Deployment by the Mars Exploration Rovers of backscatter Mössbauer spectrometers offers an incredible opportunity to (1) elucidate the iron mineralogies of rocks, soils, and atmospheric dust and (2) gain insight into the physical event by which the mineralogy came into existence and consequently acquire information having potential for yielding ancient planetary history relevant to broad issues including the question of life. Determining the mineralogy is done by subjecting raw data to reduction algorithms and generating products known as Mössbauer parameters, which are highly characteristic. Mixed mineralogies are treated through deconvolution. Through being able to exploit Mössbauer measurements made at different temperatures during the Martian diurnal variation, and utilizing spectra obtained of the same sample with both the 6.4 and 14.4 keV energy channels, characteristics beyond oxidation state and stoichiometry are accessible. These include temperature sensitive magnetic transitions that can be either abrupt or gradual, variation of Mössbauer parameters with temperature, and variation of composition with depth, all of which can be dependent upon past processing affecting initial crystallization or subsequent alteration (“weathering”). The goal is to arrive at an understanding of environments including the very ancient.
If you can't find a tool you're looking for, please click the link at the top of the page to "Go to old article view". Alternatively, view our Knowledge Base articles for additional help. Your feedback is important to us, so please let us know if you have comments or ideas for improvement.
 The conduct of Mars Exploration Rover (MER) surface operations is constrained by practicalities of uplink and downlink communication opportunities per Martian solar day (sol), duration of evaluation and decision opportunities for the science team per sol, management for a particular sol of traverse and instrumental interrogation options relative to mission goals, and vagaries. The details of the Martian diurnal temperature cycle are of significance in the conduct of Mössbauer spectroscopy as will be shown here. All of this will impact how much of the full capability of the Mössbauer spectrometers can be realized. Consequently, familiarity with Mössbauer technique is mandatory for those nonspecialists involved in surface operations science extraction management. This is the goal considered during preparation of this document. Of course much of this can be said for the other instrumental methodologies deployed by MER. Also, the paper has been prepared from the perspective of personal engagement in the technique in order to pursue issues in planetary science and astrobiology. While the reader is spared from mathematical detail, references are provided in that regard. We have utilized some backscatter spectra obtained in the course of development of the MER instrument in order to introduce the reader to the process by which data can be interpreted. Also there is discussion of spectra of terrestrial impact and hydrothermally derived materials that might serve as analogs for what might be present on Mars.
2. Mössbauer Spectroscopy Fundamentals in the Mars Exploration Rover (MER) Context
 Iron Mössbauer spectroscopy as deployed by the Mars Exploration Rovers utilizes the extremely narrow energy width (∼1 part in 1012) of a nuclear resonance involving gamma radiation emitted from a suitable source and 57Fe nuclei resident in a target sample. While Mössbauer spectroscopy done in the laboratory is usually a transmission technique, the MER instrument is a backscatter system. This makes it ideal for interrogation of samples on Mars with gamma radiation. The response is backscattering of that radiation and also emission of X-ray radiation through an internal conversion process that is subsequent to the 57Fe nucleus being excited to a short-lived state that decays by various processes, involving electrons and X rays. For the MER instrument the Mössbauer event is sensed by either resonance radiation of a 14.4 keV nuclear gamma ray or emission of a 6.4 keV X ray during restructuring of the external electron cloud following internal conversion. As will be shown later, this makes for some interesting possibilities in the analysis of surface structure of a sample.
 The Mössbauer effect [Mössbauer, 1958] occurs because both the emitting nucleus and the absorbing/scattering nucleus are constrained from recoiling by the atoms being incorporated into the rigidity of a solid-state structure, which, if the recoil occurred, would result in a substantial energy shift. This is because photons possess momentum, and when emitted by a source, that source must recoil with equal and opposite momentum (conservation of momentum). This also happens when a photon is scattered or absorbed. In the Mössbauer effect, the recoil momentum that would have been accrued to the nucleus, instead can be distributed over the mass of the crystal in which the atom containing the nucleus is resident. The portion of emitting or scattered/absorbing nuclei involved in the phenomenon is called the recoil-free fraction. The emitted photon from a nucleus relieved of recoil by this mechanism (the Mössbauer effect), is consequently extremely monoenergetic. This means the effect does not occur in gases or liquids. It is also of historical interest to note the effect was discovered for an isotope nucleus other than 57Fe, namely 191Ir.
 Because of the extremely monoenergetic character of the resonance, Mössbauer spectroscopy permits detection of subtle interactions with the nucleus by extranuclear electric and magnetic fields. These hyperfine interactions are the consequence of the atomic and molecular environment, as well as imposed conditions such as an external magnetic field. For 57Fe to emit a gamma ray it must be in an excited state. The most practical means of obtaining excited 57Fe nuclei is to derive the species from radioactive 57Co which enjoys a capture of an inner shell K electron resulting in a weak force interaction that causes a proton converting to a neutron with the emission of neutrino. The half-life for this process is 270 days, which is adequate for the duration of MER mission from launch to termination. Through this process, the nucleus becomes 57Fe and importantly it is in an excited state. The 57Co is manufactured by bombarding 56Fe with protons using an accelerator, followed by chemical separation of 57Co and incorporating it into a rigid solid-state lattice of the appropriate habit. Typically rhodium is used for the host lattice, as is the case for the MER instrument.
 The nuclear decay process resulting in emission of the gamma ray involved in 57Fe Mössbauer spectroscopy is best described by a simple diagram (Figure 1). The following two aspects of the process should be noted: (1) Gamma radiation other than the 14.4 keV emission utilized for Mössbauer spectroscopy also occurs, and (2) 91% of the time the process results in 14.4 keV emission.
 Key to the use of the Mössbauer effect as a spectroscopic tool is the <1 part in 1012 narrowness in energy for emission of the source. Given the 14.4 keV transition energy from 57Fe in the excited state (after transformation from 57Co), this translates to <1.4 × 10−8 electron volts! There are two ramifications because of this remarkable circumstance.
 1. The 57Fe nuclei resident in the atoms at an abundance of 2.14% of the total iron, of the sample, are shifted out of the resonance condition relative to the incident gamma ray energy by a variety of influences characteristic of the condition of the atom in which the nucleus is resident: its valence state, the molecular structure in which that atom participates, including magnetic ordering, and externally imposed physical conditions such as a magnetic field. Although not relevant to MER, a dramatic demonstration of the degree of sensitivity is that the resonance can even be affected in a detectable manner over a sufficient vertical distance because of a difference in gravity between source and sample, the relativistic gravitational redshift.
 2. The energy of the source photons can be readily increased or decreased in a controlled fashion by the Doppler effect using translation velocities in the millimeter per second range. The relationship is E = E0 × (1 + v/c). Thus a resonance with a sample 57Fe nucleus can be scanned for by simple translation of the source relative to the sample. Detection of energy shifts induced in the sample nuclei can be accommodated through Doppler shifting the energy of the source by velocities of not much more than ±10 mm s−1 from stationary. This is readily accomplished through electromechanical engineering not unlike that involved in a loudspeaker, even for a miniature system such as the MER instrument, which will be described later.
 Implementation of Mössbauer spectroscopy involves a 57Co source; an electro-mechanical transducer for shifting the source energy; electronics for imposing an appropriate velocity pattern on the transducer; sample; an energy dispersive detector, as in the MER instrument, in position to view backscatter radiation or as usually done in the laboratory, on the side opposite of that being illuminated (transmission mode); electronic circuitry that passes through only detector pulses indicative of 14.4 and 6.4 keV radiation signaling the Mössbauer event and filtering out all other pulses; electronics for binning event counts according to the velocity of the transducer at the moment; and, of course, the means for data storage, output, and processing. Figure 2 exhibits the architecture of an instrument system in schematic form. The flow of events is the following:
 1. The 57Co nuclei transform to excited 57Fe nuclei.
 2. The excited 57Fe nuclei emit incredibly mono-energetic gamma ray photons at ∼14.4 keV, with the precise energy of a photon dependent upon the velocity of the source being moved relative to a sample by the electromechanical transducer. Because a velocity difference between the source and sample is involved, the measurement must be performed under conditions where vibrations and other disturbances do not interfere with the measurement.
 3. If the energy of the gamma ray photon is coincident with an energy level difference of a 57Fe nucleus resident in the sample there is a resonance interaction. Otherwise, the photon is absorbed by other processes involving all variety of atoms comprising the sample or in the case of a “thin” sample just passes through.
 4. The resonance interaction raises the sample 57Fe nucleus to an excited state that has a short half-life of 98 ns, a duration which has implications with regard to magnetically ordered species. This is followed by deexcitation through the emission of a 14.4 keV photon in any of all possible directions (4π geometry) signaling the resonance scattering event or by other processes involving electrons and X rays (this is called internal conversion) including the emission of a 6.4 keV X ray in all possible directions (4π geometry) signaling the event. There are useful implications for this double signaling.
 5. For the MER instrument the presence of the resonance scattered 14.4 keV gamma ray photon is sensed directly by placing detectors around the source-sample axis with the source behind, and shielded, so as to face the sample to collect backscattered radiation only, hence the term backscatter Mössbauer spectroscopy. The detectors also sense 6.4 keV internal conversion X-ray photons emitted in their direction. Typically, in the laboratory, Mössbauer events are inferred by a decrease at certain velocities in photon flux after passing through the sample, hence the term transmission Mössbauer spectroscopy.
 6. The detectors are energy dispersive devices where the height of the electrical output pulse is indicative of the incident photon energy. The MER instrument has solid-state devices and amplifiers, which are described later. Laboratory instruments employ either the “classic” gas-filled proportional counter or solid state devices. The output pulses are directed to “window” filters (single-channel analyzers) so that only pulses in a narrow range of height pass through, eliminating signals other than those indicative of 14.4 keV gamma ray or 6.4 keV X-ray photons. These filters also direct pulses of these specific events respectively into separate pathways for separate analysis.
 7. After exiting the energy window pass filters, the pulses are binned according to the transducer/source velocity associated with each pulse. A record of the raw Mössbauer spectrum is accumulated in the form of event counts versus transducer/source velocity over a range between maximum negative velocity (downshift in energy) and maximum positive velocity (upshift in energy) induced by the Doppler effect. Because both 14.4 keV gamma ray photons and 6.4 keV X-ray photons are utilized, two separate records exist.
 For the more familiar spectroscopies the attention of the analyst is drawn to spectral structure including determination of “peak” positions, often rendering the analysis to a process of pattern matching. While there are obvious spectral characters for Mössbauer spectroscopy, elucidation of molecular structure and other physical aspects such as oxidation state and magnetic properties is carried out by reduction of the data to obtain rather specific products called the Mössbauer parameters. When an iron Mössbauer spectrum is obtained and analyzed, these parameters, characteristic of the mineralogy involved, can be determined. Important for MER are (1) the chemical isomer shift, designated as IS and expressed relative to an α-iron metal standard at 300 K, which informs about the valence state because it is related to the density of the electron charge at the Fe57 nucleus, (2) the electric quadrupole splitting designated as QS, which provides information regarding the local crystal structure in the nuclear environment due to distortion from cubic symmetry, and with the IS the valence state, and (3) the effective magnetic hyperfine field at the nucleus, designated by Bhf, which is a consequence of the magnetic state of the substance. In the case of substances that are not magnetically ordered and have no externally imposed magnetic field (i.e., Bhf = 0), the Mössbauer spectrum exhibits a two-line, or doublet, character. The spectrum can also be a singlet when QS = 0, as in the case for spinels and Fe+6 compounds. When a substance is magnetically ordered (Bhf ≠ 0), there is a dramatic difference in that the spectrum contains of six lines (sextet). So for Mössbauer spectroscopy it is the pattern, and its structure, that matters more than the individual peaks.
 Hyperfine interactions occur because the nucleus is an extended object and is nonspherical. The chemical isomer shift (IS) comes about because there is a difference in the energy of an isolated (“bare”) nucleus and one in the environment of the collective electric field of the electron cloud that exists in real atoms. The configuration of the electron cloud is determined by the nature of the species, including degree of oxidation. Isomer shifts are reported relative to a standard material, usually α-iron foil. Sodium nitroprusside is also used for this purpose. Either of these can be enriched in 57Fe if desired. There is a 0.257 mm s−1 isomer shift of sodium nitroprusside relative to iron foil. For mathematical details regarding isomer shift, the reader with mineralogical interest is referred to Hawthorne .
 Quadrupole splitting (QS) occurs when the upper energy level (excited state) involved in the Mössbauer event is split into two levels by action of the nonuniform electric field of the electron cloud upon the quadrupole moment of the nucleus. Because the nucleus is composed of protons and neutrons, the condition for the whole nucleus is that of a cooperative process from all involved charges leading in the case of the excited 57Fe nucleus to nonspherical charge distribution, a nuclear quadrupole moment. Only the excited state of the nucleus has a nonspherical charge distribution. Again the collective electric field of the electron cloud is involved in that it splits the upper energy level (hyperfine splitting) rather that just inducing a shift as for the chemical isomer shift. The difference between the two components of the doublet, measured in millimeters per second, is the quadrupole splitting and is very characteristic of the species because of the defining structure of the electron cloud. The two components of the doublet usually have near-identical intensities, however the nature of a sample including texture and anisotropy of the stiffness of the host lattice, can significantly affect the relative intensities. As before the interested reader is referred to Hawthorne  for an introduction to mathematical detail as related to the physics.
 Study of magnetic properties is a fascinating application of the spectroscopy and is expected to be a considerable interest on Mars where magnetically ordered species such hematite are expected in at least one of the MER landing site candidates. Magnetic nature as a temperature sensitive condition is of interest given the extremes in the Martian diurnal cycle. For dust samples collected on the MER magnets [Madsen et al., 2003], there can be opportunity for examining the effect of an imposed magnetic field. For those familiar with the optical Zeeman effect, what follows can be viewed as an extension to the nuclear situation, nuclear Zeeman effect.
 Because the nucleus has a magnetic dipole moment, it will interact with a magnetic field that can be either imposed in a macroscopic situation, a laboratory magnet or other generated field, or is a consequence of a microscopic cause involving magnetic ordering in an ensemble of the mineral structure in which the nucleus resides. The latter involving magnetic ordered minerals will be of considerable interest on Mars; however, also the former involving an imposed magnetic field will be of interest when spectra of dust on the permanent magnets are obtained. Material such as atmospheric dust adhering to these instrumental components can be interrogated because the spectrometer is manipulated with the robotic arm called the Instrument Deployment Device (IDD).
 The lower energy level (ground state) involved in the Mössbauer event is split into two levels under a magnetic field while the upper level (excited state) becomes four levels as shown in a portion of Figure 3. The diagram also denotes the spins associated with each energy level, important factors for understanding the spectrum that results. A simple combination of two lower levels and four upper levels suggests eight possible transition pathways; however, there are quantum mechanical selection rules that a physical system adheres to. For a high-probability transition, “permitted,” to occur, there must be a resulting spin difference between the two levels involved of either 0 or ±1. Otherwise the transition has such a low probability of occurring that it is described as being “forbidden“. Transitions involving a spin difference of 0 would be those for levels spins of +1/2 and +1/2, and −1/2 and −1/2. Those involving a spin difference of ±1 would be +1/2 and −1/2, −1/2 and +1/2, +1/2 and +3/2, and −1/2 and −3/2. The total number of high probability or “permitted” transitions is therefore six, resulting in six spectral peaks (Figure 3, bottom) where for the purpose of schematic illustration are shown in intensity ratios of 3:2:1:1:2:3.
 There is an interesting caveat regarding the nuclear Zeeman effect and it is that the magnetic field must have a duration at least as long as the time the 57Fe nucleus requires to sense the magnetic field, a little more than 10 ns. While an imposed macroscopic field is expected to be present over a considerably longer time, microscopic situations might in some cases be of very short duration because of thermal conditions. The Martian diurnal temperature cycle can serve as an important experimental variable in such cases as will be described later.
3. Brief Introduction to the MER Mössbauer Spectrometer
 As a backscatter instrument, the MER spectrometer has geometry that is different from what is generally utilized with the laboratory setting where transmission spectroscopy is the usual case. Backscatter geometry also results in greater angular relationships between source-sample-detector than is the norm in transmission studies. There are data analysis issues in this regard, such as cosine effects (“cosine smearing”), that distort the spectrum and consequently must be compensated for. Such data analysis issues go beyond the purpose of this tutorial, which is to acquaint the nonspecialist with the potential of the technique in a planetary surface setting. The MER instrument is miniaturized and consequently, from translation amplitude and resonance considerations, the transducer functions normally at the frequency of 24.32 Hz and can be set to other frequencies as desired. This is higher than the ∼5–10 Hz at which larger laboratory instruments generally function. The MER instrument, however, can be set at other translation frequencies if desired. Because the instrument is in interplanetary transit for an appreciable fraction of the 57Co half-life, the source is in a sense “overloaded” prior to launch to ensure sufficient intensity during surface operations of at least 90 sols after landing. While all of these attributes render the MER instrument somewhat unique, the utilization of the device is not much different than what a Mössbauer user experiences and in some aspects there are distinct advantages to be realized. It should be a “jewel” of an instrument to work with.
 Because the MER spectrometer is described in detail elsewhere [Klingelhöfer et al., 2003], only an outline of the basic characteristics of the instrument, particularly those aspects relevant to the tutorial intent of this paper, will be covered here. It consists of three components: the sensor head on the turret of the IDD, a flexprint cable bundle, and an electronics board resident in the warm environment box of the rover (its “body”). The sensor head is in the shape of a rectangular prism that is capped with an annular contact plate that mates the instrument to the sample surfaces and indicates this with a coupled switch. The contact plate can be inserted into the hole made by the Rock Abrasion Tool (RAT), which is on the IDD, for preparing a fresh surface down into a depth of ∼5 mm. A proper mating with the target rock is thus insured. This contact plate is instrumented with a temperature sensor to ascertain sample surface temperature. A second temperature sensor resides internally at the other end of the instrument.
 Within the instrument is the electromechanical transducer that translates back and forth in a velocity profile specified by the waveform generated by the electronics. At the forward end facing the sample, and mounted on the end of the transducer, is a 57Co source incorporated into a rhodium matrix. The main source strength at prelaunch loading, to take place 90 days prior to launch, is ∼300 mCi (1.11 × 1010 Bq). It is surrounded by a collimator composed of lead (Pb), red brass (Cu, Sn), and tantalum (Ta) to constrain the geometry of the emitted radiation away from the detectors and toward the target sample. This device is constructed in the form of what is called a graded shield to minimize the escape of Compton scattered and fluorescent X radiation arising in the collimator itself. On the opposite end of the transducer is a second 57Co source of 5–15 mCi (1.85 × 108–5.55 × 108 Bq), that serves in concert with a stationary α-Fe/α-Fe2O3 combination absorber intimate with a solid-state energy dispersive detector, for obtaining a Mössbauer spectrum to provide internal velocity monitoring and calibration. Note that this is a transmission spectrum because of the geometry involved. First use of the internal calibration on Mars will take place during the first 5 sols prior to rover egress from the landing system.
 Fronting the instrument to view backscattered 14.4 keV and internal conversion 6.4 keV radiations are four square solid-state silicon PIN detectors. A fifth Si PIN detector serves in the internal monitoring and calibration system mentioned in the previous paragraph. All detectors are energy dispersive in that pulses generated by incident radiation have magnitudes proportional to the energy of the radiation. Figure 4 shows in schematic form these key aspects of the sensor head. Figure 5 is an image of a geometric model chosen for simplicity to show the placement of the Si-PIN detector view apertures in the context of the contact plate. The combined geometry of the detector aperture placement and the annular contact plate results in a field of view on the target for the sensor head that is 1.0–1.5 cm in diameter. The flight instrument incorporates a contact switch to indicate placement of the contact plate on the sample, and a temperature sensor resides on the contact plate. It should be noted that the annular contact plate fits within the circular area that can be abraded on a rock by the RAT that also resides on the IDD.
 External to the sensor head is a calibration target, called the Composition Calibration Target (CCT) that is accessed by the head being manipulated with the robotic arm (IDD). The CCT is first utilized upon rover egress from the landing system to check the instrument's condition, with a succession of measurements being made over a temperature range imposed by the Martian diurnal cycle. Later during mission surface operations, the sensor head is returned periodically to the CCT to assess the condition of the instrument. The CCT has a composition of natural terrestrial magnetite and is 4.6 cm in diameter with a thickness of 0.48 cm.
 In addition to being able to access the CCT, two of the permanent magnet assemblies of the Magnetic Properties Experiments [Madsen et al., 2003] are available by IDD manipulation of the sensor head. These magnets are for study of atmospheric dust that settles out or surface material that is deposited by aeolian action. Active areas of the Capture Magnet and the Filter Magnet are 25 mm across, which is compatible with the field of view of the sensor head when placed in contact position. The samarium-cobalt (Sm2Co17) composition may contain iron impurities but are covered by the high-purity aluminum of the mount structures rendering them “not apparent“ to resonant 57Fe Mössbauer spectroscopy. To enhance the ability of the magnets to preferentially capture magnetic particles from the airborne dust, the magnets are tilted 45 degrees with respect to the z axis of the rover. Estimation of nonmagnetic material that does collect can be made by identifying differences in populations of magnetic and nonmagnetic (in some cases iron containing) minerals on the two different magnets. Such analysis and identification will be made using the Mössbauer and Alpha Particle X-Ray (APX) spectrometers. The Capture Magnet has a field strength of ∼0.4 Tesla (T) and a field gradient of ∼350 Tm−1, while the Filter Magnet exhibits a field strength of ∼0.2 T and a field gradient of ∼33 Tm−1.
4. Elucidation of Mineralogy and Oxidation States Via Mössbauer Parameters
 As stated earlier, Mössbauer spectroscopy yields rather specific data products called Mössbauer parameters. These are the IS, QS, and the Bhf. The precise values of the principal spectral parameters, IS, QS, and Bhf, serve as a “fingerprint” in identifying the particular mineral phase or molecular structure [see, e.g., Bancroft, 1973; Mitra, 1992; McCammon, 1995]. The first two are expressed in millimeters per second, while the unit of the third is tesla, the derivation of which will be explained shortly. Figure 6 illustrates the three distinct spectral characters in terms of each responsible physical transition process. A “real” spectrum is likely to have overlapping components of at least two and often all of these spectral characters for each of the species resident in the sample. These all too common circumstances demand spectral deconvolution algorithms to extract the individual spectral contributions of all iron bearing resident species, and very significantly derive the characteristic Mössbauer parameters that define what is present in the sample!
 Because oxidation state is of such general interest, it is worth discussing its determination first. Quadrupole splitting is very characteristic of the structure of the electron cloud surrounding the nucleus and its interaction with the nuclear quadrupole moment. The QS is thus expected to be different for different states of oxidation. When the quadrupole splitting is small, the ferric ion (Fe3+) is usually indicated while a more widely separated doublet suggests ferrous iron (Fe2+). The result is that even simple inspection of raw data product indicates the oxidation state of dominant species. Mixtures of oxidation states can often be recognized quickly with experience. This can provide the analyst with a “quick-look” impression of what comprehensive processing will precisely yield in terms of major species. Having said this, there are some exceptions to these simple guidelines. Pyrite, for example, has almost the same Mössbauer parameters as lepidocrocite at 300 K but has a different oxidation state.
 A nice example is the 14.4 keV backscatter spectrum (Figure 7) of an olivine (variety (var.) Forsterite) sample, which shows a doublet feature of obvious wide quadrupole splitting for ferrous iron, in this case ∼3 mm s−1, the profile of which also has superimposed upon it a less split doublet due to a less abundant ferric component. This blending manifests itself in the left-hand peak of the data being higher than that of the right-hand and furthermore a shoulder on its right-hand slope. It also illustrates the effect due to isomer shift in that spacing of the two peaks attributable to the dominant ferrous fraction is not symmetric about the zero velocity point. Deconvolution reveals that indeed two spectral components exist, a ferrous and a ferric, the profiles of which are shown by solid lines imposed on the count points distribution. The precise values for the IS and OS for the ferrous doublet provide the identification of olivine in the sample.
 A montage (Figure 8) of five spectra adapted from Morris et al.  and Klingelhöfer et al. , of various samples acquired with the Miniature Mössbauer Spectrometer (MIMOS) Mössbauer backscatter spectrometer which was a precursor to the Athena MER instrument, provides the reader with an opportunity for a “reading” exercise with a variety of backscatter spectra having overlapping component spectral character. The spectrum in the middle of the montage of a sample containing both siderite (FeCO3) and goethite (α-FeOOH), readily exhibits quadrupole splitting of ∼3 mm s−1 indicative of ferrous iron of siderite, reminiscent of what was shown previously in Figure 7. On the other hand, the spectrum second from the top, of jarosite exhibits an obvious lesser degree of quadrupole splitting which is a consequence of the iron being ferric. Progressing to the top spectrum of palagonitic tephra, along with the obvious indication of ferric-induced quadrupole splitting, there is a contribution of ferrous-caused quadrupole splitting although it is a lesser contributor to the overall spectrum. This manifests itself in that the palagonite spectrum shows the left of center (−v) peak to be enhanced, indicating a contribution due to blending with the left side (−v) companion to the obvious right side (+v) ferrous peak. In the middle spectrum of the sample of siderite and goethite, it can be recognized that there must be a contribution of ferric-caused (from goethite) quadrupole splitting that renders the prominent left side (−v) feature asymmetric in profile, and partly fills in the range between both prominent peaks. Interestingly, these cases are reversed in regard to the dominant and additive contributions. Attention is further called to the “plateau” upon which the ferrous and ferric peaks reside, indicating the presence of something in addition to the siderite and goethite. While the discerning eye reveals much upon inspection, the deconvolution process renders apparent all individual spectral contributions and, again very importantly, establishes the precise values for IS and QS necessary to define what the responsible species are.
 Going to the bottom two spectra in Figure 8 and examining them just in the vicinity of zero velocity, one can see that ferric-induced quadrupole splitting is present for the basalt (second from bottom), although not to the degree in millimters per second exhibited by jarosite. The impact melt rock (Manicouagan impact structure, Canada) has in its spectrum a prominent peak at ∼0 mm s−1. This is indicative of ferric clay minerals or nanophase iron oxide, an interesting material, the spectroscopy of which and its relevance to Mars will be described later.
 Because of the interest in knowing the relative abundances of the ferrous and ferric components of a target sample, estimates can be made from the relative areas of the deconvoluted spectral signatures for these states. These areas are related to the abundance of the species as well as the strengths of their Mössbauer resonance, which is described in terms of the recoil-free fraction. Measurements of the resonance strengths of materials of interest for Mars are required for precise assessment, which if not already accomplished can be a postmission activity.
 What remains to be considered regarding the spectra displayed, as examples in Figure 8 are the sextet patterns very apparent in three of the five shown. As previously explained the sextet of peaks is a manifestation of the nuclear Zeeman effect induced by a magnetic field, namely the hyperfine magnetic field Bhf, which in these cases is intrinsic to magnetic ordered species being present. For transmission measurements which usually involve powdered material dispersed in a transparent medium, the intensities of the sextet peaks are commonly in the ratio of 3:2:1:1:2:3 reflecting absence of anisotropy or texture because of the sample being randomized during preparation. This does not have to be the case for a sample in the natural state resulting in a distribution of intensities different from that just stated. In addition, quadrupole shift perturbation can also result in reordering of ratios. In the case of backscatter spectroscopy, and also for thick absorbers in transmission, the intensities of the features of a sextet can deviate from simple ratios. Both situations will be discussed.
 The “quadrupole shift” refers to a small change in energy of the four excited Zeeman split levels. The origin of this shift is identical to that of the quadrupole splitting defined earlier for the case of the doublet spectrum. Determining the value of quadrupole shift perturbing the sextet pattern is a simple procedure. You start by labeling the sextet peaks in order from −v to +v as 1, 2, 3, 4, 5, and 6 corresponding to the transitions indicated in Figure 9 where quadrupole splitting and nuclear Zeeman effect are combined in an example. Since the quadrupole splitting is precisely twice the quadrupole shift, the QS value is typically used to refer to both, so that the quadrupole shift has a value of QS/2, as indicated in Figure 9.
 The value for the effective quadrupole splitting QS, for the magnetically split case, is calculated by
where Δm,n is the difference in millimeters per second between peaks m and n, namely
Referring to Figure 9, which illustrates the case for QS < 0, the effect of quadrupole interaction on the sextet is to shift the inner four peaks of the sextet toward the right (in the +v direction) and the outer two peaks toward the left (in the −v direction) by the same amount. There are some cases such as magnetic ordered siderite at T < 35 K which is an example not expected on Mars, where the quadrupole interaction may be so great as to reorder the peaks so intensities from left (−v) to right (+v) are in the ratios of 2:3:1:1:2:3. If this kind of situation were to be encountered it would be recognized by inspection.
 To determine the Mössbauer parameter Bhf, the separation (in millimeters per second) of the sextet between the dominant intensity peaks is measured. When the quadrupole interaction is small, these are the two outer peaks of the sextet (3, 3) as shown in Figure 6. However, when the contribution of quadrupole splitting is great, the appropriate peak pair is selected on the basis of having the dominant intensities relative to the others. For the case of the outermost peaks being the dominant pair (3, 3) because of a small quadrupole interaction, the value for Bhf is calculated by
where in this case Δ3,3 is the millemeters per second difference between the two dominant peaks of the sextet, and Bhf is expressed in tesla. Reexamining the central spectrum of Figure 8, the presence of a poorly defined sextet should now be evident to the reader. It is not as spread out as much in velocity range as the three distinct examples of sextets, indicating a lower value for Bhf. Deconvolution using Mössbauer fitting algorithms would reveal the sextet to be that of goethite based upon the determined Bhf. It is interesting to note that this assignment was only made in concert with Pancam filter photometry and was not evident when the sample was also subjected to thermal emission spectroscopy (MiniTES) and Raman spectroscopy as well [Morris et al., 1998]. This is an example of why a suite of instruments is such a necessity for in-situ investigations as will be done with MER.
 It is instructive to examine examples of two more backscatter spectra, the first of which is for a slab of magnetite from the UAB Planetary Materials Laboratory collection (Figure 10). The data point pattern clearly indicates a near superposition of two sextets, which is more evident in the −v range and clearly brought into view by deconvolution. The presence of two sextets for Fe3O4 shows the resident 57Fe nuclei are in either of two hyperfine field sites. The second spectrum (Figure 11) is of the external Composition Calibration Target (CCT) previously discussed in section 3 and shows not only the sextet spectrum of magnetite but a doublet as well which on the basis of a small degree of quadrupole splitting would be caused by the presence of a ferric (Fe3+) constituent in addition to the magnetite.
Figure 12 displays a backscatter spectrum of specular hematite, a magnetic ordered mineral of interest for Mars, providing an example of how the intensities of sextet features can behave differently in backscatter measurements from the 3:2:1:1:2:3 ratio exhibited commonly in transmission spectroscopy. Inspection clearly shows the two component pairs of the sextet (1 and 2, and 5 and 6) are of comparable intensities. This is different from the example of magnetite displayed previously in Figure 11 (top), where both sextets under casual inspection seem to appear to display ratios commonly seen in transmission measurements. What is demonstrated by these two spectra is the role of directional anisotropy because there is an angular dependence involved in nuclear transition probabilities [Hawthorne, 1988]. The lesson of Figures 11 and 12 is that when considering intensity ratios in backscatter spectroscopy, there are issues in the realm of radiative transfer that must be considered. This includes acquiring a full understanding of backscatter spectroscopy as an in situ technique, something that demands more laboratory investigation of a broad range of natural and laboratory-produced examples coupled with theoretical modeling. This need is largely the consequence of the backscatter technique having been previously an esoteric technique compared to transmission studies which dominate the body of the literature.
 Finally, in Figure 13 are displayed the transmission spectra obtained at different temperatures of the combined α-Fe and α-Fe2O3 internal velocity calibration reference material incorporated into the MER Mössbauer spectrometer sensor head. Such spectra will be available for monitoring the condition of the instrument and in particular for establishing the actual velocity profile of the transducer. As a tutorial, the spectra show the data as it is obtained in the back and forth stroke cycle of the transducer as a function of channel number. This is the way data appears in a “mirror image” fashion prior to it being velocity calibrated and then being folded over. Both α-Fe (foil) and α-Fe2O3 are well understood materials from the standpoint of Mössbauer characterization and thus serve to calibrate the channel/velocity model in a precise manner making possible accurate determination of IS, QS, and Bhf for Martian samples, and hence assignment of mineralogy on the basis of the Mössbauer parameters. Figure 13 will be referred to again in section 5.
5. Exploitation of the Martian Diurnal Temperature Variation
 Mössbauer spectra exhibit variation with sample temperature making it an important experimental variable. The long history of laboratory experience, albeit in the transmission mode, is readily translated for use in assessing MER Mössbauer data and is suggestive of opportunities for gaining insight into the physical events by which the mineralogy came into existence. This, of course, is of relevance to the understanding of ancient planetary history including aspects that bear on the questions regarding life.
 The character of the Martian diurnal temperature variation is an important experimental variable and will impact the planning for acquiring temperature correlated data. The Pathfinder and Viking missions represent the three in situ experiences available for temperature measurements, consequently establishing the character for the MER landing sites early on and predicting future trends in temperature change will be a significant activity in preparation for carrying out Mössbauer spectroscopy for the duration of surface operations. Figure 14 adapted from the Pathfinder site atmospheric temperature measurements of Schofield et al.  displays the temperature variation, and isolates two distinct periods during which good quality data can be acquired under relatively isothermal conditions that are the extremes. This does not preclude obtaining data during other portions of the diurnal variation but represents the optimum situation. Meteorological results from Vikings 1 and 2 [Hess et al., 1977] show similar characters although being different in terms of temperature extremes. It should be pointed out that temperature measurements are automatically part of the instruments data stream.
 While the rovers will be immobile during the temperature minimum when acquisition of Mössbauer data is a highly likely activity, utilization of the upper temperature extreme will require justification and planning because it is during the likely period of mobility. Also, acquisition of data on the same target sample at the two temperature extremes (and in between) means not only is the rover immobile during most of the sol, but the Instrument Deployment Device (IDD) must either remain in fixed position or be capable to return the sensor head to the previous position in an assured manner. Because there is no priority for order of measurements at different temperatures, there is possible flexibility in planning the activity.
 Quadrupole splitting (QS) and the magnetic hyperfine field (Bhf) can vary with temperature in a continuous manner well within the measurement capabilities of the MER instrument. Magnetic transitions, should these occur within the range of diurnal temperature variation, will display more abrupt spectral change as magnetic ordering occurs, ceases, or changes character at specific temperatures. When normally crystalline materials capable of magnetic order below a transition temperature (Curie or Néel point) are instead found in the nanophase form (in the size range of less than ∼30 nm), there is a gradual spectral change from a nonsextet pattern to sextet character with decreasing temperature, a situation that if encountered on Mars will be of considerable interest.
Figure 13, previously discussed in terms of the material serving as a standard for the internal velocity calibration of the instrument, is also instructive in showing how a temperature caused magnetic transition results in a distinct spectral difference under a minimal change in temperature. The species component of the calibration standard that is involved is α-Fe2O3, and the temperatures for which it occurs are both of interest for Mars. Under inspection, it can be seen that in either of the sides of the mirror images displayed channel-wise for the four spectra are more than six peaks. This immediately alerts to the presence of a blend of spectra. The −17°C spectrum has 10 discernable peak,s while the spectra obtained for lower sample temperatures show 11. Both situations are a consequence of blending of individual sextets of α-Fe (foil) and α-Fe2O3. The effect of the transition is very evident in the figure, between the spectra associated with −17°C (256 K) and −34°C (239 K), particularly for the extreme left-hand peak on the left side of the mirror images as well as on the opposite side for the extreme right feature. Further inspection reveals how other peaks in the sextet of the species undergoing the transition, in this case α-Fe2O3, are affected. At first glance, because of the complexity of the spectra, it might appear that the hyperfine field (Bhf) increases at the lower temperatures relative to that of −17°C; however, this is not the whole story. There is a change in Bhf, but it is modest. What happens as well is a pronounced shifting of the positions of the α-Fe2O3 peaks of the sextet. The inner four peaks shift toward −v with respect to the outer pair. This is an example of how deconvolution and extracting individual spectra for respective component species is important in the analysis. The example is also instructive in that the other constituent species (α-Fe foil) does not show such abrupt behavior in the temperature range. Encounter of mixed constituents is a likely “real world” scenario for Mars.
 The temperature at which α-Fe2O3 undergoes the transition is the Morin transition around 260 K [Morin, 1950] making its observation likely, should this species be a constituent of a target sample. In the case of hematite, particularly just near the Morin transition, the reduction in Bhf that results from an absence of a Morin transition is typically ∼0.8 T. The temperature at which the transition occurs is dependent upon the specific nature of the hematite, including particle size and the presence of impurities, and can be completely suppressed in certain cases. On Mars it will probably be a matter of experiencing a sufficiently high sample temperature during the diurnal variation (a heat of the sol situation) that is also coincident with Mössbauer interrogation. Other substances exhibit changes between magnetic states such as the Verwey transition of magnetite (Fe3O4), which because it occurs between 119 and 116 K [Sawatzky et al., 1969] will not be observed should that material be encountered.
 While opportunity to observe abrupt spectral changes with respect to narrow differences in temperature is probably limited to Fe2O3, gradual changes for other species can be contemplated for Mars. Quadrupole splitting (QS) and the magnetic hyperfine field (Bhf) can vary with temperature in a continuous manner, the detection of which is well within the measurement capabilities of the MER instrument and can be expected for the Martian diurnal range should certain species be present. Siderite (FeCO3) is an interesting example of discernable variation with temperature. The general functional character in this regard is displayed in Figure 15 as a fit to data adapted from Wade et al. . Superimposed on the functional relationship of quadrupole splitting with temperature, is a box outlining the measurement domain expected for Mars. Within the box the variation in measurement is about ±0.01 mms−1. Because the QS of a species such as this can vary considerably, knowing the temperature of the target sample using the sensor incorporated into the contact plate is essential for interpreting the determined QS parameter as an identifier.
Figure 16, adapted from Mørup et al. , shows how the measured Bhf for goethite varies in a continuous manner with temperature including over the range between expected Martian temperature extremes of ∼260 and ∼180 K as shown by the superimposed box outlining the measurement domain expected for Mars. The Bhf for 180 K is ∼14% greater than for 260 K. Again knowing the sample temperature using the contact plate is important, and knowing how QS and Bhf vary with temperature has to draw upon laboratory studies previously carried out and new ones in response to stimulation from the MER missions. Establishment of Mössbauer parameters in the laboratory can be accomplished by both transmission and backscatter spectroscopy efforts (Figure 17).
 In general, an iron containing material capable of magnetic order will exhibit a sextet spectrum at temperatures below the critical temperature (Curie or Néel point) and otherwise above it. For crystalline materials that can be considered of interest for Mars, such critical temperatures are outside of the expected range of the extremes of diurnal variation which are ∼260 and ∼180 K. Both hematite (Tcrit ≈ 960 K) and siderite (Tcrit ≈ 35 K) are such examples. There can be a situation when the spectrum of a material below the critical temperature does not exhibit a sextet pattern as would be normally expected, and further more shows a gradual shift in this nonsextet pattern to a sextet with decrease in temperature (the opposite occurring with increasing temperature). This condition has implications for understanding the processing of the mineralogy that brought it about.
 The adopted terminology for such a material is “superparamagnetic,” and the respective spectral signatures are referred to as paramagnetic and magnetic. It is a consequence of the material being in what has been described as the nanophase form, meaning particle sizes less than ∼30 nm which for many materials is less than the size scale of magnetic domains. The term nanocrystalline has also been used in the literature. Often such particles are in a state of uniform magnetization with a net magnetic moment possible 105 times that of a single atom (hence “superparamagnetic”). Exceptions include antiferromagnets, where superparamagnetic relaxation is very common and there is essentially no net magnetic moment. For sufficiently small particles at higher temperatures, the relaxation time of the magnetic moments of these particles is smaller than the nuclear Larmor precession time (∼10 ns) which is the timescale over which the 57Fe nucleus senses the hyperfine magnetic field, and instead of nuclear Zeeman splitting resulting in a sextet, there is a collapse to the spectrum characteristic of the paramagnetic state [see Mørup et al., 1983]. This condition has been demonstrated for extraterrestrial material, initially by Wdowiak and Agresti  for the Orgueil carbonaceous chondrite. The physics is very much in the realm of statistical mechanics in that thermal and size distribution functions are involved.
 Mössbauer spectroscopy has revealed the superparamagnetic spectral signature for a distinct iron-rich narrow layer of the terrestrial Cretaceous-Tertiary (KT) boundary clay demonstrating its nanophase nature [Wdowiak et al., 2001]. From the tutorial standpoint, spectra obtained over a range of specific temperatures from 300 K to as low as 12 K for samples collected at North American and European sites all show to varying degrees the gradual change from nonsextet to sextet character with decrease in temperature. The phenomenon is independent of specific iron mineralogy, which seems to be locality specific. Figure 1 displays this phenomenon for samples collected at El Mimbral II, Mexico; Bochil, Mexico; and Moscow Landing, Alabama. The array of El Mimbral spectra, which provided the initial recognition, is so remarkable that it can serve as a textbook example of the superparamagnetic signature! Ninety-nine percent of the resident iron is incorporated into the nanophase state for this sample, which was examined as collected with no treatment other than dispersal into the molten wax that was solidified into a disk for transmission spectroscopy. The displayed examples show considerable spectral dynamic within the expected extremes of the Martian diurnal temperature variation, however nanoparticle distributions skewed toward very small particle sizes might not demonstrate such a dynamic until temperatures are below ∼180 K depending on the identity of the species. Interesting effects for goethite are possible at Martian temperatures because of different particle sizes and different degrees of interaction among nanoparticles.
 We have put forth the hypothesis that this distinct globally distributed component of the KT boundary layer is an exoatmospheric condensate of the vaporized material in the impact fireball that plumes upward perforating the atmosphere. This has implication regarding impacts over Martian history in that just like for the KT event there should be remnants of nanophase nature for impacts on that planet. Nanophase material should persist for very long periods of time although local chemistry can alter the mineralogy from the original form, which is suggested by what we have found at the KT boundary. What remains for consideration is how differences between Martian and terrestrial tectonics and climatic conditions affect retention of a nanophase record of impact events. All of this bodes for an interesting opportunity during the MER missions and beyond.
 Production of nanophase material can occur in more than one manner including by weathering. The oxidized basalt spectrum of the montage in Figure 8 can be interpreted to infer that the ferric doublet previously mentioned is that of a nanophase component resulting from weathering although rapid quenching from a melt might be a responsible mechanism [Helgason et al., 1976]. The lesson learned is that spectral interpretation should always take place in the context of other kinds of information. A number of investigators [Geissler et al., 1993; Banin et al., 1997; Mustard and Hays, 1997] have suggested that the elemental composition of soils at the Viking and Pathfinder sites can be explained as the byproduct of weathering of mafic volcanic rocks in the presence of acidic volatiles produced during episodes of volcanism. Furthermore, lack of surface water and the low-temperature environment would have the effect of inhibiting transformation of metastable amorphous mineraloids to crystalline phases. Aside from the 65 million year old KT material, there is meteoritic evidence for the survival of nanophase material for a period in excess of 4 × 109 years [Wdowiak and Agresti, 1984].
 Hydrothermal spring systems in the terrestrial setting also yield iron-rich nanophase materials, which is of interest should such systems have once been active on Mars. This would be a localized mechanism, and the areas could have served as habitats for primitive life. Using Mössbauer spectroscopy, we have investigated the products of hydrothermal vent systems with that aspect in mind [Wade et al., 1999]. Drawn from that work are two examples demonstrating hydrothermal system nanophase products that are displayed in Figure 18. The left side shows the result of a temperature study of material collected at Obsidian Pool, Yellowstone National Park. Quite evident is the signature of superparamagnetism with much of the dynamic occurring within a temperature range consistent with Martian diurnal variation. The dominant constituent is goethite.
 The second example (right side of Figure 18) is for material collected at Chocolate Pots, Yellowstone National Park, and again shows a dynamic evident within a temperature range expected for Mars although having a less dramatic growth of the sextet than what was exhibited for the Obsidian Pool material. On the other hand in the 300–200 K range the individual sextet peaks are sharper than for the Obsidian Pool sample. The Chocolate Pots sample contains siderite, nontronite, goethite, and hematite, all verified by X-ray diffraction [Wade et al., 1999].
 Finally, for this section the Martian bright region spectrum in the 300–2500 nm range, originally measured by Singer et al.  coupled with laboratory investigation [Morris et al., 1989, 2000, 2001] argues for a nanophase material being a component of Martian surface material including as dust that would become a suspended aerosol. Consequently the MER Magnetic Properties Experiment and the ability to interrogate collected material with the Mössbauer spectrometer would seem to have interesting prospects.
6. Utilization of 6.4 and 14.4 KeV Channel Spectroscopy to Acquire Information on Structure With Depth
 Because both 14.4 keV resonantly scattered gamma radiation and 6.4 keV internal conversion X radiation arise from the Mössbauer event, there is opportunity for accessing information that is translatable to how the structure target sample structure varies with depth. Since the count signals of these radiations are processed separately via two instrument channels, no additional provisions are required to acquire the raw data for later interpretation. Key to the methodology is the behavior of the atomic absorption coefficients (expressed as area per mass (cm2 g−1)) for 6.4 and 14.4 keV radiations as a function of atomic number (Z) of the species resident in the target sample. Figure 19 illustrates this relationship and upon inspection reveals that while absorption coefficients for both energies are somewhat similar in the range 25 ≤ Z ≤ 36, there is substantial difference at other atomic numbers which includes elements such as silicon, oxygen, magnesium, calcium, aluminum, etc.
 Typically, for minerals the 6.4 keV X rays will emanate from the near surface regions in contrast to the higher-energy gamma rays which come not only from the near surface but also significantly deeper within the sample. Generally, if the target sample is homogenous with respect to depth, the deconvolution will show the same component spectra with similar relative area. When there is structure in composition with depth, the contribution to the 6.4 keV channel spectrum from near surface species will be enhanced relative to that of the 14.4 keV channel. This is a consequence of what is displayed in Figure 19.
 For initial assessment, inspection of the spectra of the two channels can reveal whether the sample varies with depth or is homogeneous. Should the sample appear to vary with depth, based upon the degree of difference in the contribution of species to the two spectra, one can judge which species are more near surface abundant and which are more prevalent at depth. To go beyond this, one or two approaches can be brought to bear. The first is to prepare laboratory model structures, analogous to “phantoms” used in radiography, and examine them by carrying out spectroscopy with the same geometry of the MER instrument. The other possible methodology is to model the radiative transfer process involved in the backscatter technique using the MER instrument geometry. This will require an experimental understanding of the pertinent properties of candidate materials. Ideally, both methodologies should be employed and it is obvious that such activity from the practical standpoint would not occur simultaneously with surface operations. It must be stated that the body of literature is sparse in this area.
 TJW, MLW, and JIN are very grateful for David Agresti serving as a sounding board and providing useful commentary so the document would better serve its intended audience, and especially for his contribution in the UAB experimental activities. TJW acknowledges long-term support from the NASA Exobiology and Planetary Instrument Definition and Development Programs, and in particular support as a MER Athena Coinvestigator by Cornell University and the Jet Propulsion Laboratory. MLW and JIN acknowledge being welcomed into the LAPIS FIDO activity, and JIN is particularly grateful for support from the Athena Student Interns Program. GK acknowledges the support of the German Space Agency DLR and the Universities of Darmstadt and Mainz.