Notes prepared by

Professor David Littlejohn

Department of Pure and Applied Chemistry/CPACT

University of Strathclyde

Cathedral Street Glasgow

G1 1XL

UK

Revised October 2007

6.Molecular Spectrometry

6.1Introduction

Developments in spectroscopic techniques have provided the greatest advances in process analysis in recent years. Radiation in wavelength regions from x-rays to microwaves has been used successfully for a range of applications involving at-line, on-line, in-line and non-invasive analysis.

The techniques of widest applicability and greatest interest are UV-visible spectrometry, mid-IR spectrometry, near-IR spectrometry and Raman Spectrometry. Only a brief introduction to the use of each technique in process analysis can be given in this course. However, in the following sections, the main features and attributes of the techniques will be described, with examples given to illustrate typical applications.

Process gas stream analysis featured in many early reports of on-line spectroscopic measurements, but more recently, the compositional analysis of liquids, solutions and suspensions has been achieved. Invariably, the complex nature of the molecular spectra of mixtures requires that advanced mathematical procedures be used to extract quantitative information about the analytes from the multivariate data recorded. This means that simple univariate calibration procedures (e.g. absorbance at one wavelength versus concentration) are rarely feasible. It is not a coincidence that advances in spectroscopic process analysis have occurred simultaneously with the availability of high-power low-cost computers and developments in multivariate calibration procedures based on principle component analysis (PCA), partial least squares (PLS) regression, neural networks etc. Consideration of these mathematical techniques is beyond the scope of this course, but interested readers are directed to Chapter 8 on Process Chemometrics, in “Process Analytical Chemistry”, by McLennan and Kowalski, where an overview of these methods is given.

By the end of this section you should be able to:

  • Understand how molecules interact with radiation in the different regions of the electromagnetic spectrum.
  • Describe the operation of simple spectrometers.
  • Give the relative advantages and disadvantages of working in the mid-IR, near-IR and ultraviolet-visible regions.
  • State the advantages of interferometric spectroscopic instruments

The general features of the spectra produced by different optical spectrometry methods are compared in table 5 [1]:

Table 5– Comparative advantages and disadvantages of molecular

spectroscopic techniques.

Method / Advantages / Disadvantages
Fluorescence
Absorption:
Mid-Infrared
Near-Infrared
Visible
Raman Scattering / Strong signal.
Compatible with silica fibres.
Features are distinct and based on fundamental absorption by characteristic groups in molecules. Generally good signal to noise ratio.
Compatible with silica fibres. Generally good signal to noise ratio.
Compatible with silica fibres.Generally good signal to noise ratio.
Sharp, distinct spectral features. Any wavelength can be used for excitation, barring fluorescence problems. Presence of water poses no difficulty. / Broad overlapping spectral features. Some analytes do not show fluorescence.
Incompatible with silica fibres. Requires optical elements made from salts that transmit in the MIR region.
Broad overlapping spectral features. A high water concentration can obscure other species in the spectrum.
Many analytes are clear, colourless liquids and have no absorption spectrum. Broad overlapping spectral features.
Signal to noise ratio can be poor for low concentrations. Features can be obscured by fluorescence.

6.2Additional Notes on UV-Visible Spectrometry

Some characteristics of UV-visible spectrometry, with regard to process analysis, are summarised below.

  • Absorption and fluorescence measurements can be made in the UV and visible wavelength regions, although absorption is more common.
  • The UV region normally studied is 200 to 400 nm and the visible region is 400 to 800 nm.
  • The transitions involved occur between electronic energy levels in

molecules (see figure 23)

  • Many organic compounds and some inorganic compounds will absorb in the UV-visible region, but not all will fluoresce
  • Broad spectral bands are produced, which result in overlapping spectra of compounds in mixtures
  • Quantitative UV-visible spectrometry involves the application of the Beer-Lambert law
  • The detection limits for analysis by UV-visible spectrometry are lower than those of mid-IR and near-IR spectrometry, giving more sensitive determinations
  • Consequently, an ATR probe is often required when analysing strongly absorbing solutions
  • Flow-through or in situ transmission probes are suitable for the measurements of weakly absorbing spectra (i.e. low concentrations and/or low absorptivity)
  • Flow injection analysis (FIA) is sometimes used with UV-visible

spectrometry for automated sample introduction and treatment in process analysis

  • Compounds that do and do not absorb at 254 nm and higher wavelengths are given in table 5
  • Conjugated unsaturated species such as azo dyes exhibit strong absorption in the visible region (rather than the UV), as lower energy is required to excite electrons in these systems than in covalent bonds


Figure 23 – Electronic Transitions in Molecules

Table 6– UV-visible absorbing and non-absorbing compounds (at 254 nm and higher wavelengths).

Absorbing (254 nm or above) / Non-Absorbing (254 nm or above)
Halogens
Aromatic compounds:
(Phenols, xylene, azo dyes and naphthalene)
Oxidising agents:
(Sodium hypochlorite, chlorine dioxide, hydrogen peroxide, ozone and potassium permanganate)
Inorganic compounds:
(Salts of iron, nickel, manganese and copper)
Sulphur compounds:
(Hydrogen Sulphide and sulphur dioxide)
Pollutants:
(NOx, SOx, chlorine, phosgene and ozone are typical gas phase pollutants / Inorganics:
(Hydrochloric acid, carbon monoxide, carbon dioxide, hydrogen, oxygen and water.)
Saturated hydrocarbons:
(Butane, methane, ethane and propane)
Unsaturated hydrocarbons:
(Acetylene, ethylene and propylene)
Lower alcohols:
(Ethanol, methanol, n-butanol, n-propanol, isopropanol and isobutanol)
Acids:
(Acetic, butyric and propanoic)
Esters:
(Butyl acetate, ethyl acetate and vinyl acetate)
All ethers
Chlorides:
(Ethyl, methyl, and vinyl chlorides)

Note – some of the compounds that do not absorb at 254 nm or above, absorb at lower wavelengths.

ATR Technology

Under certain conditions radiation passing through a prism of high refractive index material will be totally internally reflected. An evanescent wave is created which escapes from the crystal surface. The angle at which total internal reflection occurs is called the critical angle (θc) and is governed by the following equation:

θc= sin-1 n2/n1

When a sample is brought into contact with the prism surface the evanescent wave will be attenuated in the regions of the spectrum at which the sample absorbs. The depth of penetration (dp) for a single reflection is given by the following equation:

= 0.5

Where θ is the angle of incidence, λ is the wavelength of the incident light, n1 is the refractive index of the ATR crystal, n2is the refractive index of the sample.

The benefits of ATR technology include:

  • Inert crystal materials can be used such as sapphire and diamond that can cope with harsh environments and are not brittle
  • Bubbles and particulate materials do not have a significant effect on the measurement
  • This technique can cope with large concentrations, therefore sample dilution is not necessary
  • Sample viscosity does not present any particular problem

UV-Visible Process Spectrometers

Three types of instrument are available for use in process analysis. A diode arraydetector instrument is similar to the design of laboratory-based diode array UV-visible spectrometers. The instrument is enclosed in an environmental cabinet with the sample from the process line delivered to the flow cell in the spectrometer. The advantages of this instrument are that rapid

analysis with simultaneous measurement at all wavelengths in the UV-visible spectrum is possible, producing laboratory-quality spectra. Also, the analyser can be reconfigured easily for flexible use in different applications. The disadvantages are that the instrument is costly and problems can arise in sample transfer.

An alternative is a fibre optic UV-visible spectrometer that uses a grating toscan the spectrum. A disadvantage is that the instrument is slower than the diode array detector spectrometer, but is still useful for multi-component work. An advantage is that the spectrometer is located away from the process and up to 200 m of fibre optic cable is used to transmit the source radiation to the probe head in the process liquid. Also, different types of probe head (ATR, transmission) can be used and the cost of the instrument is less than that of the spectrometer with the diode array detector.

The third spectrometer is a filter instrument (i.e. non-dispersive), which is cheaper than the other two analysers. In a typical filter analyser, two or more wavelengths are used to record the absorption measurements. Radiation passes through the sample into the photometer housing. The beam passes through a series of filters in turn (see diagram below). The filters are chosen so that strong absorption occurs at three of the wavelengths with no or low absorption at the reference wavelength (e.g. number 1 in the diagram). Changing the filters can vary the wavelengths to suit the application. The reference measurement compensates for any dirt accumulation on the cell windows and for any light source variance. The rugged nature of these analysers makes them a cheaper alternative to the full scanning systems. Filter systems can use ATR or flow-through transmission cells (as in the diagram) that bolt directly on to the process pipe.

Example application of UV-visible absorption spectrometry – dye

manufacturing

The schematic diagram below (Figure 24) illustrates a dye manufacturing process where a weak dye/saline solution is initially desalinated and then concentrated. The desalination process is carried out by passing the dye through a bank of membranes, which also concentrates the dyestuff. To keep the dye at the initial concentration, water is added. The desalination step must occur before turning the water off and concentrating the dye. The process requires two liquors to be analysed, the concentrated dye flowing into the hold tanks and the effluent from the membranes.


Figure 24 – Schematic of a dye manufacturing process

6.3 Review Questions

1.In the above example, typical dye concentrations (c) for the effluent and the concentrate are 0.1 g l–1and 200 g l–1, respectively. Calculate the pathlength (b) assuming that Beer’s law applies and that a maximum operational absorbance (A) of 1.0 and an absorptivity value (a or ε) of 20 l g-1 cm-1 are typical. Decide on the most suitable type of measurement probe in each case.

2.If the refractive index values of the concentrated solution and a 3-reflection sapphire ATR crystal are 1.3 (n2) and 1.76 (n1), respectively, and absorption occurs at 589 nm, calculate the angle of reflection at the crystal surface required for analysis in the above example.

Note In this example only one absorbing species was present. If more than one compound can cause absorption at the measurement wavelength, quantitative analysis requires the use of multivariate calibration methods for accurate determination of the species of interest.

3.Comment on the different types of UV-visible spectrometers used in on-line or in-line process analysis, including their advantages and disadvantages.

6.4Additional Notes on Mid-Infrared Spectrometry

Some features of mid-infrared (MIR) spectroscopy are as follows:

  • The wavenumber region normally covered is 4000 – 400 cm-1
  • The vast majority of organic molecules exhibit MIR absorption; a net

change in dipole moment must occur during the vibrational process

  • Allows the direct monitoring of vibrations of functional groups in

molecules

  • Spectra contain frequencies corresponding to the fundamental vibrations of virtually all the functional groups of organic molecules
  • Qualitative and quantitative information can be gained using functional group frequencies and the Beer – Lambert law, respectively
  • The MIR spectra of mixtures of compounds can be complex and

multivariate calibration may be required to quantify species

  • Spectral lines are narrower and more distinct than for NIR spectra
  • Mid-IR frequencies are much more strongly absorbed than in the near-IR region
  • Sample pathlengths have to be short; ATR crystal probes have a sample pathlength of 10 – 100 μm and so are useful in the MIR region
  • Strong absorption also limits the materials that can be used to transmit the radiation. Attenuation is high and fibre runs are a few metres at most

Mid-infrared Process Spectrometers

A comparison of the characteristics of mid-infrared analysers is given in table 7.

Table 7 – Comparison of the characteristics of MIR Instruments

Filter IR Instruments / Dispersive IR Instruments / FT-IR Instruments
Low signal to noise ratio
Medium energy throughput
Stray radiation is a problem
Compound specific application
Monochromatic or limited sequential polychromatic detection
Applicable to the analysis of gases / Low signal to noise ratio
Low energy throughput
Stray radiation is a problem
Generic application
Sequential polychromatic detection
Applicable to the analysis of gases or liquids / High signal to noise ratio
High energy throughput
No stray radiation
Generic application
Simultaneous
polychromatic detection
Applicable to the analysis of gases, liquids or solids

As mentioned in section 5 on optical fibres, coupling of a MIR spectrometer to a remote probe is restricted by the problems of fragile, high cost or poor transmission waveguides. Useful results have been achieved on-line, when liquid is passed from a reactor via a pump to a spectrometer with a flow cell and ATR attachment (see following example). This has the disadvantage of requiring a sampling system. The advantage is that no waveguides are required. Some attempts have been made to use hollow highly reflective tubes to transmit MIR radiation from a spectrometer to an in-line probe. However, a new approach has been to miniaturise the MIR spectrometer and put it on the end of a short probe containing an ATR crystal connected to the spectrometer at the other end by chalcogenide optical fibre (see figure). The advantages are good transmission, simplicity and robustness. The disadvantages are that only a limited range is covered (1000 – 2000 cm-1), resolution is limited using a 128 array pyroelectric detector and only ATR measurements are permitted. An alternative procedure is to use the newer, more robust polycrystalline silver halide fibres with a conventional MIR process analysis spectrometer to achieve in-line analysis with an ATR probe.

For gas analysis, sample is extracted from the process and pumped to a near-by spectrometer fitted with a gas cell. This cell is 30 – 50 cm long, but has the possibility of allowing the radiation to be reflected back and forth several times, which increases the overall path length (in Beer’s law).

Example application of on-line MIR spectrometry

Prior to the availability of silver halide and chalcogenide fibres for in-line analysis by MIR spectrometry, an on-line arrangement was required. For example, a FT-IR spectrometer was set up to monitor the reaction of butan-2-ol with but-2-enoic acid. The experimental set-up and conditions are given in figure 25.

Figure 25 Experimental set-up for on-line monitoring of an esterification reaction by MIR spectrometry.

The FT-IR analysis was performed every 10 min. Samples were also extracted for analysis by GC every 30 min, to provide reference data. The formation of ester was monitored by spectral measurements at 1185 and 1100 cm-1(C – O ester stretching vibrations) and at 1720 and 1700 cm-1(C = O stretching). It was possible to study the kinetics of the reaction through changes in the FT-IR absorbance at different wavenumber regions. A full quantitative analysis would require the preparation of a multivariate calibration model.

6.5Review Questions

  1. Comment on the advantages and disadvantages of MIR spectrometry in process analysis.

2.Describe different ways that an ATR crystal can be used with an MIR spectrometer for process analysis. Comment on the advantages and disadvantages of each approach.

6.6Additional Notes on Near-Infrared Spectrometry

NIR analysis involves the absorption of infrared radiation with consequent vibrational excitation of the sample. The difference in comparison with the MIR region lies in the fact that overtone and combination bands of the fundamentals are the frequencies absorbed (typically 12800 – 4000 cm-1). C – H, O – H, and N – H are the main bands analysed in process analysis. NIR absorptions are generally weak, broad and poorly resolved which, almost inevitably, means that multivariate calibration is required for quantitative work. Univariate calibration with the Beer – Lambert law should always be the starting point, but rarely is the calibration model simple.

Features of Near-IR Spectroscopy:

  • Bands are due to combinations and harmonics (overtones). See table 8 for data on N - H absorption
  • As near infrared absorptions are weak, long sample pathlengths are required. This has the advantage of giving good flow through the sample cell and reduces the probability of blockages.
  • NIR wavelengths have very low attenuation through silica fibres allowing very long fibre runs to be used
  • Liquid phase spectra are dependent upon temperature
  • Vapour phase spectra can be obtained. These are immune to hydrogen bonding effects, which can be seen in liquid phase spectra.
  • Can be used for water detection at the ppm or percentage levels in organic liquid
  • Near IR can be used to analyse solids. Since absorption is weak, more light is subject to diffuse reflectance from the solid to the spectrometer.
  • NIR can be used with dispersive instruments or with an FTIR

spectrometer.

Near-Infrared Process Spectrometers

The instrumentation used for NIR is almost identical to that of UV-visible spectrometry, with the difference being that the source, filters/grating and detectors have to be changed. NIR lends itself to fibre optics and can utilise large pathlengths due to low absorptivity values. The signal-to-noise ratio (S/N) is far better in NIR than in MIR, owing to a higher source of output intensity and more sensitive detectors.

Table 8 – Data for N – H bonds in NIR

Nature of bond / Frequency
(cm-1) / Wavelength
(nm) / Molar absorptivity
(dm3 mol-1 cm-1)
Combination (stretch and bend) / 5072 / 1971.5 / 1.63
1st overtone (symmetric stretch) / 6698 / 1493.0 / 1.21
1st (asymmetric stretch) / 6094 / 1448.5 / 0.156
2nd overtone
(symmetric stretch) / 9794 / 1021.0 / 0.038

Note - the absorptivity values are less than for MIR transitions and many orders of magnitude less than encountered in UV-visible spectrometry