Part I Chemical Engineering Section 2 (ex-ET)

ANALYTICAL CHEMISTRY
8 lectures, Lent Term 2012
Prof. Clemens Kaminski

Course outline 1. What is Analytical Chemistry? 2. General features of molecular spectroscopy 3. Ultraviolet/visible spectroscopy 4. Infrared spectroscopy 5. Microwave spectroscopy 6. Nuclear magnetic resonance spectroscopy 7. Methods of elemental analysis 8. Mass spectrometry 9. Chromatography Text books These lecture notes contain all you need to know about analytical chemistry for examination purposes. You can find out more (if you want to) from almost any textbook with Physical Chemistry or Analytical Chemistry in the title. Examples paper: One examples paper will be issued to test understanding and aid exam preparation.

CFK/PJB

Analytical Chemistry: Page 1

What is Analytical Chemistry?

Analytical Chemistry is concerned with answering the questions: What chemical species are present in a sample?

How much of each chemical species is present?

Analytical Chemistry is vital in the following areas: Quality control in the process industries o of starting materials o of intermediates o of products confirmation of purity identification of impurities Environmental analysis o Monitoring and control of pollutants in streams that are to be released to the environment (in gas, liquid or solid form) o Measurement of pollutants in the environment (air/river/ground) NOx, SOx, hydrocarbons in atmosphere Organic chemicals (polychlorinated biphenyls, detergents) Toxic heavy metals (lead, cadmium, mercury) Clinical and biological studies o Measurement of nutrients, including trace metals o Measurement of naturally produced chemicals (cholesterol, sugar, urea) o Measurement of drug levels in body Geological assays o Measurement of metal concentrations in ores and minerals o Measurement of oil/gas concentrations in rocks Fundamental and applied research o Chemical engineering: how much conversion (or separation) do we obtain under these conditions? o Organic molecule synthesis: what compound have we made?

Analytical Chemistry is thus vital in the process industries and in research laboratories. Qualitative analysis is the identification of elements, functional groups, or particular compounds in a sample. Quantitative analysis is the determination of the amount of a particular element, species or compound in a sample.

CFK/PJB

Analytical Chemistry: Page 2

Analytical Chemists need to be good at careful accurate measurements, statistics and error analysis: samples of known concentration must often be prepared for calibration purposes samples must not become contaminated for environmental analysis, more than one measurement is often performed on more than one sample to draw conclusions. On a process plant, analytical chemistry is normally performed off-line: a sample of product is removed and sent to the lab for testing might take hours or days for plant control purposes, we may need to infer composition indirectly o e.g. from T and P measurements and a model of how conversion (or separation) varies with T and P numerous off-line analytical techniques. However, an increasing number of analytical techniques can now be performed on-

line:
sample the process stream in situ the plant can then be controlled using the direct composition measurement fewer techniques, and most will only work for certain reactions/products.

Classical (old-fashioned) analytical chemistry is based on techniques such as: Titration: volume of a standard reagent reacting with the sample is measured o Acid-base titrations: e.g. monitor the colour of a solution containing a pH-sensitive indicator as an acid (or base) is added. o Complexation titrations: e.g. monitor the pH of a solution whilst reagent EDTA, ethylenediaminetetraacetic acid (HOOCCH2)2NCH2CH2N(CH2COOH)2, is added: EDTA reacts in a 1:1 molar ratio with almost all metal cations (except alkali metals), enabling the metal cation concentration to be determined. Gravimetry: measurements based on mass. Simple examples are: o Mass lost on heating of a solid gives the amount of water of crystallisation. o Mass of precipitate formed during a reaction can be measured. For instance, adding excess silver nitrate solution to determine the concentration of chloride ions present. Electrochemical methods:

o pH measurement. o Ion-selective electrodes. Modern analytical chemistry is largely based on instrumental techniques. In this lecture course, we shall discuss: Molecular spectroscopy techniques: first part of course

CFK/PJB

Analytical Chemistry: Page 3

Other analytical techniques: last two lectures

CFK/PJB

Analytical Chemistry: Page 4

General features of molecular spectroscopy

Quantum mechanics tells us that all energy levels are quantised. We can probe the separation between energy levels by spectroscopy: o Test of quantum mechanics and our theories of bonding o Provides structural information Absorption and Emission The ground

2.1

state of a molecule is the one of lowest energy. An excited state of a molecule is one of higher energy.

Excitation refers to the process in which the molecule goes from a low to high energy state: it requires the addition of energy by photon absorption. Relaxation is the process by which a molecule falls from a high to low energy state: it involves the removal of energy by photon emission.

Whether the transition is permitted or not depends on: o The frequency of the photon: we require E = h o Selection rules: we require that the electromagnetic radiation interact with the molecule, and that angular momentum is conserved as well as energy. [Aside: photons have an intrinsic angular momentum] For example: an electron jumping between atomic orbitals has to obey: n = anything ; l = 1 ; ml = 0, 1 This means an electron in the 1s orbital of a hydrogen atom could move to 2p, 3p, or 4p orbitals by absorption of light of appropriate frequency, but not to 2s, 3s, 4s, or 3d, 4d atomic orbitals.

CFK/PJB

Analytical Chemistry: Page 5

2.2

Schematic diagram of an absorption spectrometer

The

monochromator causes light of just a single frequency of light to be

detected; it may be before or after the sample cell. The monochromator is adjusted (e.g. by rotation of the prism) so that the frequency of light reaching the detected is scanned. This simple diagram is sufficient for this course. However, in practice better methods have been developed than this basic set-up: o Tunable diode lasers may be used: in this case, the light source is monochromatic but its frequency can be tuned. o Fourier transform infrared (FTIR) spectrometers use an interferometer technique. o NMR spectrometers use a very short pulse of radiation containing a distribution of frequencies.

2.3

Factors affecting intensities of spectral lines

Transition Probability
o This depends on the precise quantum mechanical wavefunctions of the initial and final states (beyond the level of this course). Some transitions may have zero probability in that case, they are said to violate selection rules.

The Population of States

o The initial population of an energy level obviously affects spectral intensities. o At thermal equilibrium, the relative populations of two energy levels may be obtained from the Boltzmann factor: upper E Carrie r k is Boltzmanns constant, 1.38066 x 1023 J/K solven t Analytical Chemistry: Page 6

CFK/PJB

When E >> kT, then only the lower level has significant population When E << kT, then the energy levels will have the almost identical populations

Path length and concentration of sample

o This is summarised by the Beer-Lambert law: log10 (I / I0) = [conc] L Transmittance = I / I0 Absorbance = log10 (I / I0) [formerly known as optical density, O.D.] o Hence an alternative expression of the Beer-Lambert law is: Absorbance = [conc] L o The constant of proportionality is termed the molar absorption coefficient [former term is extinction coefficient]. has units; if theyre not quoted, assume that they are in mol1 dm3 cm1

2.4

Spectroscopic units
A variety of different units are commonly used in spectroscopy to represent the energy difference between the levels. You need to be familiar with: o J (Joules), the SI unit of energy o (frequency, often expressed in MHz); related by E = h o (wavelength, often expressed in nm), related by E = hc / . o 1/ or (reciprocal wavelength, or wavenumbers, often expressed in cm1); related by E = h c (1/ ). o eV (electron volts), related by 1 eV = 1.602 x 1019 J (i.e. the charge of an electron)

CFK/PJB

Analytical Chemistry: Page 7

2.5

Regions of the electromagnetic spectrum

Depending on the wavelength used a variety of different structural may be obtained.

information

CFK/PJB

Analytical Chemistry: Page 8

Ultraviolet/visible (UV/vis) spectroscopy

Absorption in the UV/visible region is associated with transitions between electronic energy levels. o Colours of compounds/solutions arise in this way. The transition of interest is normally that between the highest occupied

molecular orbital (HOMO) and the lowest unoccupied molecular orbital

(LUMO). o Other transitions involve greater energy separations, and so are further away from the visible region; band overlap for transitions at higher energies tends to result in uninformative spectra. o Wavelengths below ~200 nm cannot easily be studied for instrumental reasons the sample cell window absorbs radiation at these wavelengths. Energy molecular orbitals

LUMO HOMO E for transition of interest

The UV spectrum is normally plotted as the absorbance against wavelength; peak positions are identified by quoting max and . Typical UV spectrum:

CFK/PJB

Analytical Chemistry: Page 9

(from http://www.cem.msu.edu/~reusch/VirtualText/Spectrpy/UV-Vis/spectrum.htm#uv3)

Aside: due to simultaneous vibrational and rotational transitions (see Sections 3 and 4), UV spectra normally consist of fairly broad peaks. The absorbing groups in a molecule are called chromophores. Two isolated chromophores in a molecule give roughly independent absorptions for each one: o e.g. CH3CH2CNS: max= 245 nm and = 800 SNCCH2CH2CH2CNS: max= 247 nm and = 2000

In organic chemistry, -conjugated systems (when multiple bonds are separated by a single bond) tend to give particularly informative spectra. o Overlap of adjacent orbitals results in a decrease in the energy gap between the occupied orbital and the unoccupied * antibonding orbital. o This results in an increase in absorption wavelength (even into the visible region for greatly conjugated systems e.g. organic dyes), and normally an increase in the intensity as well. General rule: increased conjugation increases max and . o Aromatic systems exhibit conjugation, but tend to give complex spectra, frequently with more than one absorption band. Conjugation with lone pairs (n- conjugation) can also result in spectral transitions, though these are much weaker than those originating from overlap of orbitals.

CFK/PJB

Analytical Chemistry: Page 10

Example UV spectrum results:

CFK/PJB

Analytical Chemistry: Page 11

3.1

Uses of UV/vis spectroscopy Sample is usually liquid. Can follow changes in colour/composition quite rapidly (timescale may be down to ~1 s). Can measure concentration of any coloured compound, or any compound that absorbs in the UV region. Beer-Lambert law is useful, though for accurate work absorbance will be measured on solutions of known concentration for calibration purposes. Reasonably straightforward to do the measurement on-line: o Can study the process fluid through a glass window, or using a fibre optic cable. o In practice, a technique called attenuated total reflectance is likely to be used if the sample absorbs strongly. This technique is limited: o It doesnt work if the sample doesnt absorb in the UV/vis region! o Its not good if the sample contains several species that absorb in the UV/vis region: the absorption bands are broad and so overlap too much.

CFK/PJB

Analytical Chemistry: Page 12

Infrared spectroscopy
Absorption in the infrared region is associated with transitions between vibrational energy levels. Diatomic molecules: ideal case Let us consider a diatomic molecule first (such as HCl) as it contains just a single bond. Imagine the bond behaves like a perfect spring with a force constant k m1 spring constant = k m2

4.1

This is often called the simple

harmonic oscillator (SHO)

approximation. We can solve the Schrdinger equation exactly in this case to derive the energy levels of the molecule The vibrational energy levels are characterised by a quantum number v. o E = h 0 (v + ) v = 0, 1, 2, o spring] is the reduced [Note that this corresponds to the vibration frequency of the

mass, given by

Energy

vibrational energy levels v=4 v=3 v=2 v=1 v=0 E =0 E = 9/2 h 0 E = 7/2 h 0 E = 5/2 h 0 E = 3/2 h 0 E = 1/2 h 0

At room temperature, E significantly populated.

>> kT, implying that only the v = 0 quantum level is

CFK/PJB

Analytical Chemistry: Page 13

The first selection rule is that v

= 1

o This means that for a diatomic molecule we expect to see a single absorption peak corresponding to E = h 0 o The peak tells us 0, the bond vibration frequency, from which we get information about k and/or . The second selection rule is that the bond has to have a permanent dipole

moment:
o Variation of the molecules dipole moment upon vibration is needed to interact with the oscillating electric vector of the electromagnetic radiation. o This means that diatomic molecules such as O2 and N2 wont show any absorption in the infrared region (and so arent greenhouse gases); a molecule such as HCl will show absorption in the infrared region. 4.2 Diatomic molecules: real case Real bonds do not behave as ideal springs; they behave as anharmonic oscillators. Potential energy diagram:

The simple harmonic oscillator SHO model discussed earlier provides a good approximation to the real case at the lowest energy level when r is always close to requilibrium.

CFK/PJB

Analytical Chemistry: Page 14

For real molecules, the bond will break at high energies before v . o In practice this means that the vibrational energy levels get closer together as quantum number v goes up. Selection rules for real diatomic molecules: o v = 1 (as for SHO), but v = 2, 3, are now weakly allowed as well (transition probability is small but is now greater than zero). o Molecule still needs to have a dipole moment for interaction with photon to occur. Its still the case that only the v = 0 level is significantly populated at room temperature. Spectrum will thus show an absorption band at 0, plus a weak band at ~20 and possibly ~30 Energy vibrational energy levels v=4 v=3 v=2 v=1 v=0 E =0 E = ~ 7/2 h 0 E = ~ 5/2 h 0 E = ~ 3/2 h 0 E = 1/2 h 0

Note that the ground state of the molecule (i.e. the state of lowest energy) doesnt have E = 0. Two consequences of this are: o Even at a temperature of absolute zero, bonds will still have a non-zero vibrational energy that depends on k and : The molecule is said to possess zero-point energy. Atoms still move by vibrations at absolute zero; whilst seen here from the maths, its a consequence of the Heisenberg uncertainty principle. o Consider bonds involving different isotopes, e.g. compare OH and OD bonds: They involve the same number of electrons, and so have identical force constants and bond lengths. However, they have different zero-point energies because of different . They will have different bond dissociation energies, as this is the energy required to take the bond from its zero-point energy up to an energy corresponding to the atoms being widely separated.

CFK/PJB

Analytical Chemistry: Page 15

CFK/PJB

Analytical Chemistry: Page 16

4.3

Polyatomic molecules

Polyatomic molecules have more than one vibrational mode.

o The location of all N atoms in a molecule needs 3N parameters to specify it. These are normally specified by: Translational position: specify the centre of gravity of the molecule using 3 parameters. Rotational motion: requires 2 parameters for a linear molecule; 3 parameters for a non-linear molecule. Vibrational modes: there will thus be 3N5 of these for a linear molecule, and 3N6 of these for a non-linear molecule. The vibrational modes in polyatomic molecules may involve movement of all the atoms, rather than just a single bond vibration. Each vibrational mode will have its own frequency. For example, the three vibrational modes of the water molecule are:
Symmetic stretch Asymmetric stretch Bending mode

1 = 3655 cm H O H

2 = 1595 cm H O H

3 = 3755 cm1 H O H

Each vibrational mode can be considered individually. We therefore expect absorptions at frequencies corresponding to 1, 2, 3, providing the dipole moment of the molecule is changing during the vibration (as was the case for diatomic molecules). o All three vibrational modes of H2O are IR-active o The symmetric stretch of CO2 is IR-inactive as it doesnt change the dipole moment); the other vibrational modes of CO2 are IR-active o Note that H2O and CO2 are greenhouse gases because they absorb in the infrared region. Overtone and combination bands (e.g. 21, 1+3) are very weakly allowed (as was the case for real diatomic molecules).

CFK/PJB

Analytical Chemistry: Page 17

4.4

Uses of IR spectroscopy The frequencies of some vibrational modes are almost independent

of the

structure of the compound in which they are located.

o For instance, OH stretching vibrations occur at 3200-3600 cm1 Detecting an absorption band in this frequency range is evidence that the sample contains OH functional groups [Aside: its quite a broad range in this case because OH bond strengths are affected by the extent of hydrogen bonding to them] We can thus use IR spectroscopy to identify structural groups present in the sample:

From our discussion above, we note that: o Bonds involving lighter atoms

absorb at higher frequency

than bonds involving heavier atoms: CH > CC OH > OD o Stronger bonds absorb at higher frequencies: CC 2150 cm1 C=C 1650 cm1 o Bonds with large dipole moments give strong absorptions; those without give weak (or no) absorption: C=O strong C=C often weak Organic chemists used to be very good at knowing precise vibrational frequencies of different functional groups and how there were affected by substituents. o For instance, theyd know 1710 cm1 was likely to be a C=O group in a ketone, while recognising 1730 cm1 was more likely to be a C=O group of an aldehdye. However, structural identification is now almost always done by comparison of the spectrum obtained with one in a database: a fingerprint method of

identification.

CFK/PJB

Analytical Chemistry: Page 18

Example IR spectrum:

(from http://www.cem.msu.edu/~reusch/VirtualText/Spectrpy/InfraRed/infrared.htm#ir1)

The sample for IR spectroscopy may be solid/liquid/gas. For lab measurements, the sample is held in a special cell: o the windows shouldnt absorb above ~500 cm1 (e.g. cant use glass; KBr is quite common) o the path length is usually short (because absorbances tend to be strong). Beer-Lambert law can be used to estimate concentration; this is difficult in practice to do accurately due to scattered radiation affecting the baseline. Limitation: need to identify a band from the component of interest that isnt overlapping with bands from any other components that may be present; usually okay for simple mixtures. Rough cost estimate of basic spectrometer: 12k Timescale: Typically 10 s for a modern instrument, but it depends on sensitivity and resolution required. On-line measurement of process fluids is not straightforward: o Most of the materials used for cell windows arent resistant to chemicals (e.g. KBr dissolves in water). o Normal fibre optic cables absorb in the infrared region so we cant use them. o Special materials for fibre optic cables that dont absorb above, say, 800 cm1 are being developed; thus far they tend to be expensive and react with acid/alkali.

CFK/PJB

Analytical Chemistry: Page 19

Alternative on-line measurement technique: investigate the near-infrared (NIR) region instead (wavenumbers between 4500 and 12,000 cm1): o Fibre optic cables do exist that dont absorb in this region, so remote sensing is possible. o Well only see weak overtone and combination bands (e.g. 21 and 1 + 2) rather than fundamental vibrations; spectra are far harder to interpret. There tend to be lots of weak overlapping bands in this region, Calibrations involve running pure components and developing a mathematical model of the behaviour for mixtures very time consuming process o The NIR method is now used for continuous monitoring of some bulk chemicals in industry (e.g. gasoline; polymer melts), and is just beginning to be used in food and pharmaceutical industry.

Example: NIR spectra of C-H stretching overtone region for water-ethanol mixtures. (www.axsun.com)

New techniques are being devised for on-line process analysis one called Encoded Photometric Infrared (EPIR) spectroscopy has great potential, but its too early to say how useful it will be.

CFK/PJB

Analytical Chemistry: Page 20

4.5

Raman vibrational spectroscopy

Raman spectroscopy is another way of doing vibrational spectroscopy.

Let us consider what happens when a laser emitting fixed frequency light is fired at a sample. Most photons scattered by a sample will be at the same frequency as the incident light: o Photons interact with the sample. o During interaction, the molecule is raised from the v = 0 vibrational level to a socalled virtual state. o Usually the molecule then relaxes back down to the v = 0 level. o The scattered photon thus has the same frequency as before this is termed elastic scattering or Rayleigh scattering. However, a very small number of photons (~1 in 107) will be scattered at a different (usually lower) frequency than the incident light: o This effect is called the Raman effect. o During the photon interaction with the sample, the molecule in the virtual state may relax back to the v =1 vibrational state. o In this case, photons will have a scattered frequency of vlaser 0 where 0 is the vibration frequency. o Hence we can measure the vibrational frequency 0. o Other transitions may also occur (e.g. from initial v = 1 level to level v = 0 or v = 2). The selection rule for Raman spectroscopy is different to that for infrared vibrational spectroscopy: o Raman bands require there to be a change in polarizability of the molecule upon vibration; theres no need for there to be a changing dipole moment. o As a result, bands that are inactive in IR spectroscopy are normally active in Raman spectroscopy. o Similarly bands that are weak in IR spectroscopy (because they dont change dipole moment much) are usually strong in Raman spectroscopy.

CFK/PJB

Analytical Chemistry: Page 21

Example: IR and Raman spectrum of L-cystine illustrating the different selection rules (www.jascofrance.fr):

Another advantage: light at the laser frequency can be focussed on to a small area of sample easily Raman microscopy can record a vibrational spectrum on just a small part of the sample (down to ~1 m). Main limitations: o if the laser also promotes electrons into a higher energy level, then light will be emitted as the electron relaxes back down hence there may be interference from sample fluorescence o interference from background radiation o quantifying signal the Beer-Lambert law doesnt apply o its not good for complex mixtures Rough cost estimate of basic spectrometer: 15k ? Raman spectroscopy is beginning to be used for on-line process analysis: o the laser can be in the near-infrared region meaning that it can travel through glass windows, or be transported by fibre optic cables; the latter makes possible remote on-line sensing of the process fluid.

CFK/PJB

Analytical Chemistry: Page 22

Example: in-situ Raman spectra of the polymerisation of styrene (C6H5CH=CH2) in a batch reactor as a function of time (in minutes). (http://www.surrey.ac.uk/PRC/Facilities/raman.htm)

CFK/PJB

Analytical Chemistry: Page 23

Microwave Spectroscopy
This technique is not really useful in analytical chemistry, but is covered briefly here for completeness and because it can impact vibrational spectra. Absorption in the microwave region is associated with transitions between rotational energy levels. o This is how microwave ovens work. Pure rotational spectroscopy We shall only consider linear molecules in this section (e.g. diatomic molecules, or CO2). In the same way that we write down and solve the Schrdinger equation for vibrations we can also do so for pure rotational motion. For rigid linear molecules this gives energy levels characterised by the quantum numbers J and MJ o E = B J (J+1) J = 0, 1, 2, 3, o MJ = J, J1, J2, , J o [i.e. there are 2J+1 values of MJ] where I is the moment of inertia Note that I = r2 for a diatomic molecule, where is the reduced mass Energy rotational energy levels J=5 E = 30B 11 levels

5.1

J=4 J=3 J=2 J=1 J=0

E = 20B E = 12B E = 6B E = 2B E=0

9 levels 7 levels 5 levels 3 levels 1 level

Selection rules:
o J = 1 [to conserve angular momentum] o Molecule needs a permanent dipole moment (to interact with EMR) Population of levels depends on the degeneracy (number of levels having the same energy) and the Boltzmann factor:

CFK/PJB

Analytical Chemistry: Page 24

o Several of the lowest energy levels are occupied at room temperature o Indeed, differentiation of the above equation for N(J) with respect to J shows that the most populated J level corresponds to:

Hence we will observe a series of spectral lines if we do microwave spectroscopy, at energies corresponding to:

o E (J J+1) = 2B(J+1)
o Measurement enables parameter B to be determined o Extremely accurate method of measuring moment of inertias (and thus bond lengths) of gaseous molecules. Because the upper rotational levels are occupied, we can measure microwave emission spectra from remote objects: o Method for identifying molecules in planetary atmospheres o Method for estimated temperature of remote objects, as the intensity distribution of the lines in the spectrum depends on temperature. For non-rigid molecules (i.e. real ones!): o Analysis is similar to above, but it is found that bond lengths increase very slightly the faster the molecule rotate, meaning the apparent value of B decreases slightly as J goes up.

5.2

Infrared spectroscopy revisited: vibrational-rotational spectroscopy Vibrations and rotations can be treated as being independent of each other as they occur on different timescales. For a diatomic molecule, the selection rule is: v = 0, 1 (for SHO) and J = 1 and molecule needs a permanent dipole moment Thus pure rotational spectra may be observed, but pure vibrational spectra

are forbidden despite our discussion in Section 4!

o Each transition from vibrational level v = 0 to v = 1 has to be accompanied by a rotational change J = 1 in order to conserve angular momentum.

CFK/PJB

Analytical Chemistry: Page 25

o For instance, from the v = 0 J = 3 level we get: J=5

Energy

v=1 vibrational level

J=4 J=3 J=2 J=1 J=0

J=5

v=0 vibrational level

J=4 J=3 J=2 J=1 J=0

This means infrared transitions of diatomic molecules actually obey: o E = h0 2B(J+1) J = 0,1,2,3, (up to the last thermally populated J level) o We thus expect to get a series of peaks on either side of 0 For gas-phase samples, this so-called rotational fine structure is often observed when recording infrared spectra.

CFK/PJB

Analytical Chemistry: Page 26

Example 1: here is the background spectrum of air in an IR spectrometer. Note: o The baseline is not flat o Lots of rotational fine structure on the water vibrational modes o Some evidence of fine structure on the CO2 asymmetric stretch

Example 2: IR spectrum of gases in a plume from a volcanic eruption: remote measurement is possible using the sun as the source of infrared light. [Spectrum from Love et al., OSA topical conference, 1991]

For liquid samples, intermolecular collisions usually mean that the too broad to see rotational fine structure in infrared spectra.

linewidths are

CFK/PJB

Analytical Chemistry: Page 27

5.3

Energy diagram for a diatomic molecule

CFK/PJB

Analytical Chemistry: Page 28

Nuclear magnetic resonance (NMR) spectroscopy

This is usually the best method for the structural molecules. The sample normally needs to be a

identification of organic

liquid.

o Where possible, dissolve solids to make a solution before doing NMR spectroscopy. o Far more limited information can be obtained on solids (or gases). 6.1 Simplified background theory In the same way that an electron has angular momentum (described by spin), nuclei may also possess angular momentum and so be described as having spin. o The nuclear spin quantum number, I, gives the total angular momentum. The value of I for a particular nucleus depends on the detailed arrangement of protons and neutrons within the nucleus.
2 H has I = 1/2 H (0.015% natural abundance) has I = 1 12 13 o C has I = 0 C (1.1% natural abundance) has I = 1/2 17 o 16O has I = 0 O (0.04% natural abundance) has I = 5/2 The z-component of nuclear angular momentum is given by the quantum number mI, which can take values I, I1, ..., I. Ordinarily the mI quantum levels all have the same energy (i.e. are degenerate). However they will split into 2I+1 different energies if a large magnetic field B0 is applied. Transitions between these energy levels is termed nuclear magnetic resonance (NMR) spectroscopy. The selection rule is mI=1.

CFK/PJB

Analytical Chemistry: Page 29

Energy

Example: an I = 3/2 nucleus

mI = 3/2 E = h mI = 1/2 E = h mI = 1/2 E = h mI = 3/2 Increasing B0

4 degenerate levels when B0 = 0

We shall concentrate only on the case of nuclei having I = , the most important of which are 1H and 13C. o Nuclei with I = 0 will not show NMR spectra. o Nuclei with I > tend to give broad uninformative spectral lines.

Example: an I = 1/2 nucleus Energy mI = 1/2 E = (h/2) B0 (1) No magnetic field Sample in strong magnetic field B0 The separation of the two energy levels for a spin I = nucleus is given by: E = (h/2) B0 (1) o h = Plancks constant o = magnetogyric ratio of the nucleus (ratio of magnetic moment to angular momentum): For 1H: = 26.752 x 107 rad T1 s1 For 13C: = 6.7283 x 107 rad T1 s1 Analytical Chemistry: Page 30 mI = 1/2

CFK/PJB

o B0 is the strength of the applied magnetic field (SI unit = Tesla) o is a very small shielding term which will depend on the precise chemical environment of the nucleus under investigation. The separation between the energy levels is far smaller than the other techniques discussed so far. o We need to use large B0 values to get a measureable population difference between levels; it turns out NMR signal intensity is proportional to B02. o Modern NMR spectrometers in research laboratories employ superconducting magnets: Typical field strength is 4.7-9.4 Tesla, corresponding to a 1H resonance frequencies of 200-400 MHz. Highest available field strength is 21.1 Tesla, corresponding to a 1H resonance frequency of 900 MHz.

CFK/PJB

Analytical Chemistry: Page 31

The exact separation of energy levels for a nucleus in a molecule depends slightly on its chemical environment because of the shielding term: o Electrons immediately surrounding the nucleus will tend to circulate in a particular direction in the applied magnetic field. o This circulation induces a very small magnetic field B at the nucleus which opposes the large applied magnetic field. o The result is that the nucleus actually experiencing a magnetic field B0 (1) e B0 B

There are also some other effects that cause chemical environment to give a shielding term that we dont have time to discuss (e.g. hydrogens attached to aromatic rings have less shielding than might be expected). Hence nuclei in different chemical environments in the molecule will have peaks at different frequencies in the NMR spectrum. o Those with fewer electrons around them will have less shielding. The peaks in NMR spectra are usually quoted using the chemical shift

scale in parts per million (ppm).

o This is effectively a dimensionless frequency scale relative to a standard reference substance:

o The reference sample for both 1H and 13C NMR spectra is a compound known as TMS, tetramethylsilane, Si(CH3)4, o The chemical shift scale is used because it is independent of B0 value this is helpful because B0 is rarely known exactly as it changes slightly day by day for a superconducting magnet.

CFK/PJB

Analytical Chemistry: Page 32

6.2

H NMR spectroscopy

Peaks in 1H NMR spectra

are quantitative i.e. their area is proportional to

the number of hydrogens in that chemical environment in the molecule. Example: low resolution NMR spectrum of ethyl acetate (CH3COOCH2CH3) chemical shift, frequency, shielding,

High-resolution 1H NMR spectra show an important additional effect called spin-

spin coupling, or alternatively J-coupling.

o The quantum state of a 1H spin is slightly affected by the spin states of hydrogen nuclei that arent chemically equivalent to it, provided that they are within 3 bonds of it. Example 1:

o Consider nucleus HA: Spin HB may be in the same or the opposite direction to it. Hence HA sites in a sample have two slightly different energy states. o NMR signal from HA will therefore be a 1:1 doublet due to coupling to HB. o Similarly, the NMR signal from HB will be a 1:1 doublet due to coupling to HA. J /Hz J /Hz

/ppm CFK/PJB

HA

HB Analytical Chemistry: Page 33

Example 2:

o Nucleus HA can now couple to two chemically equivalent HB nuclei: Both HB spins may be in same direction as HA (probability ). Both HB spins may be in opposite direction as HA (probability ). One HB may be in same direction as HA, and one opposite (probability ). o NMR signal from HA will therefore be a 1:2:1 triplet due to coupling to HB. o Note that one HB nucleus does not couple to the other HB nucleus because they are in chemically equivalent environments. o The NMR signal from HB will therefore be a 1:1 doublet due to coupling to HA J /Hz J J

/ppm

HA

HB

Example 3:

o Nucleus HA now can couple to three chemically equivalent HB nuclei. o The resulting pattern for HA will be a 1:3:3:1 quartet (reflecting the probabilities of the different possible spin states), while the signal from HB will be a 1:1 doublet due to coupling to HA. J /Hz

J J J

/ppm

HA

HB

CFK/PJB

Analytical Chemistry: Page 34

Expert NMR spectroscopists are able to predict J values for different functional groups and molecular conformations. o Values are normally in the range 1-12 Hz for a 3 bond coupling. Example: high-resolution spectrum of ethyl acetate at a medium magnetic field (1H frequency 100 MHz)

7 Hz

100 Hz Example: high-resolution spectrum of ethyl acetate at a high (1H frequency 500 MHz)

magnetic field

7 Hz

500 Hz

Timescale for 1H NMR experiment is ~1 minute, but setting up may take more time depending on the experiment being performed. Each chemically distinct environment gives rise to a chemical shift (fixed in ppm), which may then split up due to J-coupling (fixed in Hz). Peak areas give relative amounts of each H environment.

CFK/PJB

Analytical Chemistry: Page 35

Functional groups can be identified (via chemical shifts and intensities). Linkages between functional groups can be identified (via J-coupling). NMR experts are good at identifying molecules from chemical shift and J-coupling patterns. Nowadays, identification of unknown molecules is often by comparison with: o Established databases of NMR spectra o The results of computer programs that predict NMR spectra. For very large molecules (even proteins!), there are huge numbers of spectral lines: o NMR techniques have been developed that allow assignment of these, e.g. 2dimensional and 3-dimensional spectra reduce overlap between peaks.

Example 1H NMR spectrum: vitamin K1 (C31H46O2) (1H frequency 400 MHz) (from http://riodb01.ibase.aist.go.jp/sdbs/)

CFK/PJB

Analytical Chemistry: Page 36

6.3

13

C NMR spectroscopy

The first use of 13C NMR spectroscopy is to give the number of chemically inequivalent carbon atoms in the sample. Note that symmetry considerations may mean that this is less than the number of carbon atoms in the molecule.

3 carbons (2:2:4)

5 carbons (2:2:2:1:1)

4 carbons (2:2:2:2)

The actual 13C chemical shifts provide structural information. For example: o C=O carbonyl 160-220 ppm o C=C alkene/aromatic 100-150 ppm o Saturated C alkyl 10-50 ppm As chemical shifts are affected by the shielding given by the surrounding electrons, substituents have an inductive effect. For example: o COH 50-80 ppm (compared to 10-50 ppm without OH group) Identification of structural groups is therefore possible: o By comparison with chemical shift values in reference books o By comparison with databases of 13C NMR data o By comparison with results of NMR prediction programs J-Coupling to other nuclear spins is possible: o Probability of 13C of interest being bonded to another 13C nucleus is small (as the natural abundance of 13C is only 1.1%) o Coupling to 13C of interest to any 1H bonded to it will happen this will result in a splitting up of each 13C peak according to how many hydrogens are directly bonded to the carbon. o J-coupling values in this case are ~120 Hz. o 13C multiplets due to coupling to 1H are:

CFK/PJB

Analytical Chemistry: Page 37

13

C NMR spectra take a long time to acquire if good signal-to-noise is desired (timescale is hours per spectrum). Its therefore common to record the spectrum using a technique called continuous o Almost all 13C NMR spectra are recorded this way o This eliminates the J-coupling to H nuclei, and so concentrates multiplet signal intensities into a single peak o The sample is irradiated with a broad bandwidth pulse whilst the 13C NMR spectrum is recorded such that all protons rapidly cycle between their spin up and spin down states, so that they become decoupled from 13C. o Coupling to protons other than that of H can still take place. o The result is a significant saving in the time it takes to record the spectrum. However, doing the experiment in this way means that the 13C NMR spectra dont have absolutely reliable relative intensities.

proton decoupling:

Example 13C NMR spectrum (proton decoupled): riboflavin (vitamin B2, C17H20N4O6) (from http://riodb01.ibase.aist.go.jp/sdbs/)

CFK/PJB

Analytical Chemistry: Page 38

6.4

Industrial on-line NMR spectroscopy The NMR techniques described above are those used in research and specialist analytical laboratories. They cant be implemented on a chemical plant: theyre too expensive and the hardware is not robust enough. Theres also a problem measuring NMR spectra of flowing samples the nuclei under investigation need to be in the magnet for a certain length of time before they can be measured by NMR. o Slow flow only, or need to stop the flow for the measurement. On a chemical plant, we can use permanent magnets of far lower magnetic field strength (corresponding to a 1H frequency of, say, 20 MHz) o Cant do high-resolution spectroscopy but can sometimes separate low-resolution signals: for instance, we can get a measurement of the ratio of aromatic to aliphatic 1H environments in gasoline. o 1H NMR signal intensity can be used to estimate the water content in solids. o 1H NMR relaxation times (how long spins take to move from the upper energy level to the lower energy level) of liquids in porous solids gives some information on the pore sizes present; this is used in oil-well logging. Magnetic resonance imaging (MRI) is also being developed for use in a process plant setting.

CFK/PJB

Analytical Chemistry: Page 39

Methods of elemental analysis

In this section, three principal methods of elemental

analysis are described.

They arent the only methods for doing elemental analysis: o For example, titration, gravimetry, and electrochemical measurements can also be used. CHN Analysis Organic compounds are normally analysed by flash

7.1

combustion of a small

sample (typically 1-2 mg) in oxygen unlike spectroscopic techniques, this is destructive test. An exact amount of sample is weighed inside a small tin capsule. The capsule is introduced into the analysers furnace which is at 950C. The tin capsule combusts, elevating the temperature to >1800C. At this temperature, the sample is vaporised and forms CO2, H2O and a mixture of N2, NO, NO2. CxHyNzO + O2 xCO2 + y/2 H2O + z NO2 + O2 The product gases are then analysed: o Instruments differ on whether they try to remove SOx, halogens, phosphorus etc. before analysing the product gases, or whether they try to measure them. o The product gases are normally reduced (using copper at high temperature) to remove O2 and convert NOx gases to N2 before analysis. o Some instruments separate the product gases: This can be by using chromatography (see Section 9), or by having a water and a CO2 trap (using Mg(ClO4)2 and NaOH respectively). A thermal conductivity detector is normally used to measure the gas concentrations. o A few instruments use infrared spectroscopy of the mixture as the detection technique. From the product gas concentration, the original mass fractions of C, H, and N in the sample are calculated. With modern CHN analysers: o Little attention is required by operator, other than a daily calibration of instrument. o Analysis is complete in 15 minutes. Oxygen contents of the sample can only be obtained indirectly using this method (e.g. by comparing the sample mass used, and that calculated for C+H+N components).

CFK/PJB

Analytical Chemistry: Page 40

o For a direct measurement of oxygen, pyrolysis (=heating in absence of air) in the presence of platinum-carbon can be used: this gives CO, which is then converted to CO2 and detected quantitatively. 7.2 Atomic spectroscopy

Atomic absorption spectroscopy (AAS) is a very sensitive method

for detecting the presence and concentration of about 70 elements. A sample of solution is vaporized to its constituent atoms in a hot flame. A hollow cathode lamp containing the element under investigation emits lights at specific frequencies for that element. The photons will be absorbed by the sample if the frequencies correspond to an allowed transition of atoms in the flame. o Recall selection rule for transitions of electrons between atomic orbitals is: n = anything ; l = 1 ; ml = 0, 1 The absorbance obeys the Beer-Lambert law to a good approximation, and so concentrations can be determined quantitatively. In practice the spectrometer will be calibrated on solutions of known concentration beforehand to achieve high accuracy. Atomic orbitals F L A M E

Light at (chosen for specific element of interest)

Measure absorption h

Solution of sample Advantages of AAS: o Sensitivity to a wide range of elements (typically down to 1 ppm) o High accuracy if care is taken over sample preparation and calibration. Disadvantages of AAS: o Some solid samples are difficult to get into solution form. o Need a hollow cathode lamp for sharp monochromatic lines for each element. o Different atoms require different flame temperatures to achieve reliable results (e.g. air/acetylene 2250C; NO/acetylene 2955C). Analytical Chemistry: Page 41

CFK/PJB

o Other factors influencing absorption may need to be taken into account: excited states of the atom under investigation (e.g. ionisation of Na/K) potential interferences from other elements present. Atomic emission spectroscopy (AES) is a similar process: Again a solution of sample is introduced into a hot flame. This time the intensity of a particular frequency emitted is measured. The advantage here is that there is no need for individual lamps for each element. The disadvantage is that it is far less sensitive than AAS. Inductively Coupled Plasma (ICP) spectrometers use a plasma o o o o (temperature >7000 K) rather than a flame: o Under these conditions, atomic emission spectra (ICP-AES) can be measured with similar sensitivity to AAS, while removing many of the chemical interferences present in a flame. o Mass spectrometry (ICP-MS) may be used to measure mass of atoms/ions present in the plasma directly using a quadrupole mass spectrometer [see Section 8].

CFK/PJB

Analytical Chemistry: Page 42

7.3

X-ray fluorescence (XRF)

This is a method of elemental analysis for solid samples. Its useful for those that dont easily dissolve, e.g. some minerals and ores. X-rays are fired at the sample, and these knock out a core electron from an inner quantum level (e.g. from the 1s, 2s or 2p atomic orbital). An electron then falls from an upper energy level to fill the core-level vacancy; this is accompanied by the emission of a photon of frequency h. The wavelengths of emitted photons enable the elements present in the sample to be determined, while the intensities give the amount of each element present (after careful calibration). This method is only appropriate for elements with atomic numbers greater

than ~20.

Gram quantities of a solid sample are required. h

4s4p 3d 3p 3s

2p 2s

Hole created by X-rays

1s

CFK/PJB

Analytical Chemistry: Page 43

Mass spectrometry (MS) Mass spectrometry is a widely used technique in analytical chemistry
o It can measure the molecular weight of components present in a sample o It can identify particular chemicals o After calibration, it can be used quantitatively. All mass spectrometers involve the following steps: o Production of ions in the gas phase o Separation of the ions according to their mass-to-charge (m/z) ratio. o

Detection of ions

There are several different ways of performing each step.

8.1

Production of ions: five methods Electron Ionisation (EI) is the most common method of ionisation for small organic molecules: o The sample must first be vaporized (by heat or a spark) if it isnt already in gas phase. Some sample decomposition may occur for thermally unstable samples during this step. o The sample is then bombarded with electrons that knock an electron out: M + e M+ + 2e o The parent ion is always produced in a vibrationally excited state and so might fragment (often in a fairly predictable manner) to smaller ions before it reaches the detector. o EI is a harsh ionisation method: fragmentation may be so extensive that the parent ion is absent from the spectrum. o The fragmentation pattern provides a fingerprint method of sample identification by comparison of results with MS databases. o Example: EI MS of vinyl chloride

[Figure from //www.cem.msu.edu/~reusch/VirtualText/Spectrpy/MassSpec/]

CFK/PJB

Analytical Chemistry: Page 44

Chemical Ionisation (CI) is a common alternative ionisation approach for

small organic molecules: o Again, the sample is first be vaporized (by heat or a spark) o A small amount of methane is present in the ionisation chamber, which is ionised by electron impact. o This then reacts to form species such as CH5+ that can donate H+ to the sample of interest: CH4 + e CH4+ + 2e CH4+ + CH4 CH5+ + CH3 M + CH5+ MH+ + CH4 o Ionisation of M to give MH+ by this method is thus by proton transfer. o This is a softer method of ionisation than EI, so less fragmentation occurs. o The parent ion [MH]+ is likely to be the most prominent peak; note that it will be at a molecular weight one greater than that of molecule M. o If the methane concentration is too high, then sometimes the species [MC2H9]+ is detected as well.

Fast-Atom-Bombardment (FAB) may be used to ionise medium-sized

organic molecules: o High-energy atoms hitting the sample can vaporise and ionise it in the same step Since no heating is required, this method can be used to analyse samples that arent thermally stable. o The parent ion [MH]+ is normally detected together with some useful fragmentation. o The method works for quite large molecules (up to about 5000 Da).

For very large macromolecules such as proteins, the above methods dont work well: o multiple fragmentation makes interpretation difficult/impossible. o its difficult to separate species with high m/z ratios (e.g. above 5000 Da).

Electrospray ionisation (ESI) has now become common for MS of

proteins: o A solution of the sample passes through a metal capillary that has a high applied voltage, and is then sprayed out to produce droplets (10 m) with a very high charge. o The droplets shrink as the solvent evaporates, so Coulombic repulsion between the charges increases until it causes each droplet to break up into smaller droplets.

CFK/PJB

Analytical Chemistry: Page 45

o Continuing the shrinking/breaking up process eventually leads to the molecule of interest being a lone ion it may be singly or multiply charged. o This is a very soft method of ionisation, and so negligible fragmentation occurs. o For proteins, the method produces a range of ions, e.g. from [MH10]10+ to [MH20]20+ o Example: ESI MS of myoglobin (mass 16955 Da).

[Figure from http://www.chm.bris.ac.uk/ms/theory/]

Matrix-Assisted Laser Desorption Ionisation (MALDI) is another ionisation method used for very large molecules: o The sample is dispersed in a solid matrix to form a solid solution. o A laser is then used to disintegrate the solid solution the matrix material is chosen so that it absorbs the laser wavelength. o Clusters ejected from the surface break up to give the sample molecule in [MH]+ or [MNa]+ form; some are multiply charged ions. o There is little fragmentation using this technique.

CFK/PJB

Analytical Chemistry: Page 46

8.2

Analysis and detection of ions: three methods Once the ions have been formed, we need to separate them according to m/z ratio. This part of the spectrometer will need to be at a good vacuum because we dont want the ions to hit other molecules before reaching the detector. The analysis method used will depend on mass resolution required, mass range, scan rate, and detection limit required.

Magnetic-sector instruments: this is the traditional method, but is now

becoming rarer. o These accelerate ions through a voltage V and then deflect them by a magnetic field (B) through a radius of curvature r. o Only ions of correct mass-to-charge ratio (m/z) will reach the detector:

o Scanning either the magnetic field or the accelerating voltage produces a spectrum.

[from http://www.cem.msu.edu/~reusch/VirtualText/Spectrpy/MassSpec/masspec1.htm#ms1]

Time-of-flight (TOF) method:

o The ions are accelerated through a voltage (V) and the time taken for them to travel a specific distance to reach the detector is measured. o The kinetic energy that each ion has is given by:

o Hence ions of low m/z will have large v and reach the detector first. o TOF instruments are usually very sensitive, and can scan a large mass range.

CFK/PJB

Analytical Chemistry: Page 47

Quadrupole detectors:
o The ions pass through an area with four hyperbolic magnetic poles created by a radiofrequency field. o Only certain ions can take a stable path through the field and be detected. o Scanning the rf field (by increasing voltage) allows a mass spectrum to be quickly and easily recorded, but resolution is more limited than the other methods.

8.3

Uses of MS Low-resolution mass spectrometry gives integer masses for peaks: o Useful for structural identification High-resolution mass spectrometry can be very accurate: o It can distinguish between CO, N2 and C2H4; these have exact molecular weights of 27.9949, 28.0062 and 28.0313 mass units respectively. o Even more useful for structural identification. For small organic molecules, the fragmentation pattern in EI MS often provides structural information: o For instance a peak at M+16 may correspond to loss of NH2 suggesting an acid amide to be present. o Fragmentation can permit identification of compounds by comparison with database of known mass spectra. The different isotopes of the elements are very important in MS: o For instance, molecules containing a single chlorine atom will give separate MS peaks for the 35Cl and 37Cl isotopes (in a relative ratio of ~3:1) o The natural abundance of 13C is 1.1%; MS peaks due to ions with one (or more) 13 C isotope in them will be observed, particularly for medium and large molecules. MS is used sometimes on-line in the process industries: o Quadrupole detectors arent too expensive and are fairly robust o The most common use is for analysis of gas-phase species; only difficulty is reducing pressure from process operating conditions to good vacuum inside instrument.

CFK/PJB

Analytical Chemistry: Page 48

Chromatography
All the techniques discussed so far are limited in one important respect: o They can only easily analyse pure compounds, or simple mixtures at best. What do we do if the sample is a complicated mixture? Chromatography is the key separation technique used by analytical chemists when the sample is a mixture. After calibration, chromatography can be used for structural identification and quantitative measurement as well as simply being a separation technique. Chromatography is also a chemical engineering unit operation for purification of highvalue chemicals, particularly in the pharmaceutical and biotechnology industries. Chromatography is based on the physical separation of individual chemical components in a sample: o The sample is present in a mobile or carrier phase: may be gas, liquid, or even supercritical fluid. o The sample is separated into components due to differences in affinity for a stationary phase.

9.1

Gas chromatography (GC)

For GC, the carrier phase is an inert gas (e.g. He, Ar, N2). The sample needs to be vaporised if its not already a gas, and injected as a pulse into the carrier stream. The stationary phase is usually a column containing the stationary phase on a fused silica support: o The column is usually very narrow (say 2 mm diameter) and may be 1-10 m long; physically it will look like a coiled loop. o There are many types of silica and modified silica so different separations can be achieved. Separation is based on the components having different retention times on the column: o affected by boiling points of the substances to be separated. o affected by selective adsorption of a component onto the stationary phase. The column is located in an oven, the temperature of which can be controlled. For good separations in a reasonable length of time, its common for the temperature of the oven to be increased over the course of the experiment.

CFK/PJB

Analytical Chemistry: Page 49

GC experimental configuration:

[from http://teaching.shu.ac.uk/hwb/chemistry/tutorials/chrom/gaschrm.htm]

There are a range of detectors available. Well mention the two most common.

Flame ionisation detectors (FID):

o These are widely used for analysis of organic compounds. o Gas at the column exit is mixed with hydrogen and air and burnt. Any organic compounds present produce ions and electrons in the flame making it capable of conducting electricity. The FID measures the current response to an electric potential at the burner tip. o The FID response has high sensitivity, a large linear response range, and low noise. It is also robust and easy to use, but it destroys the sample. o For hydrocarbons, the FID peak areas are proportional to the number of carbon atoms present in that component of the sample. Thermal conductivity detectors (TCD): o These compare the thermal conductivity of the gas at the column exit with a reference flow of carrier gas (usually He). Any change is due to the presence of sample compounds. o Disadvantage: TCDs are slightly less sensitive than FIDs and have slightly lower resolution (as they have a larger dead volume). o Advantage: TCDs can be used to detect any compound (i.e. not just hydrocarbons), and the sample isnt destroyed. o Because the thermal conductivity of organic compounds tend to be similar to each other, TCD peak areas for hydrocarbons are roughly proportional to the concentration of that component.

CFK/PJB

Analytical Chemistry: Page 50

Example: GC of a mixture of air + 9 hydrocarbons:

(a) column at 45C separate components 1-5; dont get components 6-9. (b) column at 145C doesnt separate 1-4; does separate components 5-8. (c) column heated a linear rate from 30-200C separates components 1-9. Theres also less band spreading as a function of time. From http://www.uft.uni-bremen.de/chemie/pdf/GC_Intro_Christian_Jungnickel.pdf

45C

time,min 145C

time,min

CFK/PJB

Analytical Chemistry: Page 51

T ramp: 30-200C

9.2

Liquid chromatography (LC)

Note modern instruments often use the acronym HPLC (for high-performance or high-pressure LC). In this case the sample is in the liquid phase (by dissolving into solution if its not already a liquid). The stationary phase is a solid packed into a column; it can be a liquid-coated solid. Different detector systems are used (e.g. UV, fluorescence, refractometry). A variety of different separation mechanisms are used. Liquid-Solid separations are based on the intermolecular interactions between sample molecules and the solid phase. o For instance, these may be polar interactions or hydrogen bonding interactions. o The normal case is that the solid has hydroxyl groups at the surface and so has an affinity for polar groups: Less polar molecules will pass through the column faster than polar molecules if the surface of the solid likes polar species. o In reverse phase chromatography, the stationary phase is made hydrophobic (e.g. silica with n-alkyl chains covalently bound to its surface): In this case, hydrophobic compounds will have longer retention times than hydrophilic ones. o The retention times are critically affected by the polarity of the solvent (carrier phase).

CFK/PJB

Analytical Chemistry: Page 52

Mixtures that dont separate when using one solvent as carrier phase may separate easily using a solvent of different polarity. o This technique is widely used in synthetic organic chemistry labs as a bench-scale purification technique.

Carrier solvent Inject sample

Carrier solvent

Carrier solvent

Carrier solvent

3 2 1 3

Packed column

2 1

3 2 1

Collect product 1

Liquid-Liquid separations are based on the partition of the sample between two liquid phases. Size-Exclusion chromatography is based on the molecular size of the compounds present. o The stationary phase consists of solid beads containing small pores. o Large compounds cant enter inside the beads and thus will elute first. o Smaller compounds enter the beads and will have longer retention times. Ion-exchange chromatography operates on the basis of selective exchange of ions in the sample with those in the stationary phase.

CFK/PJB

Analytical Chemistry: Page 53

o The column consists of a polymer matrix bearing certain ionic functional groups: e.g. MSO3H+ for the case of cation exchange (anion exchange columns also exist). o Molecules capable of ion-exchange will be retained at these sites : e.g. if they contain cations or acidic hydrogens in this example. o Molecules retained on the column can be subsequently collected by changing the properties of the mobile phase: e.g. by changing the pH so that the carrier liquid will displace sample molecules attached to the column. Affinity chromatography uses immobilized biochemicals that have a specific affinity to the compound of interest.

CFK/PJB

Analytical Chemistry: Page 54

9.4

Hyphenated techniques These are simply a combination of the techniques that weve discussed so far. Examples include: o GC MS o LC MS o LC NMR With these techniques, separation and sample identification of complex mixtures can be performed in a single piece of equipment. This represents the state of the art in modern analytical chemistry.

CFK/PJB

Analytical Chemistry: Page 55