Liquid Sample Introduction in ICP Spectrometry (eBook)

Chapter Specifications of a Sample Introduction System to be Used with an ICP

2.1. Introduction

Like for a flame, liquid sample is usually introduced into the plasma in the form of an aerosol consisting of fine polydisperse droplets, along with some vapour produced during the droplet transport in the spray chamber. A vapour is a gas phase in a state of equilibrium with a liquid of identical matter below its boiling point. When the droplets enter the plasma at a rate of several millions per second, they undergo three major processes – desolvation, volatilization and atomization – before reaching the state of free atoms. Desolvation is the evaporation of the solvent from the droplets that leads to a suspension of a dry aerosol, that is desolvated particles. During the evaporation process, droplets and dry particles are surrounded by a solvent vapour cloud. When solvents evaporate, they remain near their boiling point, that is far from the plasma temperature. Volatilization produces the conversion of the dry aerosol into a gas or a vapour. Solute particles have a broad range of boiling points with values comparable to what can be observed within the plasma. Two transfer mechanisms are involved: heat transfer from the plasma that leads to surface boiling and followed by mass transfer from the surface of the droplet or particle. Atomization is the conversion of volatilized analytes into free atoms. It should be noted that the word atomization is also used to describe the conversion of a liquid into a spray or a mist. The sample introduction system is, then, called an atomizer, which is obviously wrong because no atoms are produced. Nebulizer is most appropriate as derived from the Latin nebula (mist).

Once free atoms are obtained, they undergo partial ionization, and both atoms and ions are excited. Timescale for the first three processes is of the order of ms, whereas excitation and ionization are significantly faster. Because an inductively coupled plasma (ICP) uses hf power of the order of kilowatts and reaches kinetic temperature in the range 4000–7000 K, it could be thought that this source can absorb a large amount of sample without any problem of evaporation and atomization. However, an argon-based ICP suffers from some severe limitations. A high-frequency field is mostly coupled with the periphery of the plasma, because of the so-called skin effect. Energy diffuses to the plasma centre through collisions with losses, and only a small fraction is available along the axis. The energy consumed for desolvation and vaporization is actually a small fraction of the forward power. Besides, argon has a poor thermal conductivity, and residence time of the sample is rather short. Moreover, there has been a constant trend to decrease the available power of the hf generator for cost and size reasons.

An ideal sample introduction system should lead to complete desolvation and volatilization in the plasma. For that, its design must consider the efficiency of the heat transfer from the plasma to the droplets, which is related to the power reaching the sample and the residence time of the droplets. The residence time is, in turn, linked to the carrier gas flow rate, the injector’s i.d., and the temperature along the axis of the plasma.

Practically, the design must take into account the answers to some key questions such as:

• What is the maximum droplet diameter ensuring a complete desolvation and volatilization?
• What is the maximum amount of aerosol, in other words, the plasma loading, acceptable by the plasma?
• Is it possible to avoid any change in the sample stoichiometry?
• Is the system sensitive to any change in the physical properties of the solution, for example viscosity, surface tension, density and volatility in the case of organic solvents?
• Is desolvation and vaporization sensitive to any change of the element concen-trations in a droplet, particularly those of the major elements?
• Is there any benefit to keep water, or to remove it before introduction into the plasma?
• How to minimize noise because of liquid delivery, aerosol production, filtration and transport in the introduction system, drain and processes in the plasma?

Obviously, the answers to these questions depend upon plasma characteristics, torch design and operating parameters such as:

• the physical properties of the plasma gas such as ionization energy and thermal conductivity, which in turn are related to the selected plasma gas,
• the energy coupled to plasma, which depends on the generator design and frequency, and on the available hf power and the coupling efficiency,
• the plasma gas flow rate, that is the amount of gas per time unit, usually expressed in L min−1,
• the residence time in the plasma that depends upon the carrier gas velocity resulting from the gas flow rate and the injector i.d., and upon the temperature along the central channel,
• the desolvation, evaporation and atomization timescale,
• the solvent loading accepted by the plasma.

2.2. Physical Properties of a Plasma

The physical properties of a plasma, such as ionization energy and thermal conductivity, depend obviously upon the gas. Rare gases are usually used to generate plasma, because they emit only atomic spectra in emission spectrometry, and relatively simple spectra in mass spectrometry, although some rare gas–based molecular ions can survive during their transport to the ion detector. The highest ionization energy is for He (24.6 eV), followed by Ne (21.56 eV). The He gas has been used to sustain ICPs, but the cost can be a limitation, kinetic temperature is lower than that of an Ar plasma, and introduction of a sample is more challenging [153, 190, 227, 337, 367, 416, 449, 513]. Although providing interesting results [674], the Ne gas cannot be used on a routine basis because of its unacceptable cost. The argon gas, which has lower ionization energy (15.76 eV), exhibits some advantages. It is the most common of the rare gases, with 1% in air, which means that its cost is acceptable as a subproduct of air distillation. However, it is only easily available in industrialized countries. Some properties are important for sample injection and atomization. Because of the increase in the number of free electrons when the temperature goes up, the viscosity will move from 2×10−5 kg m−1 s−1 at room temperature to 20×10−5 kg m−1 s−1 at 6000 K, that is a factor of 10. This means that it would be difficult to inject a sample by using a cold Ar gas because of the penetration in a medium that is significantly more viscous. However, thanks to the use of a hf field, a skin effect is observed; that is the energy is mostly deposited at the periphery of the plasma, and then translated by particle collisions to the axis of the plasma. In other words, the energy, and therefore the temperature, is lower along the axis of the plasma, and so is the viscosity. The axis of the plasma is a zone where it is easier to introduce a sample by forcing the penetration with a gas having a sufficient speed. There is creation of a virtual central channel, which explains the success of an ICP. Sample introduction in other types of plasma such as microwave-induced plasmas or direct-current plasmas is far more challenging.

Another crucial parameter is the thermal conductivity, that is the capability to transfer heat. Molecular gases, air, N2, O2 and H2 exhibit thermal conductivities at 5000 K that are 5.8, 5.4, 4.5 and 31.6 times higher than that of Ar, respectively. Even He has a thermal conductivity that is 10.1 times higher. This means that when using Ar, residence time of the sample must be longer to compensate for a lower thermal conductivity.

2.3. Energy Delivered by the Plasma

There has been a constant trend to decrease the size and the cost of ICP spectrometers. One means of obtaining this reduction is through the use of low-power hf generators. Currently, the maximum power that can be used on commercially available ICP systems is in the 1.5–2 kW range, and the operating power recommended by ICP manufacturers is typically below 1.5 kW (1000 W in ICP–AES and 1200 W in ICP–MS). This has to be compared with early work in ICP–AES, where 7 MHz, 15 kW (Radyne) and 5.4 MHz, 6.6 kW (STEL) hf generators were used. Increase in the generator frequency was also observed along with the decrease in the power: 36 MHz, 5 kW (Radyne), 50 MHz, 2 kW (Philips), 64 MHz, 4.3 kW (Durr) and 27 MHz, 2.5 kW (PlasmaTherm). Currently, commercially available ICP instruments make use of 27 or 40 MHz (Tables 2.1 and 2.2). The operating frequency determines the skin depth, that is the penetration of the energy at the periphery of the plasma, and the coupling efficiency [265]. Coupling efficiency is also...