Inductively Coupled Plasma-Atomic Emission Spectroscopy

Inductively Coupled Plasma-Atomic Emission Spectroscopy 1 Theory ICP-AES, or Inductively Coupled Plasma-Atomic Emission Spectroscopy (also known as ICP-OES, Optical Emission Spectroscopy), is a type of emission spectroscopy that is often used to detect the presence of trace metals in a sample. Through the use of the eponymous Inductively Couple Plasma, an ICP-AES produces excited ions and atoms (by ionization in an intense electromagnetic field) that emit detectable amounts of light at characteristic wavelengths, with intensities proportional to the concentration of the ion. As indicated by the name, the spectra is measured and analyzed by an atomic emission spectrometer (AES) using concentration-intensity correlations that are similar to how the Beer-Lambert Law applies to AAS. ICP-AES is invaluable for its ability to record the spectra of multiple trace elements simultaneously, assuming they do not significantly overlap in characteristic wavelength, as well as the minimization of matrix effects due to the extreme nature of the plasma. If good wavelengths are chosen for the metals used, the amount of interference between spectra will be minimal; however, it is possible for interference effects to cause two spectra to overlap, wildly skewing readings. Nowadays, the presence of commercial ICP-AES software makes calculation of optimal wavelengths to minimize and account for overlap, through the use of proprietary software and interelement correction factors[1] . Like AAS, AES is widely used to measure the concentration or presence of trace elements in samples, and sees much modern use in environmental testing and protein analysis. 2 Results and Discussion This section contains only tabulated results from the Appendix. Derivations can be found in Appendix A on page 5. Graphs can be found in Appendix B on page 9. Raw data can be found in Appendix C on page 13. 1 2.1 Results 2.1.1 Unknown Concentrations Concentration (ppb) Metal Brass-58 Pre-Rain Creek Post-Rain Creek Difference Iron 57.3940291081 49.8540606889 50.6463112024 0.7922505135 Zinc 60.7833747246 -0.5819309431 -0.4453814841 0.136549459 Nickel 10.2397539775 11.0803296422 11.7444576539 0.6641280117 Manganese 2.5480694188 1.7674494946 1.7157049908 -0.0517445038 Copper 798.5072809986 10.2752139061 9.5453224703 -0.7298914358 Chromium -5.0852474686 -29.0073438581 10.750504919 39.7578487772 2.1.2 Brass-58 Percent Composition Metal % Total Copper 79.9% Zinc 6.08% Iron 5.74% Nickel 1.02% Manganese 0.255% Chromium -0.509% Total 92.5% 2.2 Discussion Using our calibration curves, we were able to successfully create calibration curves with good fit and use them to calculate values for our unknown compositions that, if not correct, are eminently reasonable. However, this process was not without its flaws: it is apparent by looking at the raw data that a large majority of our calibration data had to be thrown out due to a complete lack of fitting that we can only assume is due to instrumental or operator error, though our “standard 3” contained valuable information. Moreover, we were unable to gather any significant data regarding the presence or absence 2 of magnesium, and the abnormally high magnesium readings for our calibration curve only emphasize the anomalous fact that our unknown magnesium intensity values were so high that they were beyond the instrument’s capability to detect. As a result of these anomalies, we were forced to discard our magnesium data. However, excepting this, our data is quite reasonable. Some of our values as calculated predict negative quantities of compounds, something that is obviously impossible, but these compounds are without exception already near-zero in concentration, and so they can be explained away as instrumental variation. Looking first at our brass sample, we can see a clear abundance – almost 80% – of copper in the sample, something we might expect based on the general composition of brass as copper and zinc. In line with this reasoning, zinc is the next most abundant element in the sample, at about 6%. This number is comparatively small, and worryingly close to the next metal’s % composition (Iron, 5.74%), but it is not outside the realm of plausibility. The total of all the percent compositions of the brass sample is 92.5% – not 100%, but this is to be expected considering that we are only looking at a limited subset of all metals and it is plausible that the alloy contains metals we did not test for. Looking next at our creek sample, we can notice several reassuring trends. A majority of the metals tested for have near-identical concentrations before and after rainfall, something that is not outside the realm of belief considering that we do not know for sure the effects of rainfall on metal content. By comparing to experimentally-determined metal concentrations on the North Fork of Strawberry Creek, as measured during a 2006 water quality report[3] , we can determine that our measurements, though not identical, are of the same order of magnitude as the report’s values. We would not expect to see identical values in two different areas of the creek eight years apart, but it is reassuring to see that our calculations are of the same order of magnitude that we could reasonably expect a creek to have. For example, our concentration of iron is about 50 ppb, in line with the report’s 34 ppb, and our value of about 10 ppb for copper is not dissimilar from the report’s measurement of 4.0 ppb. There is the issue of our pre-rainfall chromium concentration, which is illogically low at -29 ppb, something we are unable to explain 3 outside of experimental or instrumental error – it’s possible that this is due to interelement interference that was not accounted for by the software. It is worth pointing out that the water quality report shows a chromium concentration of < 1 ppb, and while -29 ppb is certainly nonsensical we may be justified in simply rounding the concentration up to around 0 ppb. Though there is no easy quantitative way to determine and circumvent the presence of interelement interferences, this problem is neatly solved by the preexistence of interelement interference tables which can be used to determine characteristic wavelengths for metals that do not appear on the spectra of other elements. Many commercial ICP/OES software algorithms, including Perkin-Elmers MSF (Multi-component Spectral Fitting), and Varian/Agilents FACT (Fast Automated Curve-fitting Technique) automatically perform interference correction, or it can be calculated using interelement correction factors[1] . 2.3 Accuracy and Error All in all, our data is a mixed bag. Many of our calibration curves were worthless, failing to produce positive correlations at all, and had to be binned. However, the calibration curve used for the majority of our calculations is exceptionally clean, with R2 values that are strictly greater than 0.999. However, our unknown samples did not come out so neatly, and while many of our final concentrations are plausible there are several that are not. Additionally, unknown errors in our instrumental method led to the loss of our magnesium intensity data for our unknowns, a mysterious occurrence considering that the Mg data was apparently too high for the detector to read yet none of our samples should have contained such significant quantities of magnesium. Despite all of these setbacks, however, we still acquired final data values that we deemed acceptable. 4 References [1] ICP/OES ural Interference Maxims Resources. [Online]; Wisconsin Department of Nat- http://dnr.wi.gov/regulations/labcert/documents/ICP- OES Interference correction.pdf (accessed Mar 29, 2014). [2] Optima 7000 DV ICP-OES Manual ; Perkin-Elmer Informatics. http://www.perkinelmer.com/Content/relatedmaterials/productnotes/prd optima7000dv-icp-oes.pdf (accessed March 25, 2014). [3] University of California, Berkeley Strawberry Creek Water Quality 2006 Status Report [Online]; UC Berkeley Office of Environmental Health and Safety. http://strawberrycreek.berkeley.edu/naturalhistory/documents/SC2006WQStatuspdf10.02.2006.pd (accessed Mar 19, 2014). A Calculations A.1 Detection Limits Detection limits for this machine (Perkin-Elmer Optima 7000 DV ICP-AES) vary by element and were obtained from the product manual[2] . Detection limits for magnesium were not available, unfortunately, but luckily our samples contained an excess of magnesium, saturating the detector and rendering it unable to give a reading (so not knowing this doesn’t effect our results). However, this does mean that we were unable to determine the magnesium values in our creekwater samples. Element Wavelength (nm) Det. Lim. (ppb) Cr 267.716 0.25 Cu 224.700 0.9 Fe 259.939 0.2 Mn 257.610 0.03 Ni 231.604 0.4 5 A.2 A.2.1 Calibration Standards Masses This section deals with the specific salts we used to create our calibration samples, and the calculation of the precise concentrations of the metal ions in the solutions of these salts. Metal Salt g Salt (mg) Salt ( mol ) g Metal ( mol ) Metal (mmol) Metal (mg) Ni(NO3 )2 ·6 H2 O 183.66 290.795 58.693 0.63158 37.0695 Mn 62.22 54.938 54.938 1.133 62.220 MgSO4 261.2 120.37 24.305 2.170 52.741 Cu 53.300 63.546 63.546 0.839 53.300 Zn 52.990 65.38 65.38 0.810 52.990 CrCl3 ·6 H2 O 277.09 266.45 51.996 1.040 54.072 FeSO4 ·7 H2 O 265.45 278.01 55.845 0.954 53.322 A.2.2 Concentrations All solutions were prepared to a volume of 50.00 mL, then serially diluted a thousandfold down to an initial concentration of approximately 1000 ppb. This initial solution was diluted by a factor of half, then half again, then by two-fifths, producing calibration concentrations of approximately 500, 250, and 100 ppb. Precise concentration values vary from this figure, as the initial amount of salt added was not exact. 6 Concentration (ppb) Metal [Salt] Initial Ni(NO3 )2 ·6 H2 O 741.390 370.695 185.347 74.139 Mn 1244.4 622.2 311.1 124.44 MgSO4 1054.825 527.412 263.706 105.483 Cu 1066 533 266.5 106.6 Zn 1059.8 529.9 264.95 105.98 CrCl3 ·6 H2 O 1081.448 540.724 270.362 108.145 FeSO4 ·7 H2 O 1066.440 533.220 266.610 106.644 A.2.3 Dilution 1 Dilution 2 Dilution 3 Graphs and Curve-fitting By plotting the concentrations of each metal against the intensity of the measurement (intensity data can be found in Appendix C on page 13), we can create very good fits that allow us to extrapolate or interpolate the concentration of each metal in our unknown sample. Because of the wide variation in intensities, each metal has been given its own graph to allow for greater readability. For brevity, the graphs have been placed in Appendix B on page 9, while a summarized table of regression data is provided here for further calculation. Metal Slope Intercept R2 Nickel 92.672 -836.761 0.99944 Manganese 2917.194 -6252.767 0.99996 Magnesium 10248.772 1002551.090 0.99321 Copper 1259.992 20111.762 0.99999 Zinc 265.998 -1500.619 0.99987 Chromium 375.684 35554.204 0.99942 Iron 568.193 -27081.210 0.99977 7 A.2.4 Calculating Unknown Concentrations By taking the individual intensities of each element in our unknown samples (Brass 58, pre-rainfall creekwater, and post-rainfall creekwater) and putting them through the appropriate inverse functions (y = mx + b → x = y−b ), m we can easily derive the concen- trations of specific elements in our sample. Concentration (ppb) Metal Brass-58 Pre-Rain Creek Post-Rain Creek Difference Iron 57.3940291081 49.8540606889 50.6463112024 0.7922505135 Zinc 60.7833747246 -0.5819309431 -0.4453814841 0.136549459 Nickel 10.2397539775 11.0803296422 11.7444576539 0.6641280117 Manganese 2.5480694188 1.7674494946 1.7157049908 -0.0517445038 Copper 798.5072809986 10.2752139061 9.5453224703 -0.7298914358 Chromium -5.0852474686 -29.0073438581 10.750504919 39.7578487772 Magnesium N/A N/A N/A N/A A.2.5 Calculating Brass Percent Composition Based on a dilution factor of 1×106 times from the pure metal to the AES solution, we can determine the percentage composition of the brass-58 alloy as follows: Metal % Total Copper 79.9% Zinc 6.08% Iron 5.74% Nickel 1.02% Manganese 0.255% Chromium -0.509% Total 92.5% A similar composition breakdown can be created in an identical fashion for the creekwater unknowns, though of course since creekwater is not composed of pure metal abso8 lute concentration is more valuable than % composition, and since the creekwater was not diluted like the brass-58 sample further calculations are unnecessary. B Graphs 9 10 11 12 C Raw Data Sample ID Elem Wavelength Int (Corr) 1000 ppb brass std. Fe 259.939 53562.97356 1000 ppb brass std. Zn 213.857 350455.3171 1000 ppb brass std. Cu 324.752 727346.4845 1000 ppb brass std. Cr 283.563 371458.5782 1000 ppb brass std. Mg 279.553 13011258.27 500 ppb brass std. Fe 259.939 1951111.54 500 ppb brass std. Zn 213.857 212155.8309 500 ppb brass std. Cu 324.752 669912.1389 500 ppb brass std. Cr 283.563 571276.7991 500 ppb brass std. Mg 279.553 6273275.127 100 ppb brass std. Fe 259.939 101024.1545 100 ppb brass std. Zn 213.857 41424.21525 100 ppb brass std. Cu 324.752 92673.36258 100 ppb brass std. Cr 283.563 55431.62597 100 ppb brass std. Mg 279.553 1333372.371 1000 ppb creek std. Fe 259.939 584086.5927 1000 ppb creek std. Zn 213.857 357603.8548 1000 ppb creek std. Ni 341.476 98466.77701 1000 ppb creek std. Mn 257.61 3283494.973 1000 ppb creek std. Cu 324.752 1261947.342 500 ppb creek std. Fe 259.939 2584402.889 500 ppb creek std. Zn 213.857 24059.20028 500 ppb creek std. Ni 341.476 408.0885826 500 ppb creek std. Mn 257.61 3541.554524 500 ppb creek std. Cu 324.752 -18918.11446 250 ppb creek std. Fe 259.939 71770.81809 13 250 ppb creek std. Zn 213.857 86209.27897 250 ppb creek std. Ni 341.476 22331.0908 250 ppb creek std. Mn 257.61 701104.7696 250 ppb creek std. Cu 324.752 -6642.393715 100 ppb creek std. Fe 259.939 114768.3646 100 ppb creek std. Zn 213.857 51194.35715 100 ppb creek std. Ni 341.476 136333.8881 100 ppb creek std. Mn 257.61 603083.6916 100 ppb creek std. Cu 324.752 50992.65963 Brass 58 Fe 259.939 9506.242361 Brass 58 Zn 213.857 20258.8005 Brass 58 Ni 341.476 -88.83286601 Brass 58 Mn 257.61 1551.540778 Brass 58 Cu 324.752 -19286.60787 Brass 58 Cr 283.563 -28524.36487 Brass 58 Mg 279.553 -13989.41236 Pre Rain Creek Fe 259.939 413.5290421 Pre Rain Creek Zn 213.857 1446.663662 Pre Rain Creek Ni 341.476 -63.71230508 Pre Rain Creek Mn 257.61 1453.407284 Pre Rain Creek Cu 324.752 -22150.50774 Pre Rain Creek Cr 283.563 -27999.17551 Pre Rain Creek Mg 279.553 34425.24012 Post Rain Creek Fe 259.939 2567.677028 Post Rain Creek Zn 213.857 1421.788384 Post Rain Creek Ni 341.476 -86.07095416 Post Rain Creek Mn 257.61 1440.924438 Post Rain Creek Cu 324.752 -22131.4051 14 Post Rain Creek Cr 283.563 -27042.73665 Post Rain Creek Mg 279.553 4439.796171 1000 ppb std 3 Fe 259.939 394438.2914 1000 ppb std 3 Zn 213.857 354864.5174 1000 ppb std 3 Ni 341.476 97165.61924 1000 ppb std 3 Mn 257.61 3107903.083 1000 ppb std 3 Cu 324.752 1355719.205 1000 ppb std 3 Cr 283.563 439457.3669 1000 ppb std 3 Mg 279.553 12145232.71 500 ppb std 3 Fe 259.939 181651.4753 500 ppb std 3 Zn 213.857 174652.6958 500 ppb std 3 Ni 341.476 47045.24541 500 ppb std 3 Mn 257.61 1540296.514 500 ppb std 3 Cu 324.752 686492.1173 500 ppb std 3 Cr 283.563 244347.1294 500 ppb std 3 Mg 279.553 5934200.923 250 ppb std 3 Fe 259.939 81680.22402 250 ppb std 3 Zn 213.857 89444.17021 250 ppb std 3 Ni 341.476 24738.99611 250 ppb std 3 Mn 257.61 766792.045 250 ppb std 3 Cu 324.752 355779.4225 250 ppb std 3 Cr 283.563 136334.5832 250 ppb std 3 Mg 279.553 3878993.738 100 ppb std 3 Fe 259.939 13222.49232 100 ppb std 3 Zn 213.857 33443.68741 100 ppb std 3 Ni 341.476 8538.423244 100 ppb std 3 Mn 257.61 312995.8772 100 ppb std 3 Cu 324.752 152834.8798 15 100 ppb std 3 Cr 283.563 73701.54913 100 ppb std 3 Mg 279.553 2271731.809 Brass 58 Fe 259.939 5529.675581 Brass 58 Zn 213.857 14667.63711 Brass 58 Ni 341.476 112.1774806 Brass 58 Mn 257.61 1180.44582 Brass 58 Cu 324.752 1026224.548 Brass 58 Cr 283.563 33643.75789 Brass 58 Mg 279.553 N/A Pre Rain Creek Fe 259.939 1245.518305 Pre Rain Creek Zn 213.857 -1655.411467 Pre Rain Creek Ni 341.476 190.0753086 Pre Rain Creek Mn 257.61 -1096.773939 Pre Rain Creek Cu 324.752 33058.44932 Pre Rain Creek Cr 283.563 24656.60903 Pre Rain Creek Mg 279.553 N/A Post Rain Creek Fe 259.939 1695.669501 Post Rain Creek Zn 213.857 -1619.089584 Post Rain Creek Ni 341.476 251.6213797 Post Rain Creek Mn 257.61 -1247.722695 Post Rain Creek Cu 324.752 32138.79195 Post Rain Creek Cr 283.563 39592.99669 Post Rain Creek Mg 279.553 N/A 16