the linearity of these calibrations. Comparison of the linear regression y-intercept to the actual value (blank value) measured should be included in this discussion.
2. Use the calibration plots for AA of Ca and Mg to determine the concentrations in the sand samples. Include the unknown sand sample in this analysis, assume the sample weight for this sample to be exactly 0.5 g. Do not report the concentrations of the solutions. Report the concentrations in the sand, in µg/g. Prepare a table for the report showing the concentrations of Mg and Ca in the sand samples.
3. Use the calibration plot for emission of Ca to calculate the concentration of Ca in the unknown sample solution. Once again, do not report a solution concentration. Report the concentration of Ca in the sand in µg/g. Asssume a sample weight of exactly 0.5 g.
VARIANCE
When measurements are made, there are always indeterminate errors. It is, however, variance and not standard deviation that is propagated through any experiment or calculation. This is provided the variances are expressed as comparable quantities. Thus, when a series of events or operations is performed consecutively, the total variance introduced is the sum of the variance due to each operation. In an instrumental analysis samples are collected, prepared for analysis, and then analyzed with the instrument. The total variance (Vtotal) in the sample analysis will be the sum of the variances for each operation:
Vtotal = Vsampling + Vprep. + Vinstr.
CALCULATION OF VARIANCE
With the results from the AA analysis, Mg and Ca in the sand in µg/g, use the following definitions to calculate the variance. Complete the calculations separately for Mg and Ca.
1. Vtotal is the total variance of Mg or Ca in the four samples analyzed. In order to calculate this, data for all of the samples must be included and each data point must have the same measurement basis. In order to achieve this, calculate the variance using only one point from each of the triplicate measurements, plus the values measured for the sand sample measured singly and the unknown sample solution. The calculation will then include four data points.
2. Vinstr. is the variance due to the instrument. Calculate this using the concentrations from the single solution that was measured three times. Calculate this value for both atomic absorption and emission. The value from AA will be used below. The value for AE will be used later to compare the two techniques
3. (Vinstr. + Vprep.) is the variance due to the instrument plus that due to the sample preparation. Calculate this using the concentrations from the grid sample that was preparared and analyzed three times.
4. Vprep. = (Vinstr. + Vprep.) - Vinstr.
5. Vsampling = Vtotal - (Vinstr. + Vprep.)
This type of data analysis is often called ANOVA, an acronym for analysis of variance. Since variance is additive in this fashion it's analysis is especially well-suited to this situation. It can aid in understanding how each of the steps in the analytical process effect the results obtained. It is hoped that this experiment will show the value in measuring the contribution of each analytical operation to the data produced. This experiment is a laboratory model of a soil sampling experiment. As such, some of the grids in the sand-box have been spiked to create a non-homogenous sampling situation. It is often the case that small regions of topsoil, like an individual farmer's field, are not homogeneous with respect to soil nutrients like Ca and Mg. The same type of situation could arise in a manufacturing situation, such as pharmaceuticals, where a continuous production process feeds a batch packaging process. In either case this inhomogenous composition could have serious consequences. The farmer could improperly fertilize his field leading to poor yields. A drug company could sell a lot of a drug that is not uniform. In both of these cases with proper sampling and proper data analysis, the nature of the situation can be understood and appropriate steps can be taken.
COMPARISON OF ABSORPTION AND EMISSION
Prepare a table comparing absorption and emission. The following values should be included: Sensitivity (the slope of the calibration curves); Vinstr.; and the concentration of Ca in the unknown sand sample.
DISCUSSION
1. Explain the rationale behind the ANOVA calculations.
2. Comment on the relative magnitudes of the different sources of variance. What conclusions can be drawn from these values?
3. Compare, quantitatively, the variances obtained from the Mg and Ca data. In order to do so, you should convert the absolute values of variance to relative values based on the ratio of the varience for each step to the total varience. For example:
4. Discuss the two techniques of atomic absoprtion and emission, using the data obtained. Remember the instrument used is optimized for absorption measurements.
5. Discuss how the precision of this analysis could be improved. Relate the results of the ANOVA analysis to this discussion.
6. You are given the task of analyzing 20 rail hopper cars of ore for Ca and Mg content. You have sufficient time and resources to collect and analyze 100 samples, although fewer samples would be preferable. You have no information concerning the uniformity of the ore within the cars or from car-to-car. Give a detailed description of how you would collect the samples.
REFERENCES
1. Skoog, D.A.; West, D.M.; Holler, F.J. Fundamentals of Analytical Chemistry, 6th Edition, Saunders College Publishing, 1992, Chapter 20, Sections 20C, 20C-1; Chapter 24, Sections 24A, 24B and subsections, 24C.
2. Stroebel, H.A.; Heineman, W.R. Chemical Instrumentation: A Systematic Approach, 3rd Edition, Wiley, New York, NY, 1989, Chapter 14.
PRE-LABORATORY ASSIGNMENT
1. Define variance (an equation is sufficient).
2. A series of standard solutions of magnesium are measured by atomic absorption. The following absorbances were collected:
[Mg] in µg/ml Absorbance
0.00 0.023
0.20 0.075
0.50 0.149
1.00 0.273
A sand sample weighing 0.5081 g was treated as specified in the experiment. An absorbance of 0.089 was measured. What is the concentration, in µg/g, of magnesium in this sand sample?
3. Given a standard solution containing 50 µg/ml Ca and 10 µg/ml Mg specify, as a detailed procedure, how to prepare 100 ml each of the following standard solutions:
Solution # [Mg] in µg/ml [Ca] µg/ml
1 0.20 1.00
2 0.50 2.50
3 1.00 5.00
Note: This dilution procedure will be used during the experiment.
4. A lot of an analytical reagent was analyzed for iron. Five bottles out of the lot were sampled. For one of the bottles, five samples were analyzed. For the remaining bottle, one analysis was performed for each. The following results were obtained ( Fe in µg/g):
Bottle #1: 5.08, 5.12, 5.03, 5.11, 5.17
Bottle #2: 5.23
Bottle #3: 5.10
Bottle #4: 5.29
Bottle #5: 5.01
Calculate the mean iron content of this lot, in µg/g.
Calculate the total variance observed for the lot
Calculate the variance due to sampling.
Comment on the uniformity of this lot of reagent with respect to iron.
(Hint: Calculation of relative standard deviation would be helpful here.)