1.4.1 X-ray fluorescence (XRF) X-ray fluorescence spectrometry (XRF) is currently the most widely used analytical technique in the determination of the major and trace element chemistry of rock samples. It is versatile and can analyse up to 80 elements over a wide range of sensitivities, detecting concentrations from 100 % down to a few parts per million. It is a rapid method and large numbers of precise analyses can be made in a relatively short space of time. The chief limitation is that elements lighter than Na (atomic number — 11) cannot be analysed by XRF. Good reviews of the XRF method are given by Norrish and Chappell (1977), Tertian and Claisse (1982), Williams (1987) and Ahmedali (1989). X-ray fluorescence spectrometry is based upon the excitation of a sample by X-rays. A primary X-ray beam excites secondary X-rays (X-ray fluorescence) which have wavelengths characteristic of the elements present in the sample. The intensity of the secondary X-rays is used to determine the concentrations of the elements present by reference to calibration standards, with appropriate corrections being made for instrumental errors and the effects the composition of the sample has on its X-ray emission intensities. Alternatively, the X-rays may be detected without 1.4.2 Neutron activation analysis (INAA and RNAA) Neutron activation analysis is a sensitive and versatile method of rock analysis, chiefly applicable to trace elements and capable of determining a large number of elements simultaneously without necessarily destroying the sample. There are two approaches. Instrumental neutron activation analysis (INAA) employs a powdered rock or mineral sample; radiochemical neutron activation analysis (RNAA) involves the chemical separation of selected elements. The range of elements analysed is given in Table 1.5 and the methods are described in detail by Muecke (1980). In instrumental neutron activation analysis (INAA), about 100 mg of powdered rock or mineral sample is placed in a neutron flux in a neutron reactor together with standards. The sample and standards are irradiated for up to about 30 hours. The neutron flux gives rise to new, short-lived radioactive isotopes of the elements present which emit gamma radiations. Particular isotopes can be identified from the gamma radiations emitted and the intensities of these radiations are proportional to the amounts of the isotopes present. The gamma-ray spectrometry (the counting) is done at set intervals (several hours, several days and several weeks) after the irradiation in order to measure isotopes with different half-lives. Corrections are made for overlapping lines in the spectrum and concentrations are determined by comparison with the standards analysed at the same time. The method is particularly sensitive for the rare earth elements, the platinum group elements and a number of high field strength elements. When elemental concentrations are below about 2 ppm, a chemical separation may be employed following the irradiation of the sample, but prior to counting. This approach, known as radiochemical neutron activation analysis (RNAA), clearly has the advantage of increased sensitivity. 1.4.3 Inductively coupled plasma emission spectrometry (ICP) Inductively coupled plasma (ICP) emission spectrometry is a comparatively new technique with enormous potential in geochemistry. In principle the method is capable of measuring most elements in the periodic table with low detection limits and good precision over several orders of magnitude. Elements are measured simultaneously and a complete analysis can be made in the space of about two minutes, making it an extremely rapid analytical method. A full description of the method and its application is given by Walsh and Howie (1980) and Thompson and Walsh (1983). ICP emission spectrometry is a 'flame1 technique with a flame temperature in the range 6000-10 000 K. It is also a solution technique and standard silicate dissolution methods are employed. The sample solution is passed as an aerosol from a nebulizer into an argon plasma. The inductively coupled plasma is a stream of argon atoms, heated by the inductive heating of a radio-frequency coil and ignited by a high-frequency Tesla spark. The sample dissociates in the argon plasma and a large number of atomic and ionic spectral lines are excited. The spectral lines are detected by a range of photomultipliers, they are compared with calibration lines, and their intensities are converted into concentrations. .4 Atomic absorption spectrophotometry {AAS) Atomic absorption spectrophotometry is based upon the observation that atoms of an element can absorb electromagnetic radiation. This occurs when the element is atomized and the wavelength oflight absorbed is specific to each element. Thus the atomic absorption spectrophotometer comprises an atomizing device, a light source and a detector. A lowering of response in the detector during the atomization of a sample in a beam of light, as a consequence of atomic absorption, can be calibrated and is sensitive at the ppm level. The sample is prepared in solution and aspirated via a nebulizer and atomized in an acetylene-air or acetylene-nitrous oxide flame. The method is described in detail by Price (1972). There are two limitations to the routine use of AAS in silicate analysis. Firstly, the sample must be prepared as a solution, and secondly it is element-specific, i.e. only one element can usually be analysed at a time, although this latter limitation has in part been overcome by fitting instruments with multiple-turret lamp holders. Table 1.5 lists the elements which can be determined by AAS with relative ease. It can be seen, for instance, that all the major elements, with the exception of P, can be measured and detection limits for Na, K, Mg and Ca are extremely low. There are methods in which all the major elements may be determined from one solution, although the cathode lamp has to be changed for each element. The trace elements Ba, Be, Co, Cr, Cu, Li, Ni, Pb, Rb, Sr, V and Zn are also readily determined. AAS cannot compete with more rapid methods of silicate analysis such as XRF and ICP. Nevertheless, because AAS is comparatively cheap both in the capital outlay and in running costs, it most frequently finds its use in one of three specific applications. (1) The determination oflight elements such as Be and Li, with atomic numbers too low to be measured by XRF. (2) Routine analysis of transition metals readily leached from soil or stream sediment. This application is commonly used in geochemical exploration. (3) Non-routine trace element analysis using specialized applications of AAS in which, for example, the sample is atomized in a graphite furnace. This allows exceptionally low detection limits to be achieved for elements difficult to measure using other analytical techniques. 1.4.5 Mass spectrometry Mass spectrometry in its various forms is the most effective method of measuring isotope ratios. It is normally preceded by the chemical separation of the element of interest. Charged ions are generated from the element to be analysed either by the bombardment of a gaseous sample with electrons (gas source) or by the volatilization of the sample on a glowing filament made of a high-melting-point metal (solid source). The ion beam is fired along a curved tube through a very powerful electromagnet which splits up the atoms according to their mass. A mass spectrum is produced in which the lighter ions are deflected with a smaller radius of curvature than heavy ions. The quantitative detection of the signal at two or more mass numbers allows an isotope ratio to be calculated. Gas source mass spectrometry is used in stable isotope studies and in argon isotope geology, whereas Analytical methods in geochemistry 15 hotope dilution mass spectrometry (IDMS) Inductively coupled plasma emission mass spectrometry (ICP-MS) Spark source mass spectrometry (SSMS) solid source mass spectrometry is used in other geochronological and isotope geology applications (Rb-Sr, Pb-Pb, U-Pb and Sm-Nd) and in trace element analysis by isotope dilution. Isotope dilution mass spectrometry is the most accurate and most sensitive of all trace element analytical techniques and is particularly suited to measuring very low concentrations. The method is described in some detail by Henderson and Pankhurst (1984) and depends upon the addition of an isotopic tracer or 'spike' to the sample. The spike contains a known concentration of a particular element whose isotopic composition is also known. If a known amount of spike and a known amount of sample are mixed, and the isotope ratio of the mixture determined, the concentration of the element in the sample can be calculated. The method is particularly useful in determining the abundances of REE at low concentrations, although four of the REE (Pr, Tb, Ho and Tm) are mono-isotopic and cannot be analysed by this method. The main disadvantage is that even with automated mass spectrometry the method is time-consuming and expensive and so is normally reserved for measurements which can be used to calibrate other more rapid methods. ICP-MS is a relatively new technique extending from the development of inductively coupled plasma emission spectrometry (Date and Jarvis, 1989). It is becoming increasingly accepted as a tool for trace element and isotopic analysis as a result of the very low detection limits and good accuracy and precision. It can be used for analysing a wide range of trace elements, in a single solution, using a small sample (Jenner et al.y 1990). Ions are extracted from the plasma through a pinhole-sized orifice into a pumped vacuum system and focused with an ion lens into a mass spectrometer. This is a less widely used analytical method in geochemistry but has been used in the analysis of trace elements. The method is described by Taylor and Gorton (1977) and its usefulness was debated by Kronberg et al. (1988) and Jochum and Hofman (1989). It is capable of the simultaneous determination of about 40 trace elements, has high sensitivity with detection limits of 1-10 ppb, requires small amounts of sample and has high precision and accuracy when the isotope dilution technique is used (± 2-5 %). The sample is mixed with spiked graphite and briquetted into rod-shaped electrodes. A vacuum discharge is generated between the two sample electrodes. Elements are detected by mass on photoplates situated in the focal plane of the mass spectrometer. The mass spectra are analysed and ion intensities determined from line blackenings on the photoplate. Elements readily analysed by XRF, INAA, IDMS, AAS, ICP and ICP-MS Atomic Symbol Element No. XRF INAA IDMS AAS ICP 1 CP-MS 1 H Hydrogen 2 He Helium 3 Li Lithium X X X 4 Be Beryllium X 5 B Boron X 6 C Carbon 7 N Nitrogen 8 O Oxygen 9 F Fluorine 10 Ne Neon 11 Na Sodium X X X 12 Mg Magnesium X X X 13 Al Aluminium X X X 14 Si Silicon X X X 15 P Phosphorus X X 16 S Sulphur X 17 CI Chlorine X 18 Ar Argon 19 K Potassium X X X 20 Ca Calcium X X X 21 Sc Scandium X X X X 22 Ti Titanium X X X 23 V Vanadium X X X 24 Cr Chromium X X X X 25 Mil Manganese X X X 26 Fe Iron X X X 27 Co Cobalt X X X X 28 Ni Nickel X X X X 29 Cu Copper X X X 30 Zn Zinc X X X 31 Ga Gallium X 32 Ge Germanium X 33 As Arsenic X 34 Se Selenium 35 Br Bromine 36 Kr Krypton 37 Rb Rubidium X X X X 38 Sr Strontium X X X X X 39 Y Yttrium X X X 40 Zr Zirconium X X X 41 Nb Niobium X X X 42 Mo Molybdenum 43 Tc Technetium 44 Ru Ruthenium X 45 Rh Rhodium 46 Pd Palladium X 47 Ag Silver X 48 Cd Cadmium 49 In Indium 50 Sn Tin X 51 Sb Antimony 52 Te Tellurium 53 I Iodine 54 Xe Xenon Table 1,5 Continued Atomic Symbol Element No. XRF INAA IDMS AAS ICP ICPMS 55 Cs Caesium X 56 Ba Barium X X x X 57 La Lanthanum X X X x X 58 Ce Cerium X X X x X 59 Pr Praseodymium X x 60 Nd Neodymium X X X X x 61 Pm Promethium 62 Sm Samarium X X X x x 63 Eu Europium X X X x 64 Gd Gadolinium X X x x 65 Tb Terbium X v 66 Dy Dysprosium X X X A x 67 Ho Holmium x 68 Er Erbium X x V 69 Tm Thulium X A j£ 70 Yb Ytterbium X X x x 71 Lu Lutetium X X x x 72 Hf Hafnium X Y 73 Ta Tantalum X A 74 W Tungsten 75 Re Rhenium X 76 Os Osmium X 77 Ir Iridium X 78 Pt Platinum X 79 Au Gold X 80 Hg Mercury 81 Tl Thallium 82 Pb Lead X X X x 83 Bi Bismuth 84 Po Polonium 85 At Astatine 86 Rn Radon 87 Fr Francium 88 Ra Radium 89 Ac Actinium 90 Th Thorium X X X x 91 Pa Proactinium 92 U Uranium X X X X Table 13.1. Average Chemical and Normative Composition of N-MORB and Trace Elements (ppm) for N-MORB and E-MORB N-MORB N-MORB 49.9} Or 1.00 Ti02 1.51 Ah 22.06 A120, 15.90 An 31.14 FeO 10.43 Di 21.04 MnO 0.17 Hy 15.55 MgO 7-56 Ql 2.47 CaO 11.62 Mt 3.89 Na20 2.61 II 2.87 K20 0.17 Ap 0.19 p.o3 0.08 N-MORB E-MORB Cs 0.007 0.063 Rh 0.56 5.04 Ba 63 37 Til 0.12 0.6 U 0.47 0.18 Nb 233 83 Ta 0-132 0.47 La 2-5 6.3 Ce 7.5 15 Sr 90 155 Nd 7.3 9 Sm 2.65 2.6 Zv 74 73 Eu 1.02 0.91 Gd 3.68 2.97 Dy 4.53 3.55 Y 2S 22 Yb 3.05 2.37 1.-U 0.455 0354 Top, major oxides from McKeftrie and O'NiqnS (1991); bottom, data from Sun and McDonowgh (1989). Discrimination diagrams for basalts 10000 32 700 16600 - 8600 2670 1860 1000 10 15 43 95 (b) Zr (ppm) igure 5.2 Discrimination diagrams for basalts based upon Ti-Zr variations, (a) Linear scale (after Pearce and Cann, 1973); the fields are as follows: A, island-arc tholeiites; B, MORB, calc-alkali basalts and island-arc tholeiites; C, calc-alkali basalt; D, MORB. The plotting coordinates (extracted from Pearce and Cann, 1973 — Figure 2) are as follows: Point Zr Ti Point Zr Ti 1 12 1700 10 113 2 400 2 18 4600 11 65 2 400 3 52 8 500 12 39 3 100 4 84 10 400 13 31 4100 5 131 13 000 14 42 6000 6 167 10900 15 78 6 400 7 135 7 100 16 113 7400 8 192 6 600 17 65 5400 9 190 3400 18 89 7 400 (b) Log scale (after Pearce, 1982) showing the fields of volcanic-arc basalts, MORB and within-plate basalts. The values given on the ordinate and abscissa can be used as a guide in drawing the field boundaries (data extracted from Pearce, 1982). TV 100 Zr Yx3 Figure 5.1 The Ti-Zr-Y discrimination diagram for basalts (after Pearce and Cann, 1973). A is the field of island-arc tholeiites, C the field of calc-alkali basalts, D is the field of within-plate basalts and B is the field of MORB, island-arc tholeiites and calc-alkali basalts. Rocks which plot in field B give an ambiguous result but can be separated by plotting on a Ti-Zr diagram (Figure 5.2) or if unaltered on a Ti-Zr-Sr diagram (Figure 5.3). The plotting coordinates, extracted from Pearce and Cann (1973 — Figure 3) are: Point Ti/100 Zr Yx3 Point Ti/100 Zr Yx3 1 24 55.5 20.5 7 19 40 41 2 28 59 13 8 10 55 35 3 50 38.5 11.5 9 10 62.5 27.5 4 48 24 28 10 16 63 21 5 39 20.5 40.5 11 29.5 45 25.5 6 30 26 44 12 42.5 30 37.5 Per rot ec Lome Associations 3 1 10 16 36 60 77 112 137 226 268 596 (a) Zr Cppm) 8.7 S 3 N 1.46 1 10 16 60 80 303 (b) Zr Cppm) Figure 5.S Discrimination diagrams for basalts based upon Zr/Y-Zr variations; both diagrams have a logarithmic scale, (a) The fields are A, volcanic-arc basalts; B, MORB; C, within-plate basalts; D, MORB and volcanic-arc basalts; E, MORB and within-plate basalts (after Pearce and Norry, 1979). The values along the ordinate and abscissa are on a logarithmic scale and are given to assist in constructing the field boundaries (data extracted from Pearce and Norry, 1979 — Figure 3). (b) Fields of continental and oceanic-arc basalts separated on the basis of a Zr/Y value of 3. The shaded area is the field of overlap between the two basalt types (after Pearce* 1983). I_I_I_L_l_I_L_I_I_|_I I I I 1 I La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu Atomic number Figure 4J9 Rare earth element abundances (log scale) in the solar system plotted against atomic number.] Data from Anders and Ebihara (1982), normalized to Si = 106 atoms. Elements with even] atomic numbers have higher abundances than those with odd atomic numbers. ] i j 20 ti 4-1 •c O u U -a o u e rt -a c -C 10 - J_I_j_I_I_I_I_L J_I_I_I_I Figure 4.20 La Ce Pr Nd Pm Srn Eu Gd Tb Dy Ho Er Tm Yb Lu Atomic number Rare earth element abundances normalized to chondritic meteorite values plotted a atomic number for an Archaean tholeiitic basalt (sample 11 — Rollinson, 1983), The sample has been normalized using five different sets of values; the numbers refer to columns of normalizing values listed in Table 4.5. Table 4.5 Chondrite values used in normalizing REE (concentrations in ppm) Wakita Ha sktn Masuda Nakamura Evensen Boynton T & M Primitive Analytical mantle method NAA NAA IDMS IDMS IDMS IDMS IDMS value Chondrite(s) analysed: Composite Composite Leedey Composite Avg.CI Avg.CI Avg, CI (Ref.) (1) (2) (3) (4) (5) (6) (7) (8) La 0.340 0.330 0.3780 OJ290 0.244 60 0.3100 0.3670 0.7080 Ce 0.910 0.880 0.9760 0.8650 0.637 90 0.8080 0.9570 1.8330 Pr 0.121 0.112 0.096 37 0.1220 0.1370 0.2780 Nd 0.640 0.600 0.7160 0.6300 0.473 80 0.6000 0.7110 1.3660 Sm 0.195 0.181 0.2300 0.2030 0.154 00 0.1950 0,2310 0.4440 Eu 0.073 0.069 0.0866 0.0770 0.058 02 0.0735 0.0870 0.1680 Gd 0.260 0.249 0.3110 0.2760 0,204 30 0.2590 0.3060 0.5950 Tb 0,047 0.047 0.037 45 0.0474 0.0580 0.1080 Dy 0.300 0.3900 0.3430 0.254 10 0.3220 0.3810 0.7370 Ho 0.078 0.070 0.056 70 0.0718 0.0851 0.1630 Er 0.200 0.200 0.2550 0.2250 0.166 00 0.2100 0.2490 0.4790 Tm 0.032 0.030 0.025 61 0.0324 0,0356 0.0740 Yb 0.220 0.200 0.2490 0.2200 0.165 10 0.2090 0.2480 0.4800 Lu 0.034 0.034 0.0387 0.0339 0,025 39 0.0322 0.0381 0.0737 Y 2.1000 (1) Wakita et aL (1971); composite of 12 chondrites. (2) Haskin et ai (1968): composite of nine chondrites. (3) Masuda et at. (1973): Leedey chondrite. (4) Nakamura (1974). (5) Evensen et ai. (1978): average of CI chondrites. (6) Boymon (1984). (7) Taylor and McLennan (1985): 1.5 x values of Evensen [column (5)], (8) McDonough et al (1991).