Silicon speciation by hyphenated techniques for environmental, biological and industrial issues: A review

Abstract

Silicon speciation in environmental, biological and industrial matrices is of considerable importance due to its wide use in many consumer and personal care products and industry. In addition, the entry of silicones in various compartments like wastes, soils, air and water highlights the need to perform exposure studies, toxicological surveys and to measure negative effects. Due to possible contamination and trace level presence of silicon compounds, challenges to determination, identification and quantification are presented. The principal species of concern include siloxanes, silanols, silanediols and silanes. State of the art of analytical methods for total silicon determination and silicon speciation are established. Atomic spectroscopic methods are mainly used to measure total Si at trace concentration levels. On the opposite, hyphenated techniques are performed for Si speciation. Particular attention is paid to chromatographic methods coupled to sensitive and selective detectors (MS, AED and ICP) allowing structural information. Liquid and gas chromatography emerge as the most widespread separation techniques. However, other procedures such as MS, NMR, IR and XRF enable a better knowledge of these species. The potential and limitations of hyphenated techniques are highlighted, particularly concerning sensitivity and selectivity. Furthermore, potential sources of contamination and analytical artifacts in silicon determination are reviewed.

---

F. Chainet

Fabien Chainet graduated from the University of Orleans with a master degree of Chemical Pollution and Environmental Risk (CPRE) in 2009. He is a Ph.D. student working on ultra and trace analysis in petroleum and derived products and more precisely in silicon speciation in hydrotreatment feeds at IFP Energies Nouvelles.

C.-P. Lienemann

Charles-Philippe Lienemann graduated from the University of Geneva, Switzerland, in 1993 and obtained his Ph.D. from the University of Lausanne, Switzerland, in 1997. He then worked at the Wolff Environment Labs in Lyon, France, before joining the IFP Energies Nouvelles in 2000 as the head of the elemental analysis lab. His research is focused on the behaviour, fate and analysis of trace metals in petroleum and related samples.

M. Courtiade

Marion Courtiade graduated from the School of Chimie Physique and Electronique (CPE) in 2004. She obtained her Ph.D. from Lyon University, which deals with trace analysis of pesticides in the environment. Since 2007, Dr Courtiade has been at the head of the GC laboratory at the IFP Energies Nouvelles and is an expert in characterization by comprehensive gas chromatography.

J. Ponthus

Jérémie Ponthus graduated in 2002 and obtained his Ph.D. dealing with Metastable Atom Bombardment Mass Spectrometry for petroleum related molecules characterisation from the University of Paris (UPMC), France, in 2005. He then joined the IFP Energies Nouvelles where he is now at the head of the Mass Spectrometry laboratory. His research is focused on the FT/MS analysis of complex mixtures from heavy crude oil, biomass and coal.

O. F. X. Donard

Olivier François Xavier Donard is a research director at the French CNRS (Centre National de la Recherche Scientifique). He is the vice president of the analytical chemistry division of the French Société Chimique de France. His research interests are related to the speciation and isotopic signature of elements in the environment, and industrial processes and their environmental chemistry.

---

Introduction

Organosilicon compounds such as siloxanes and polydimethylsiloxanes (PDMS) are widely used in a variety of industrial applications and consumer goods since their commercial introduction in 1943.¹ These molecules find applications in a wide array of common consumer products such as cosmetics, textiles, medicinal implants, pharmaceuticals, chemistry and in industry (lubricants, coatings, gels, adhesives, antifoaming agent…).^2–4

Silicon species used in these applications are polymeric silicon organo compounds formulation consisting of a backbone of alternating Si–O units with organic side chains attached to each silicon atom.³ These structures provide the species with unique high thermal stability, low surface tension, hydrophobicity, electric insulation and lubrification properties.⁴ These specificities resulted in the fact that some siloxanes have been identified as high production volume (HPV) chemicals by the US Environmental Protection Agency (USEPA)⁵ and by the Organisation for Economic Co-operation and Development (OECD).⁶ HPV products are defined as compounds produced or exported exceeding 1000 tons per year in at least a region or a country.⁶ As a result of their ever wide usage, these different products are dispersed in the environment and are present in a large variety of biological and industrial matrices (petroleum and derived petroleum products) most frequently at trace levels. Their ubiquitous occurrence in the environment requires the silicon species to be investigated to assess their environmental fate.^3,7 Both the ecological⁸ and toxicological impact of the silicon^9,10 have been recognized to be of concern, therefore, it is necessary to carry out exposure and effect studies of the species.^11–13 In order to improve the general knowledge of the occurrence, pathways and toxicity assessment of silicon species, speciation of silicon and their species must be undertaken to improve the understanding of their impact and translocation in the environment.

According to the International Union of Pure and Applied Chemistry (IUPAC), speciation is defined as the analytical process that leads to measure the distribution of an element specific chemical species in a sample.¹⁴ Silicon speciation is gaining in interest and addresses the identification and quantification of the physical and chemical forms of silicon.¹⁵

The developing domain of silicon speciation is being reviewed and Fig. 1 presents the number of publications for Si determination and speciation in its different application domains. This figure shows that environment, medicine, biology, polymer synthesis and petroleum applications are the most investigated fields of interest. Two main books deal with the analytical chemistry of silicones^1,16 but more specifically on total Si determination. Also a review was published in 2006 by Varaprath et al. on silicone analysis and its artifacts in environmental and biological samples.¹⁷

Fig. 1 Publication distribution in percentage for Si determination and speciation in different application domains.

However, silicon speciation as such is not well developed and the number of articles is small. Silicon is in general present at very low concentrations in most of the investigated matrices.^18,19 Further, contamination problems during the entire analytical procedure²⁰ as well as in the artifacts,^21,22 generates a real analytical challenge. The reason being that due to their widespread use of silicon species in many products, their stability and of the high reactivity of certain silicon species which may alter the analytical process. Therefore, a wide variety of high-performance analytical techniques have been developed and applied. Fig. 2 summarises the results of a general survey of more than 120 articles published over the last 30 years dealing with total silicon determination methods as well as the hyphenated techniques applied for silicon speciation.

Fig. 2 Number of publications for Silicon determination and speciation by different methods.

The objective of this article is to review the main analytical methods used for silicon determination and silicon speciation in environmental, biological and industrial applications. One chapter is focused on contaminations and analytical artifacts which in the case of silicon is of special concern for the validity of the data and has been a major drawback for the development of this analytical domain. In general, direct analytical methods based on atomic spectroscopic methods (AAS, ICP-OES and ICP-MS) and also NMR techniques give access to total silicon measurement. All the direct spectrometric methods do not provide information about silicon speciation (Fig. 2). Silicon speciation relies on hyphenated techniques combining the high separation potential of chromatography (e.g., LC or GC) with identification possibilities offered by mass spectrometry (MS) or the sensitivity and selectivity of an element specific detector (AED, ICP-OES and ICP-MS). This review is mainly focused on silicon speciation by hyphenated techniques particularly at trace concentrations compared to classical total determination methods for silicon. Special attention is given to the contamination problems encountered during the entire analytical process which in this case is of particular relevance for the data quality.

1 Molecules, fields of interest and analytical methods

Organosilicon compounds

The present paper deals with siloxanes (cyclic and linear), silanols (trimethylsilanol), silanediols (dimethylsilanediol) and silanes. These correspond to the four main families of organosilicon species²³ (Table 1) found in the fields of interest reported in Fig. 1. As siloxanes have long names, abbreviations created by Hurd²⁴ are employed for several species such as D_n (Table 1). Siloxanes are typically represented by the following letters “M, D, T, Q” which correspond to the number of oxygen atoms linked to an atom of silicon. Indeed, M indicates a silicon linked to one oxygen, while D, T, Q indicate a silicon linked to two, three and four oxygens respectively. Moreover, Table 1 lists formula, molecular mass, boiling point and structure of species according to the fields of interest presented in Fig. 1.

Table 1 Molecules of interest

Silicon species (Hurd Abbreviations)	Formula	M M/g mol^−1	B.P/°C	Structure	Fields of Interest ^a
Siloxanes
Polydimethylsiloxane (MDnM)	PDMS viscosity: 12,500 cSt	67,700	—		EN
					BI
Hexamethylcyclotrisiloxane (D3)	C6H18O3Si3	222	134		BF
Octamethylcyclotetrasiloxane (D4)	C8H24O4Si4	296	175
Decamethylcyclopentasiloxane (D5)	C10H30O5Si5	370	211		CP
Dodecamethylcyclohexasiloxane (D6)	C12H36O6Si5	382	245
Hexamethyldisiloxane (MM) or L2	C6H18OSi2	150	100		P
Octamethyltrisiloxane (MDM) or L3	C8H24O2Si3	236	153
Decamethyltetrasiloxane (MD2M) or L4	C10H30O3Si4	310	194		PP
Dodecamethylpentasiloxane (MD3M) or L5	C12H36O4Si5	382	210

---

Silanols
Trimethylsilanol (TMS)	C3H10SiO	90	99		EN
Triethylsilanol (TES)	C6H16Osi	132	158		PP

---

Silanediols
Dimethylsilanediol (DMSD)	C2H8O2Si	92	102		EN
					PP
Tetramethyl-1,3-disiloxanediol (TMSD)	C4H14O3Si2	166	—		EN

---

Silanes
a EN: Environmental samples (air, soil, water, sediments); BI: Breast implants; BF: Biological fluids; CP: Consumer products; P: Polymers; PP: Petroleum and derived products.
Tetramethylsilane	C4H12Si	88	27		EN
					PP
Triethylsilane	C6H16Si	116	107		PP
Diethoxydimethylsilane	C5H14O2Si	134	95		PP

---

1.2 Fields of interest

PDMS, defined as a high molecular weight (HWM) polymer is the most important silicone used for industrial and consumer applications (80%).³ Other silicones are mainly of low molecular weight (LMW) materials, also referred to as volatile methyl siloxanes (VMS) with significant vapour pressures under ambient environmental conditions.^13,25 Consequently, the behaviour of HMW and LMW siloxanes has gained in interest in many matrices. The following paragraphs report problems caused by these species.

Environment. The entry of silicones into the environment may also occur from applications resulting with wastes and volatilization (Fig. 3). As a result of their wide use, VMS (D_n and L_n) evaporate into the atmosphere^26,27 and degrade in the soils.²⁸ Additionally, some of these species can be produced by PDMS degradation in environmental conditions. Due to low solubility in water and high adsorption coefficient,⁷ PDMS ends up and degrades in landfills sewage,^29–31 biogas,^2,32 soils,^7,22,33–35 and sediments (Fig. 3).³⁶ The dimethylsilanediol (DMSD) is the main degradation product obtained by PDMS hydrolysis.^7,33,35,37 Then, it can be evaporated due to its high vapour pressure and oxidized by radicals as the other volatile species found in environment (tetramethylsilane, trimethylsilanol (TMS), L₂ and D₃-D₅).^38–40 Subsequently, this compound is biodegraded in soils where it is mineralized in end products (CO₂, SiO₂ and H₂O) (Fig. 3).⁷ Buch and Ingebringston have also shown the formation of low molecular oligomers (siloxanes and silanols) during PDMS degradation.⁴¹ In addition, Carpenter et al.⁴² reported the presence of monomerdiol (DMSD), traces of dimer and trimerdiols, octamethyltrisiloxane (L₃) and TMS during PDMS degradation in soils by hyphenated techniques.

Fig. 3 Fate of silicon in the environment.

LMW hydrolysis products are water soluble and volatile. Thus, these induce silicon molecules which can partition from the soil to the water and into the atmosphere (Fig. 3).⁷ In addition, combustion of silicon containing biogas produces the abrasive silica and causes serious damages to gas engines,^2,43 heat exchangers,⁴⁴ and catalytic exhaust gas treatment.⁴⁵

Long et al. have developed a novel on-site method for siloxane detection in biogas based on microcantilever array use.⁴⁶ Performance like high sensitivity, portability, inexpensive and less energy consuming than GC-MS promise to facilitate in-field siloxane analysis and reduce biogas cost.⁴⁶ Several applications are performed to reduce and eliminate silicon species in biogas.^47,48 Silicon speciation in the environment contributes to estimate the environmental fate and the ecological and toxicological risks occurred by siloxane compounds in vulnerable ecosystems (Fig. 3).⁸

Medicine and biology. PDMS is employed in silicone gel-filled breast implants (Fig. 3) for aesthetic surgery because it was supposed to be biologically inert.⁴⁹ This gel contains 98% of HMW and only 1–2% of LMW silicones (D₃-D₇).⁵⁰ Lykissa et al. reported that D₅-D₇ are the most concentrated of LMW molecules content in silicone implant gel.⁵¹ These compounds can be released in blood,⁵⁰ plasma,⁵⁰ and tissues^51,52 of human exposed to breast implants. Consequently, sensitive techniques have been used to quantify^50–53 silicones in order to carry out epidemiologic studies. Currently, there is no credible correlation between the increase of diseases such as cancer originating in the breast among women with cosmetic breast implants.^54,55

Consumer products. VMS siloxanes such as D_n and L_n are mostly used in consumer products (Fig. 3). Two recent works achieve the determination of organosilicon compounds levels^12,13 in personal care and cosmetic products such as fragrances, hair care products, antiperspirants, skin lotions, baby products and many other consumer goods. These results show that more than 50% of the investigated samples contained D₄, D₅ and D₆.¹² Cyclic molecules were predominant compared to linear siloxanes. Wang et al. have also reported that D₃, D₄, D₅ and D₆ are respectively found in 0.8, 4.8, 14.3 and 9.1 percent of the analyzed products.¹³ Considering these studies,^12,13 D₅ is the most abundant compound (680 mg g⁻¹ in an antiperspirant).¹³ Regarding their high concentrations, one can wonder if a potential adverse effect can happen.¹² According to Wang et al., no risk for dermal exposure of these products has been identified because of two reasons: firstly the uncertainty in adsorption efficiency penetration rates in human skin and secondly in the evaporation during products use.¹³ However, speciation is necessary to subsequently perform exposure studies (Fig. 3) to silicon compounds such as occupational exposure assessment in silicon industry,⁵⁶ exposure to ambient atmosphere^26,27 and exposure per D₄ inhalation.¹¹ Furthermore, fertility problems can occur during D₄ exposure of animals^9,10 as well as potential carcinogenicity after rat exposures to D₅.⁵⁷

Polymers. Due to the need of polymer synthesis enhancement, several workers have carried out studies dealing with PDMS stability.^58–60 Experiments have been performed using thermogravimetric analysis (TGA) and GC-MS at high temperature and under helium flow for polymer applications like rubbers, insulating materials or ceramics. Fig. 4 illustrates the distribution of D_n compounds generated by the degradation of PDMS between 400 and 500 °C in three different publications.^58–60 The production of D_n tends to decrease as the polymerization degree increases. As shown in Fig. 4, LMW siloxanes (D₃-D₆) are the major compounds. Moreover, Camino et al. suggested that mixtures obtained by degradation in the presence of air are similar to that one under He flow.⁶⁰ However, many other molecules can be generated by decomposition according to temperature, oxygen levels,^60,61 impurities,^59,62 and polymer terminal group.^63,64

Fig. 4 Low molecular cyclic siloxane (D_n) distribution generated during PDMS degradation at different temperatures (Mass percentage).

Petroleum and derived products. This field is also concerned by silicon contamination due to the fact that it causes catalyst poisoning in refining processes and in particular in hydrotreatment (Fig. 5).⁶⁵ In addition, automobile and fuel sensors as well as emission control devices are susceptible to silica poisoning because of contamination from fuel.⁶⁶ Indeed, in 1985, samples of unleaded fuels with concentrations ranging between 30 to 110 mg kg⁻¹ of Si caused troubleshooting in motor engine combustion.⁶⁶ Even if the origin was not well established, one hypothesis consists in the unintentional use of recycled and contaminated aromatic process solvent (toluene) used as octane booster by area fuel blenders.⁶⁶ Similarly, this phenomena occurred once again in the United Kingdom in 2007 with 400 000 damaged cars due to the formation of a silica layer on an O₂ probe in automobile sensors.

Fig. 5 Silicon poisoning in petroleum process.

Silicon presence in feedstocks is derived from antifoaming agents (PDMS), added in the lighter fractions of coker or visbreaker operations.^67,68Fig. 5 presents a petroleum process constituted of feed thermal cracking (Coker and Steam cracker), separation and hydrogenation units. Due to the PDMS degradation, silicon compounds (Table 1) are found at trace levels^69,70 in feeds and occur as poison which induce severe catalytic deactivation by adsorption on the catalyst surface (Fig. 5).^68,71–75 According to theoretical explanations and GC analysis of heated silicon oil, several authors suggest that cyclic oligomers (D_n) are the major breakdown products of PDMS released in petroleum products.^67,76 These explanations are in agreement with results presented in Fig. 4^58–60 due to the high temperatures in petroleum processes.⁶⁷ According to Breivik and Egebjerg cyclosiloxanes are rapidly adsorbed on the catalyst surface of hydrotreatment and caused catalyst deactivation.⁷⁶ In addition, Molnar et al. demonstrated a poisoning of olefin hydrogenation by triethylsilane between 200 and 400 °C.⁷⁵ Furthermore, recent studies have shown poisoning by hexamethyldisiloxane (L₂) used as a model poison compound during combustion of volatile organic compounds (VOC) in gas treatment.^45,77–79 These phenomena give rise to untimely catalyst replacement and economic loss. Consequently, the petroleum industry must continuously propose sensitive analytical methods in order to reach the specifications in petroleum products.⁸⁰ As far as demonstrated in the literature, only a total determination of Si is achieved in petroleum matrices.^81,82 However, speciation (identification and quantification) is necessary to give access to the chemical nature of poison species and hence, a better understanding of the poisoning processes.

1.3 Analytical methods for silicon determination and speciation

As previously mentioned, silicon compounds result in numerous problems in the environment (fate and exposure), medicine and biology (breast implants), consumer products (exposure, toxicity), polymers (stability) and catalysts (poisons) in petroleum processes and gas treatment. Consequently, two different groups of techniques used for silicon determination, are presented in Fig. 6. These techniques consisted of direct analysis to achieve total silicon determination as well as, hyphenated techniques for silicon speciation. Before selecting an analytical method, it is essential to define the matrix, the species and compound concentration ranges especially at trace levels.¹⁷ Generally, detection limits greatly depend on the preparation and one concentration. Thus, comparison between different technique performances is difficult.

Fig. 6 Analytical methods for silicon determination and speciation.

Most total silicon determinations are performed by atomic spectroscopy, NMR and other methods by infra-red spectroscopy (IR) and X-Ray fluorescence (XRF) (Fig. 2). For silicon speciation, MS methods and SEC separation have been usually employed for HMW molecules. On the other hand, the methods on the right side of Fig. 6 are used for LMW compounds. Silicon speciation requires chromatographic separation (LC, SFC and GC) coupled to a suitable detector depending on the expected detection limit. The survey presented in Fig. 2 clearly shows that GC-MS is the most commonly employed hyphenated technique for trace level Si speciation. Flame ionisation detector (FID) can also be applied but it does not provide very low detection limits. The coupling of GC and atomic detection such as AED or ICP (inductively coupled plasma) allows high sensitivity and selectivity therefore, giving access to structural information. The combination of LC and ICP is also used for the determination of a wide variety of silicon compounds as well. MS methods can also be applied to LMW compounds (Fig. 6) providing identification compared to direct methods.

2 Direct methods for total silicon determination

Total silicon can be determined by atomic spectroscopic methods and NMR (Fig. 2 and Fig. 6). The major techniques employed for silicon trace analysis are AAS, ICP-OES and ICP-MS. This chapter is mainly focused on environmental, biological, crude oil and derived samples and more particularly on papers published since 2006 for the first two fields. Other studies in environmental and biological fields have also been reviewed in The Analytical Chemistry of Silicones¹⁶ and Organosilicon Materials¹ but also in a comprehensive review on the analytical aspects of silicones published in 2006.¹⁷

2.1 Atomic spectroscopic methods

These methods can be applied with a decomposition step before silicon determination in order to use it in aqueous matrices or directly by injection of organic phases without initial sample digestion. This preparation step is mainly used for solid samples^83–87 and to limit carbon introduction in the plasma⁶⁹ allowing minimum mass interferences especially with ICP-MS.

2.1.1 Atomic absorption spectroscopy (AAS). GFAAS (graphite furnace atomic absorption spectroscopy) is usually used for silicon determination mainly in environmental and biological matrices with detection limits generally reported below 1 mg kg⁻¹. However, the atomization optimization step is crucial because of possible formation of thermally stable silicon carbide and SiO_(s).

The detection of organosilicon materials in river sediments has been reported by AAS after extraction and concentration.³⁶ More recently, Mukhtar and Limbeck developed an accurate procedure for trace silicon detection in solid environmental samples such as soils or airborne particulate matter using GFAAS.⁸⁸ Based on a preliminary treatment of samples leading to mineralization, and with the use of 20 μL sample injection volume, an instrumental LOD of 52 μg L⁻¹ was obtained, which translated to method detection limits of approximately 0.52 μg m⁻³ when considering collected air volumes.⁸⁸

AAS is also applied in medical fields in order to measure concentrations in tissues, plasma and human blood exposed to breast implants.⁸⁹ Indeed, the concentration of Si in blood of women with silicone gel-filled breast implants was found to be double (33.5/17.1 μg L⁻¹) that a women with no implants.⁹⁰ Lugowski et al. detected Si in blood and urine of individuals exposed to breast implants⁹¹ and in human tissues²⁰ with μg L⁻¹ levels by GFAAS. Detection limit (2SD) of 0.5 mg kg⁻¹ of tissue was obtained by GFAAS after heptane extraction.⁹² GFAAS has also been employed for silicon measurement in biological tissues as well.^93–96 Hornung et Krivan showed that pre-ashing solid samples improved detection limits (0.2 to 0.03 mg kg⁻¹), sample homogeneity and precision compared to direct sampling by this previous technique.⁹⁵ Moreover, a sensitive (LOD of 1.5 μg L⁻¹), simple and accurate method for the routine determination of trace silicon by Zeeman GFAAS has been reported.⁹⁷

The first technique used in metal trace analysis of petroleum products is GFAAS with detection limits usually reported in the range of 10 μg L⁻¹.⁶⁹ Amaro and Ferreira determined silicon determination in naphta (C₄–C₁₅) by GFAAS after sample dilution in toluene.⁸¹ An experimental design has been achieved to determine optimal conditions for Si determination to obtain detection limit of 15 μg L⁻¹.⁸¹ Nevertheless, problems related to the formation of thermally stable silicon carbide (SiC_(s)) and volatile SiO_(s) during pyrolysis by GFAAS can occur.^81,98–100 These constraints can be resolved by using chemical modifiers^83,101,102 or treated Zr tubes⁸⁸ to avoid analyte reactions with graphite.

Finally, several studies in food products^102–104 with detection limits ranging between 7 μg L⁻¹ and 1.8 mg L⁻¹ and industrial applications (LOD of 30 μg kg⁻¹)⁸³ have been carried out by AAS with pre-digestion in order to measure silicon levels.

2.1.2 Inductively coupled plasma optical emission spectrometry (ICP-OES). ICP-OES can be used for Si determination in environmental, biological and petroleum samples. However, detection limits varied from 3 μg kg⁻¹ to 30 µg kg⁻¹ and depending on sample preparation such as acid digestion. In addition, this latter step avoids spectral interferences observed by ICP-MS.

ICP-OES is a useful tool for environmental samples^105,106 such as water¹⁰⁷ and agriculture matrices.¹⁰⁸ For instance, the detection of total silicon has been achieved in rivers and sediments by ICP-OES after an extraction of organosiloxanes in a mixture of petroleum ether/MIBK with detection limits of 100 μg L⁻¹¹⁰⁹ and 10 μg kg⁻¹¹¹⁰ of Si respectively. Masson et al. reported quantification of Si in plant samples after digestion using samples (2 mg) issued from an inter-laboratory test with detection limit of 30 μg kg⁻¹.¹⁰⁵ More recently, silicon concentration was measured by this technique in marine sediments as a function of pH and salinity.¹⁰⁶ The results have shown a reduction of silicon levels while the pH increase.

This spectroscopic method was also applied for silicon determination in biological samples such as urine,^111,112 blood,¹¹² tissues and organs,^84,85 plasma¹¹² and serum.¹¹² Detection limits of 2 mg kg⁻¹ and 3.7 mg kg⁻¹ obtain by ICP-OES for tissue after digestion using 0.1 g of sample have been reported by Hauptkorn et al. and McConnell et al. respectively.^84,85 In 2003, Jia et al. obtained a detection limit of 0.2 mg kg⁻¹ during the analysis of silicone oil extracted in a mixture of toluene/ACN in pharmaceutical matrices by ICP-OES coupled with an ultrasonic nebulization.¹¹³ Low silicon concentrations in foods and soils are determined using the association of a microwave dissolution and an ICP-OES.¹¹⁴ Detection limits have been improved by a factor of 2 with the incorporation of tertiary amines in the acid digestion procedure.¹¹⁴ Silicon measurement was also achieved by ICP-OES after microwave acid digestion of foods and beverages due to the importance of Si in bone formation and connective tissue metabolism.¹¹⁵

In 1988, Carduner et al. determined the presence of silicon in petroleum samples and more particularly in unleaded gasolines by ICP-OES with detection limit of 100 μg L⁻¹ for Si.⁶⁶ Botto reported LOD and RSD precision for Si by ICP-OES using two nebulization systems pneumatic and ultrasonic nebulization (USN).⁸² The use of ultrasonic nebulization improved the detection limit by a factor of 10. Silicon has been detected at 2.7 μg kg⁻¹ in toluene with a USN system and at 27 μg kg⁻¹ with a pneumatic system. Toluene was an excellent diluent for oil and derived samples. Sanchez et al.^116,117 recently worked on silicon compounds determination by ICP-OES in xylene matrices. The effect of silicon chemical form was evaluated for sixteen different silicon molecules with sensitivity varying by a factor of up to 20.¹¹⁶ It was due to the liquid sample introduction system configuration, in particular the spray chamber design. These studies have shown that the application of two introduction systems, heated torch integrated sample introduction system (h-TISIS)¹¹⁶ and demountable direct injection high efficiency nebulizer (d-DIHEN) drastically reduces the influence of the chemical compound form.¹¹⁷ In addition, the response of silicon compounds also depends on the matrix but it less pronounced for high dilution factors.¹¹⁸ Finally, the determination of silicon, iron and vanadium was carried out in petroleum coke by microwave plasma torch atomic optical spectrometry (MPT-OES) after dissolution in nitric acid.¹¹⁹ This method allows a better control of trace elements which is very important when using petroleum coke as an electrode.

2.1.3 Inductively coupled plasma mass spectrometry (ICP-MS). A significant enhancement of detection limits for silicon can be obtained by replacing OES for MS.⁷⁰ Indeed, the performance of ICP-MS allows the detection of very low concentrations levels (sub-ng L⁻¹) in environmental, biological, petroleum products and in industrial samples. However, molecular interferences can also occur especially for silicon at m/z 28 (¹²C¹⁶O⁺ and ¹⁴N₂⁺) but this problem can be solved with the use of collision-reaction cell^87,88,120 or also by using a HR-ICP-MS detection.¹²¹

A procedure using a chromatographic purification coupled to multi collector (MC) ICP-MS has been described to enable dissolved silicon measurement in natural waters.¹²² Silicon isotopic measurement was also achieved by MC-ICP-MS in lake waters.¹²³ The determination of Si by laser ablation (LA) ICP-MS and X-ray fluorescence (XRF) are also compared for airborne particulate matter.¹²⁴ These results emphasized that LA-ICP-MS is the most preferable choice regarding detection limits and matrix limitations.

Several workers have carried out surveys using a microwave digestion before ICP-MS analysis of silicon in biological solid samples.^125–127 Direct determination of Si by a double focusing magnetic sector inductively coupled plasma mass spectrometer without preparation step was reported in serum.¹²⁸ Si measurements were carried out using the major ²⁸Si isotope and with a resolving power of 3000, which is sufficient to avoid spectral interferences from ¹⁴N₂⁺ and ¹²C¹⁶O⁺ during organic product injection. An interlaboratory trial has been achieved by many co-workers for the determination of silicon in biological samples.¹²⁷ Various techniques (ETAAS, XRF, ICP-OES and ICP-MS) were compared considering LOD, RSD and statistical results for each technique and for each matrix.

Due to numerous matrix effects with the introduction of organic samples in ICP-MS, the use of high dilution is recommended.⁷⁰ The detection of silicon in naphta and petroleum matrices after dilution by 10 in toluene has been achieved using a near cold plasma ICP-MS.¹²⁹ Sub-μg kg⁻¹ detection limits^120,129 were obtained but solvent purity and volatility species are pointed out as significant issues.¹²⁰

Takaku et al. have determined trace silicon in ultra high purity water for the manufacture of semiconductors by ICP-MS.¹²¹ A preconcentration procedure was applied, allowing quantification down to sub-ng L⁻¹ of Si. The variation of element concentrations such as B, Si, P and S in steels has a significant influence on the mechanical and physical properties of this material.⁸⁷ Thus, Si determination was achieved by ICP-MS¹³⁰ or DRC-ICP-QMS.^86,87 Detection limits varied from 0.2 μg L⁻¹⁸⁶ and 2 μg L⁻¹⁸⁷ for the introduction of acid solution and precision was better than 6.3%.⁸⁶ These LOD are low enough for analysis of nitric acid steel samples.¹³⁰ Finally, electrothermal vaporization (ETV) ICP-MS was applied to quantify silicon in solid polyamide samples.¹³¹ According to a comparison between other spectroscopic methods, Resano et al. concluded that ETV-ICP-MS and ETAAS have approximately the same performance concerning detection limits in the solid sample (0.3 μg g⁻¹).¹³¹ De Schrijver et al. have shown that the LODs considerably improved with direct solid sampling methods compared to those attainable after dissolution.¹³² ICP provides a very sensitive determination of total silicon but these do not provide structural information if it is not coupled to a separation technique (see chapter 4).

2.2 NMR methods

NMR has the potential to achieve both total silicon determination and molecular identification¹³³ mainly in biological fluids and tissues, environmental and petroleum products. The application of three NMR methods (¹H-NMR, ¹³C-NMR and ²⁹Si-NMR) has been widely reported for silicon determination in the Analytical Chemistry of Silicones¹³⁴ and in Organosilicon Materials.¹⁹ Varaprath et al. have also described numerous works by NMR for the determination of silicon in environmental and biological samples.¹⁷ Nevertheless, ²⁹Si NMR could not detect concentrations below 50 mg kg⁻¹ in these matrices

In 1990, Fux determined the presence of PDMS in extracts of chemicals after a soxlhet extraction with pentane. A quantification limit of 0.1 mg kg⁻¹ for PDMS was obtained by ¹H-NMR.¹³⁵ Other works by ¹H-NMR and ²⁹Si-NMR for determination of PDMS degradation products in animal models and human tissues have been related.^136–138 Indeed, Garrido et al. have shown that silicon migrates from the implant to local and distant sites, and were located in tissues¹³⁶ and blood.¹³⁷ The biodegradation of PDMS and D₄ as model compounds in lymph nodes of rats has also been studied by ²⁹Si-NMR.¹³⁸ In addition to the resonance associated with the PDMS injected, the NMR spectra showed new resonance, compared with lymph nodes control that are attributed to partially hydrolysed polysiloxanes and silica. Moreover, a critical review article of works by NMR has also reported on this problem as well.¹³⁹

The natural abundance of ²⁹Si isotope is only 4.7% and the NMR signal is weak. However, this technique is attractive for silicon speciation because of the wide distribution of ²⁹Si chemical shifts (120 ppm) from silicon nuclei. In the petroleum domain, Carduner et al. determined the presence of Si and identified the chemical nature of the molecules by ²⁹Si-NMR.⁶⁶ Indeed, a resonance was observed at −19.55 ppm that corresponded to octamethylcyclotetrasiloxane (D₄) in unleaded gasoline. Hamilton also reported the detection of PDMS hydrolysis products by ²⁹Si-NMR in the environment.¹³³ The different chemical shifts were specified for silicon compounds (D₃-D₆, HMDS, DMSD…). D₄ chemical shifts fitted the value reported by Carduner et al.⁶⁶

In spite of specified identification of chemical structure and non-destructive technique, NMR is not adapted to trace analysis of several contaminants.¹⁴⁰ Indeed, the ratio of LOD calculated between AAS and ²⁹Si-NMR is approximately 10⁴. Carduner et al.⁶⁶ and Bellama et al.¹⁴¹ also noted that concentrations lower than 60 mg kg⁻¹ of Si in unleaded gasolines and 45 mg kg⁻¹ of Si in environmental samples could not be detect by this technique.

2.3 Other methods

Considering the fact that X-ray fluorescence (XRF) and infra red spectroscopy are not adapted to trace determination of silicon, less attention is paid. However, there were applied in petroleum products, industrial and biological samples with detection limits generally above 10 mg kg⁻¹. The determination of silicon by XRF¹⁴² and IR¹⁴³ was widely reported in The Analytical Chemistry of Silicones. The application of XRF for silicon analysis in heavy oil samples allows quantification with detection limit of 12 mg kg⁻¹ and a reproducibility above 97%.¹⁴⁴ A comparison between XRF and FTIR applied to silicon determination in paper coating has shown that XRF is the more convenient combination between sensitivity and samples preparation.¹⁴⁵ However, FTIR can be possibly used as an alternative of XRF. FTIR has been particularly applied in the determination of silicon in human tissues from breast implant degradation.^146–148

3 Silicon contamination and analytical artifacts

As seen in the previous chapters, the important use of silicon compounds in a wide variety of applications (cosmetics, personal care products and consumer goods) increases the potential contamination of samples during the analysis.^12,13 The purpose of this part is to review the different sources of contamination and the analytical artifacts induced during silicon speciation due to various contamination sources.²¹ A relative estimation of silicon contamination based on publication numbers and pollution concentration is illustrated in Fig. 7. The importance of each source is discussed throughout this chapter. Generally, contamination through human way has to be minimized by taking care of the use of cosmetic and personal care products during Si determination by laboratory personal.²¹ Moreover, numerous authors described contamination problems arising from organosiloxanes present in different sources like spare parts made of silicones in various parts of the analyzer or contamination from different sources during sampling and preparation. Thus, extreme care must be taken to minimize the sources of artifacts and siloxane contaminations for speciation analysis at trace levels.¹⁴⁹

Fig. 7 Relative estimation of silicon contamination based on literature publications in percent.

3.1 Contamination in gas chromatography

Contamination in GC mainly concerned cyclic siloxanes generated by the inlet septa and bleeding column during analysis. According to Fig. 7, the degradation influence of the inlet septa is greater than the column bleeding regarding the siloxane levels. Moreover, the contamination augments with the injection temperature. The presence of septa particles located in the liner also contributes to this pollution.¹⁵⁰ Nevertheless, this phenomenon can be reduced by using specific gas chromatography tools.

Septa are made of pure heavily cross-linked PDMS and of phthalates added as stabilizers.¹⁵⁰ At high temperatures (200–300 °C), PDMS starts to degrade and produces cyclic siloxanes (D₃-D₁₃) (Fig. 4).^58–60 The trimer is reported to be the most abundant product with decreasing amounts of tetramer, pentamer, hexamer and higher oligomers.⁵⁸ In order to minimize the impact of septum bleeding, De Zeeuw recommends several points:¹⁵⁰

• Avoid the use of a septum and use valve injection such as merlin microseal valve^22,151

• Reduce the risk of septum scoring by using tapered needles and a prepierced septum (BTO septum)

• Clean the liner (ultrasonic bath) and replace in a time

• Replace the septum frequently

According to this work,¹⁵⁰ Horii and Kannan have studied five different septa between 100 and 250 °C for injection temperatures.¹²Fig. 8 shows D_n concentrations (μg L⁻¹) as a function of septa used during GC analysis of 1 μL of hexane spiked with an internal standard (M₄Q) at 200 °C. The concentrations of D₄, D₅, D₆ measured with an Agilent advanced green septum are respectively 6, 1.6, 4.6 μg L⁻¹ at 250 °C and 1.7, 0.6 and 0.3 μg L⁻¹ at 200 °C. It corresponds to a drastic reduction (79%) of D_n concentrations. The use of BTO septum at 200 °C allows a balance between low levels of D_n and a correct vaporization of compounds (Fig. 8). Overall, the background of D₆ in blanks is at least 20 times lower than the lowest concentrations found in consumer products.¹² Considering this work, Sparham et al. have applied an injection temperature at 150 °C to reduce contamination by septum.¹⁵²

Fig. 8 Instrumental background levels of D₄, D₅, D₆ in n-hexane injected into several septa of GC-MS at 200 °C (Adapted from Horii et Kannan, 2008).¹²

Septum particles can also be deposited in the liner from 200 °C to 300 °C and caused contamination.¹⁷ One inlet septum particle into the sample extract is almost half of the total amount stationary phase coating of the capillary GC column (a few milligrams for a VF-5 MS).^12,150 Consequently, a method is proposed to determine if the septum produces cyclic siloxane compounds based on the simulation of different splitless injection time with variation from 0 to 20 min.¹⁵⁰

The generation of these molecules also originated from column bleeding because most of the capillary columns are covered with PDMS. Horii et al. examined cyclic oligomer backgrounds of two low bleed columns, a DB-5 MS (5% phenyl and 95% PDMS) and a DB-1 MS (100% PDMS).¹² Another column (DB-XLB) has also been studied by Kala et al. in 1997 to minimize hexamethylcyclotrisiloxane (D₃) levels⁵¹ in blanks. The results suggested that the release of organosiloxanes from a low bleed column is minor in comparison with the release from septa (Fig. 7). The fact that D_n contamination mainly originates from septum degradation has been confirmed by De Zeeuw¹⁵⁰ and Wang et al.¹³ For example, DB-5 MS and DBWAX (phase without silicon) columns produce similar background levels of siloxanes during blank analysis. However, Varaprath et al. have also focused their research on the analytical artifacts related to siloxane analysis. For example, D₄ can be generated by interactions between water contained in biological or environmental samples and PDMS stationary phase in GC.¹⁵³ Higher cyclic siloxanes can also be produced at trace levels.

3.2 Mass spectrometry detection and artifacts

In 1982, Ende and Spitleller reported different contaminations in Mass Spectrometry.¹⁴⁹ Several elements present in laboratory (syringe, lubricants, flasks, pump fluids, septa…) contained silicones.^17,149 For instance, Carter et al. suggested that plastic pasteur pipettes contain a silicon lubricant with a molecular weight of approximately 16 500 g mol⁻¹.¹⁵⁴ General precautions to limit contamination are described but even with the highest degree of precaution, these authors suggest that analysts are not able to exclude contamination in trace silicon analysis.¹⁴⁹ Indeed, they can only minimize the risk of contamination.

In association with GC, the degradation products formed by column stationary phase decomposition are similar to the products generated by septa degradation.¹⁵⁰ These compounds induce the formation of artifact ions (m/z 207, m/z 281 and m/z 355) in MS that correspond to major fragment ion of our interest compounds as D₃ (222-CH₃), D₄ (296-CH₃) and D₅ (370-CH₃).¹⁵⁰ Other ions m/z 73 (Me₃Si⁺), m/z 147 (Me₃SiOSiOMe₂)⁺m/z 149 (plasticizer), m/z 295 (silicones with M<400) can also appear as contamination sources in the background¹⁷ although, low bleeding and MS quality columns are recommended.¹⁵⁰

3.3 Other contaminations, storage and conservation

Several authors have focused on the storage and conservation of organosilicon compounds such as silanols and silanediols because of their high reactivity.²² Sample storage in vial septa made of silicones must be avoided and silicon compounds (DMSD) need storage in plastic containers to eliminate condensation enhanced by alkaline surface.

Various authors focused their works on the contamination induced by the use of silicone/PTFE vial septa.^155–157 Chambers et al.¹⁵⁵ and Wang et al.¹⁵⁷ demonstrated that silicone/PTFE septa generate D_n compounds when they are in direct contact with the sample. Chambers et al. observed the generation of cyclic siloxanes (D₃-D₇) during MTBE (methyl tertiary butyl ether) analysis in blood at ng L⁻¹ levels by gas chromatography coupled to mass spectrometry (GC-MS) in single ion monitoring mode (SIM).¹⁵⁵ Moreover, the results obtained by Wang et al. showed that sample analysis must be made immediately after they are introduced in vials with silicone/PTFE septa¹⁵⁷ in order to avoid the contamination by D_n. On the other hand, Pattinson et al. tested nine vial septa with different solvents by GC-MS.¹⁵⁶ They concluded that septa which have a PTFE layer minimize silicon contamination. Thus, vials with butyl/PTFE caps used by Sparham et al. have shown a reduction of D_n generation and a minimization of the sample volatilization more observed with PTFE septa.¹⁵²

Storage and conservation play an important role during trace analysis. Silicon materials, such as PDMS, have a great affinity for glass and are often adsorbed on the surface.¹⁸ Thus, it is more safe to keep them at lower temperatures in teflon or plastic containers (preferentially high density polyethylene).¹⁷ This precaution is particularly true for silanols that condense and siloxanes that react in the presence of alkaline surface, acids or strong bases.¹⁸ Indeed, glass surfaces, moisture, temperature and acidity will improve condensation.²² Several authors suggest to store samples in polypropylene vials before their analysis.^12,31,158

Varaprath and Lehmann have also studied the conservation and stability of the dimethylsilanediol (DMSD) by ²⁹Si NMR and HPLC (high pressure liquid chromatography) at different concentrations.²² This compound is unstable even in pure form¹⁵⁹ and must be stored in a freezer between 4 and 8 weeks in plastic container with a drying agent in order to avoid polymerization and the presence of dimerdiol at high concentration.²² At a lower concentration (100 mg L⁻¹), HPLC results confirmed that DMSD is stable for one year.

4 Hyphenated techniques for Silicon speciation

Several hyphenated techniques (Fig. 6) have been developed for Si speciation using chromatographic methods. These separation techniques are based on different physico-chemical properties, such as volatility, molecular size, degree of aromaticity or polarity. This chapter is mainly focused on the combination of chromatographic methods (LC, GC and SFC) with different detectors reported in Fig. 6. However, other techniques such as mass spectrometry have been employed for Si speciation.

4.1 Liquid chromatography

For silicon speciation, LC has been employed in two separation techniques: high pressure liquid chromatography (HPLC) for LMW compounds and size exclusion chromatography (SEC) for HMW molecules (Fig. 6). HPLC can be performed in normal phase (NP) or reverse phase (RP). For silicon speciation, HPLC applications by reverse phase are more important than by normal phase.

4.1.1 High pressure liquid chromatography (HPLC). HPLC analysis enables the separation of polar and non polar compounds that do not elute under normal GC analysis conditions such as low volatile, highly polar or thermally unstable molecules.¹⁷ For silicon speciation, reversed-phase HPLC is carried out using a C₁₈ (octadecylsilica) stationary phase and elution with a polar solvent (methanol, acetonitrile, water).

Several researchers have employed RP-HPLC for LMW silicon molecules in environmental and biological samples using classical detectors (UV, RI, radioisotope)^160,161 or ²⁹Si NMR¹⁶² to provide molecular identification. However, most of the works have been performed by the coupling of RP-HPLC and atomic detectors (ICP-OES and ICP-MS) allowing structural information with the LC separation and selective detection (Fig. 6).^{31,154,158,163–166} Generally, a considerable improvement of LOD is observed when ICP-OES (20–500 μg L⁻¹) is replaced by ICP-MS detection (0.1–4 μg L⁻¹). However, according to Ebdon et al., ICP-OES is a more attractive detector for silicon speciation when compared to ICP-MS because mass interferences (m/z 28) are deleted.¹⁶⁶ RP-HPLC analyses for Si speciation are summarized in Table 2.

Table 2 HPLC methods for silicon speciation

Matrices	Molecules	Solvent	Injection volume/μL	Column	Mobile Phase	Flow rate/mL min^−1	Detection	LOD	Ref.
Soil	Dimethylsilanediol	THF	—	C18	H2O (100)	—	Refractive index	—	22
				25 cm × 4.6 mm × 5 μm	H2O (100) to ACN (100)
					ACN (100) to H2O (100)
Biogas	Inorganic silicon	H2O	100	C18	ACN/H2O (20/80)	0.3	ICP-OES		31
	Dimethylsilanediol			25 cm × 2 mm	ACN/H2O (5/95)			500 μg L^−1 of Si
	1,3-tetramethylsilanediol				ACN/H2O (60/40) for Flushing			30 μg L^−1 of Si
	Trimethylsilanol
Environmental	Methylsilanediols	Xylene	25	C18	ACN/H2O (1 : 10) to ACN/H2O (4 : 10)	0.2	ICP-OES	40–150 μg L^−1 of Si	158
				25 cm × 2 mm
Biological	L2 D5 and its metabolites	mobile phase	—	C18	H2O (100)	—	Radioisotope	—	161
				25 cm × 4.6 mm × 5 μm	H2O (100) to ACN (100)
					ACN (100)
					ACN (100) to H2O (100)
	D4 and its metabolites	mobile phase	—	C18	H2O (100)	—	Radioisotope	—	160
				25 cm × 4.6 mm × 5 μm	H2O (100) to ACN (100)		Refractive index
					ACN (100) to H2O (100)
Medical	Inorganic silicon	mobile phase	100	C18	MeOH–H2O (30/70)	1	ICP-OES	100 μg L^−1 of Si	166
	1,3-tetramethylsilanediol			25 cm × 2 mm				400
	L2							500
	Inorganic silicon	mobile phase	100	C18	MeOH–H2O (20/80)	0.15	ICP-MS	0,1 μg L^−1 of Si	154
	Dimethylsilanediol			25 cm × 2.5 mm × 6 μm				4
	1,3-tetramethylsilanediol							4
Polymer	L2-L5, D3-D5	AcOEt	5	C18	MeOH–H2O (95/5) to MeOH (100)	1	ICP-OES	20–50 μg L^−1 of Si	163
				15 cm × 4.6 mm × 5 μm
	L2	Acetone	5	Carbosphere 30 DS	ACN/Acetone (80/20) 30–130 °C at 2 °C min^−1 130 °C (8 min)	0.20	ICP-OES	40 μg L^−1 of Si	164
	PDMS
	Hexamethyldisilane	ACN	60	C18	ACN/CDCl3: (9 : 1) to ACN/CDCl3 (1 : 9)	0.5	^1H NMR	—	162
	L3 and L5	CDCl3		15 cm × 4.6 mm × 5 μm			^29Si NMR
	PDMS
Organic matrices	D4	Xylene	—	C18	MeOH–H2O (7 : 10) to MeOH(10)	1	ICP-OES	500 μg L^−1 of Si	165
		Hexane		30 cm × 10 μm

---

Varaprath et al. detected several metabolites of hexamethyldisiloxane (L₂) and decamethylcyclopentasiloxane (D₅)¹⁶¹ and also metabolites of D₄¹⁶⁰ in animal urine by RP-HPLC with a radioisotope detection. Metabolites were eluted on a C₁₈ column using a ACN/H₂O mobile phase. The main structures were confirmed by synthesizing ¹⁴C-labeled standards and GC-MS was applied in order to further confirm their identities.^160,161 DMSD was detected by refractive index detector during PDMS degradation in soils.^22,167 These previous detectors provide a non-destructive technique but without the sensitivity to allow trace analysis.¹⁷

As previously mentioned, Si speciation in the environment by RP-HPLC-ICP-OES has been reported (Fig. 2). Firstly, the detection and the linearity response of D₄ in xylene and hexane matrices have been achieved with detection limit of 500 μg L⁻¹.¹⁶⁵ Similarly, Biggs et al. carried out silicon molecules separation (L₂-L₅ and D₃-D₅) with improved LODs ranging from 20 and 50 μg L⁻¹ of Si with this combination.¹⁶³ According to gradient and column optimization, elution on a C₁₈ column (15 cm × 4.6 mm × 5 μm) with a gradient from MeOH–H₂O (90/10) to MeOH (100) over 8 min produces the best results.¹⁶³ Nevertheless, chromatographic separation observed respectively between D₄ and L₃ and between D₅ and L₄ is not optimal. The same authors have also indicated the improvement of thermal gradient by RP-HPLC-ICP-OES for PDMS solution in acetone with a ACN/Acetone gradient.¹⁶⁴

In 1994, the speciation of silanediols in water, sludge extracts and soils was carried out by RP-HPLC-ICP-OES with detection limits of 40 μg L⁻¹ of Si for trimerdiol and 150 μg L⁻¹ of Si for monomer and dimerdiol.¹⁵⁸ It was also shown that low injection volume and use of xylene minimize the instability of the plasma to organic solvents. Grumping and Hirner performed the separation of two silanediols and trimethylsilanol (TMS) in leachate samples by the same hyphenated technique.³¹ Detection limits of 30 μg L⁻¹ of Si for TMS and 500 μg L⁻¹ of Si for DMSD were calculated for 100 μL injection volume. Additionally, a modification of eluent by a more polar gradient (Table 2) allowed the separation of the DMSD and the silicate.³¹ Dorn and Kelly¹⁵⁸ obtained DMSD lower detection limits than Grumping and Hirner³¹ because of the different gradient application and the use of a special ICP interface and micro HPLC system.

More recently, Ebdon et al. compared the performance of RP-HPLC with a radial ICP and an axial ICP for polar silicon compounds.¹⁶⁶ Generally, it is well known that detection limits are improved using axial ICP compared to radial ICP.¹⁶⁸ However, radially viewed ICP-OES gave similar detection limits (100, 400 and 500 μg L⁻¹ of Si for inorganic silicon, TMSD and L₂ respectively) to axially viewed ICP-OES with improved chromatographic peak reproducibility for all three compounds.¹⁶⁶ In 2004, Carter et al. replaced the ICP-OES with HR-ICP-MS coupled to RP-HPLC for Si speciation.¹⁵⁴ Consequently, using similar conditions to previous study,¹⁶⁶ an improvement of detection limits for inorganic silicon compound with a 1 : 1000 ratio and for TMSD with a 1 : 100 ratio was indicated (Table 2).¹⁵⁴ These results show that ICP-MS provides better detection limits than ICP-OES for Si speciation when mass interferences are resolved. Finally, reversed-phase LC-NMR with ¹H and ²⁹Si has been developed using a gradient from ACN/CDCl₃ (9 : 1) to ACN/CDCl₃ (1 : 9) in order to determine several organosiloxane structures.¹⁶²

4.1.2 Size exclusion chromatography (SEC). This technique, also referred to as gel permeation chromatography (GPC), is commonly used for unravelling HMW molecule behaviour in environmental, biological tissues and industrial applications.¹⁶⁹ SEC separates molecules on the basis of their size, or more precisely on their hydrodynamic volume. It is employed in silicon speciation for PDMS degradation in environmental conditions by using SEC with a refractive index detector^28,41 or an ICP technique with detection limits reported below sub-mg kg⁻¹.^42,170 Several researchers have also combined SEC and mass spectrometry for identification of silicon polymers.

Hausler and Taylor have achieved PDMS separation by SEC-ICP-OES with detection limits ranging between 0.03 to 3 mg L⁻¹ of Si.^171,172 This hyphenated technique was applied in water and sludge extracts with LOD of 160 μg L⁻¹ of Si.¹⁵⁸ PDMS separation with molecular mass between 550 and 500 000 g mol⁻¹ using THF and xylene as mobile phases has been established. Xylene was chosen because it enables to easily maintain a stable plasma and allows the solubilisation of most of the PDMS polymers. Nevertheless, using THF iso-response as a function of molecular mass and upgrade in sensitivity by a factor 4 are observed. This is due to the better solubility of PDMS polymer in THF, particularly for high weight mass polymer.¹⁵⁸

Carter et al. reported a similar separation of three PDMS polymers (162, 1500 and 16 500 g mol⁻¹) by SEC-ICP-MS with detection limits of 12, 26 and 30 μg L⁻¹ of Si respectively.¹⁵⁴ Extraction efficiencies of these compounds from spiked human plasma were performed with xylene because of its relatively low vapour pressure (plasma stability ¹⁵⁸) and compatibility as an elution solvent for the SEC separation.¹⁵⁴

Maziar et al. have studied HB-PDMS (α,ω-bis(4-hydroxybutyl) polydimethylsiloxane) behaviour by SEC-MALDI-TOF/MS in order to obtain detailed information during synthesis.¹⁷³ A comparative study between automated SEC-MALDI-TOF/MS and on-line SEC-ESI-TOF/MS for PDMS characterization demonstrated an improvement of chromatographic resolution and an enhancement in low molecular weight polymer (M<550 g mol⁻¹) by this latter coupling.¹⁷⁴ In opposition, the first technique effectively reported the HMW oligomers and underestimated the LMW oligomers.¹⁷⁴

4.2 Gas chromatography

Gas chromatography has been widely used for Si speciation (Fig. 2). An enhancement in separation compared to HPLC based on the volatility and the polarity of species can be achieved by GC. Fig. 6 summarizes various detectors for silicon reported in the literature. GC is preferred to LC because of higher resolution and lower detection limits. For instance, chromatographic resolution between silicon molecules like D₄ and L₃ is better by GC-ICP-OES³¹ than by LC-ICP-OES.¹⁷⁵ GC chromatography applied for Si speciation in environmental (Table 3), biological and industrial samples (Table 4) are presented. Most of the studies is carried out on non polar columns (95% methylPolysiloxanes-5% Phenyl) with 1 μL injection volume in splitless mode.

Table 3 Gas chromatography methods for silicon speciation in environmental samples

Matrices	Molecules	Solvent	Injection	Column	Detection	LOD; Si Eq.^a	Ref.
a Si Eq.: Silicon Equivalent. b Estimated.
Soil	D4-D6	THF	1 μL	HP5	MS SIM	1 μg L^−1; 0.4 μg L^−1 of Si	22
	Silanediols (1–5)	Acetone	250 °C	30 m × 0.25 mm × 0.25 μm		0.1–1 mg L^−1; 0.04–0.4 mg L^−1 of Si
	Silicon degradation products	THF	Merlin system	HP5	MS Full Scan	—	151
		Acetone	2μL 250 °C	30 m × 0.25 mm × 0.25 μm
		Hexane
	Dimethylsilanediol	THF	1μL	HP5	MS SIM	—	33
			250 °C	30 m × 0.25 mm × 0.25 μm
	Dimethylsilanediol	THF	2μL	DB1	MS SIM	—	35
	silanols	Acetone	250 °C	60 m × 0.32 mm × 0.25 μm
	Dn
Water	D5	Hexane	3 ml in HS	DB-FFAP	MS SIM	3 ng L^−1; 1.1 ng L^−1 of Si	152
			220 °C	30 m × 0.25 mm × 0.25 μm
Air	D3-D5	—	—	HP5-MS	MS SIM	—	179
				30 m × 0.25 mm × 0.25 μm
	D3-D4	—	180 °C	200 f t OV-17	MS Full Scan	—	178
	L5			200 ft OV-101
	Tetramethylsilane			400 ft OV-101
	Alkoxysilanes	Heptane	1 μL	DB5	FID	1–5 mg L^−1; 0.12–0.62 mg L^−1 of Si	56
				30 m × 0.32 mm × 0.25 μm
Air, water, soil, sediments, biota	D3	Hexane	1 μL	CP-Sil8CB	MS SIM	50 μg L^−1; 18.9 μg L^−1 of Si	8
	D4-D6		200 °C	30 m × 0.25 mm × 0.5 μm		5 μg L^−1; 1.9 μg L^−1 of Si
	L2-L5					0.3–0.5 μg L^−1; 0.11–0.18 μg L^−1 of Si
Biogas	L2-L4, D3-D6 et Trimethylsilanol	Pentane	200 °C	RTx-1	AED	9.5 μg L^−1 of Si	32
				47 m × 0.32 mm × 1.5 μm	MS Full Scan	—
Waste sludge	L2 et D3-D6	Hexane	1 μL	VF-1MS	FID	—	29
			125 °C
Waste	D3-D5; L2-L4	Pentane	5 μL	HP1	ICP-OES	0.1 μg L^−1 of Si^b	175
	Trimethylsilanol					—
	D3-D6	Hexane	1μL	SE-54	FID	—	30
			240 °C	50 m × 0.32 mm × 0.25 μm	MS Full Scan

---

Table 4 Gas chromatography methods for silicon speciation in biological and industrial samples

Matrices	Molecules	Solvent	Injection	Column	Detection	LOD; Si Eq. ^a	Ref.
a Si Eq.: Silicon Equivalent. b Estimated.
Biological	D4	Tetrahydrofuran	Merlin system 2 μL 250 °C	HP5-MS	MS SIM	1 μg L^−1; 0.4 μg L^−1 of Si	153
				30 m × 0.25 mm × 0.25 μm
Blood, Plasma	D3-D6	Hexane	—	HP5	MS SIM	2 μg L^−1; 0.76 μg L^−1 of Si	50
			200 °C	30 m × 0.25 mm × 5 μm
	D4	Tetrahydrofuran	2 μL	ND5-MS	MS SIM	4.9 μg kg^−1; 1.9 μg kg^−1 of Si	11
			250 °C	30 m × 0.25 mm × 5 μm
Tissues	D3-D6	Hexane	—	HP5	MS SIM	—	52
			200 °C	30 m × 0.25 mm × 5 μm
Breast Implants	PDMS	Ethyl acetate	1 μL	DB1-HT	AED	80 μg L^−1 of Si	53
	Dn		320 °C	30 m × 0.32 mm × 0.2 μm	MS SIM	100 μg kg^−1; 38 μg kg^−1 of Si
	Dn and Ln	Ethyl acetate	—	Extra Low bleed DR-XLE	MS SIM	—	51
					AED	—
Medical	Silylated alcool	Pyridine	1 μL	FS-Suprem-5	ICP-MS	3 μg L^−1 of Si ^b	185
				25 m × 0.25 mm × 0.25 μm
Consumption products	D4-D7	Ethyl acetate	1 μL 200 °C	Rxi5-MS	MS SIM	117 μg kg^−1; 44 μg kg^−1 of Si	12
	L5-L14	Hexane		30 m × 0.25 mm × 0.25 μm		18 μg kg^−1; 6.8 μg kg^−1 of Si
	D3	Methanol	280 °C	DB5-MS	MS SIM	120 μg kg^−1; 45.3 μg kg^−1 of Si	13
	D4-D6	Acetone hexane		30 m × 0.25 mm × 0.25 μm		80 μg kg^−1; 30.2 μg kg^−1 of Si
Polymer	D3-D8	Toluène	1 μL	MDN-5S	MS Full Scan	—	184
			250 °C	30 m × 0.25 mm × 0.25 μm
	Dn et Ln	Dichloromethane chloroform	200 °C	DB5	MS Full Scan	—	177
				30 m × 0.25 mm × 0.25 μm
	D3-D10	—	—	CP-Sil 8	MS Full Scan	—	183
				30 m × 0.25 mm × 0.25 μm
	Degradation resin products	Tricholormethane	250 °C	Anabond 225	MS Full Scan	—	182
				50 m × 0.32 mm × 1 μm
	D3-D6	—	200 °C	100% methylsilicone	MS Full Scan	—	181

---

4.2.1 GC-FID. This coupling was mainly used during workers exposure in the silicon industry and in biogas samples but it is not very adapted to trace analysis due to poor detection limits. Firstly, three alkoxysilanes have been detected by GC-FID using heptane as dilution solvent with detection limits ranging from 1 to 5 mg L⁻¹ in order to carry out worker exposure to silicon compounds.⁵⁶ More recently, the determination of cyclic siloxanes (D₃-D₆) in waste samples has been performed by GC-FID after a successful extraction by XAD-2 resins using hexane.³⁰ Note that the identification is possible by matching the retention time with analytical standards. However, completed identification was achieved by GC-MS.³⁰ After a liquid–liquid extraction with hexane, Dewil et al. have separated D₄ and D₅ by GC-FID on a non-polar column in waste sludge.²⁹ Excellent correlation coefficients (R² > 0.999) were obtained up to 1000 mg L⁻¹. Popat et al. have also identified D₄ in biogas with detection limit of 1 mg m⁻³.¹⁷⁶

4.2.2 GC-MS. According to Fig. 2, GC-MS is the most hyphenated technique employed for Si speciation in environmental, biological and industrial matrices. Mass spectrometry (MS) detection can be mainly performed in full scan (FS) mode especially for qualitative analysis and in SIM (selected ion monitoring) mode for trace quantitative measurement (Table 3 and Table 4). GC-MS is one of the most sensitive technique when it is carried out using the latter mode¹⁷ with detection limits generally ranged between 1 and 100 μg kg⁻¹. Reactive polar compounds such as silanols and silanediols caused analysis problem which could be solved by CI (chemical ionization) use or derivatization procedures.

D_n (D₄-D₆) and silanediols (mono-pentamer) were extracted from soils with polar solvent (THF, acetone) and identified using SIM mode via two ionization methods.²² Detection limits of 1 μg L⁻¹ with a signal to noise ratio of 5 for D_n and between 100 μg L⁻¹ and 1 mg L⁻¹ for diols have been calculated. In order to differentiate these compounds which have the same fragmentation patterns in their mass spectrum, some authors^22,177 have preferred to work using chemical ionization (CI) instead of electron impact ionization (EI).

Siloxanes lose a methyl group (M-15) under EI conditions. On the other hand, silanediols (excepted DMSD and dimerdiol) lose a water molecule (18) in addition to a methyl group (M-33), trimer, tetramer and pentamer diols have a mass difference of 18 with D₃, D₄, D₅ respectively. Thus, their mass spectrum is similar (M-15).²² Although these compounds can be distinguished by their retention time in GC, CI helps to identify these molecules. For instance, the identification of D₄ (m/z 314) and tetramerdiol (m/z 332) with a mixture of reagent gas CH₄:NH₃ (90 : 10) [M + NH₄]⁺²² and D_n compounds using isobutane [M + H]⁺¹⁷⁷ were carried out.

Other VMS (D₃-D₅, L₅ and tetramethylsilane) were detected by GC-MS in ambient atmospheric extracts.^178,179 More recently, D₅ analysis was performed in river water and treated waste water by hexane extraction and head space (HS) coupled to GC-MS SIM using a DB-FFAP column.¹⁵² The use of internal standard in SIM mode (m/z 73, m/z 267, m/z 355 for D₅ and m/z 360 for ¹³C₅-D₅) provides a quantification limit of 10 ng L⁻¹ for D₅ in water. In 2005, concentrations of cyclic and linear oligomers were reported by GC-MS SIM in the nordic environment (water, sediments, air, sludge, biota…) of six countries.⁸ Detection limits have been calculated for various matrices after an hexane extraction. For instance, detection limits for D₃, D₄-D₆ and L₂-L₅ were found to be 50, 5 and between 0.3 and 0.5 μg kg⁻¹ in biota samples respectively.

Analysis of polar silicon compounds, such as silanols or silanediols may require derivatization by sylilating agents before GC analysis due to their instability (see chapter 3).^17,22 Basically, if stationary phase contents siloxanes, the formation of reactive hydrogen atoms can appear in presence of moisture¹⁷ and promotes silanol condensation into the GC column.¹⁸⁰ Consequently, this reaction compromises severely the sensitivity and degenerates the peak shapes.²² Several workers have capped OH functions of the molecules with trimethylsilyl groups,¹⁸⁰ by usually using BSTFA (bis(trimethylsilyl)-trifluoroacetamide).¹⁵¹ However, Varaprath et al. have shown that species like DMSD can be analyzed without derivatization reaction.²² The procedure described a 1 μL BSA (bis(trimethylsilyl)acetamide) injection volume into GC column before analysis in order to deplete all reactive surfaces.²² In opposition, Lehmann et al. have achieved the direct detection of DMSD, silanols and D_n without derivatization step by GC-MS SIM during PDMS degradation in soil.^33,35

Cyclic and linear compounds were also analyzed by GC-MS in FS mode for qualitative analysis and in SIM mode for quantitative study in biological matrices. After extraction by ethyl acetate and THF, detection limits of 100 μg kg⁻¹ for tissue⁵³ and 5 μg kg⁻¹¹¹ for plasma samples have been respectively reported by SIM mode. Several workers quantified D_n in blood and plasma using a widely internal standard M₄Q (tetrakis(trimethylsiloxy)silane) for Si speciation.^11,12,50,153 Extraction yields above 90% with THF and a 2μL injection volume gave access to very low detection limits around 1 μg L⁻¹.¹⁵³ In addition, linearity range between 1 to 16 μg L⁻¹ of D₄ compared to internal standard was obtained. Considering the same analytical method, Flassbeck et al.⁵⁰ calculated detection limit of 2 μg L⁻¹ after an hexane extraction of D_n from blood and plasma, which is similar to the values reported in the literature.^11,22,153

As previously mentioned in chapter 1, silicon molecules are used in consumer products. Horii and Kannan measured high VMS (D_n and L_n) levels in 76 cosmetic and personal care products by SIM mode.¹² After an extraction by a mixture of ethyl acetate–hexane, 1 μL of analyte injected in a Rxi-5 MS column at 200 °C has allowed the acquisition of detection limits ranging between 18 (L_n) and 120 μg kg⁻¹ (D_n). Details in contamination problems were also given (see chapter 3 for more details).¹² More recently, cVMS have been analyzed in 252 consumer products by GC-MS (fragrances, hair care products, antiperspirants, lotions…)¹³ using similar conditions to Horii and Kannan.¹² Detection limits between 120 μg kg⁻¹ (D₃) and 80 μg kg⁻¹ (D₄-D₆) and calibration curve ranging from 50 μg kg⁻¹ to 10 mg kg⁻¹ were reported.¹³ According to a comparison between D₅ and D₆ levels in skin lotions by Horii and Kannan¹² (35.3 mg g⁻¹ and 6.3 mg g⁻¹) and Wang et al.¹³ (47.3 mg g⁻¹ and 6.5 mg g⁻¹), these studies are completely in agreement.

GC-MS has also been applied for Si speciation mainly in FS mode for qualitative analysis. Silicon compounds (siloxanes and chlorosilanes) were separated by GC and identified by MS in EI and in CI for silicon rubbers and silicon resins respectively.^181,182 Furthermore, D_n characterization was performed by GC-MS when PDMS was submitted to corona discharges.¹⁸³ Wacholz et al. have also analyzed a mixture of cyclic and linear siloxanes by GC-MS and GC-FTIR.¹⁷⁷ More recently, thermal stability of polysiloxanes has been studied by HS-GC-MS in toluene.¹⁸⁴ These results have shown the presence of D_n, phthalates and ethylhexanoic acid. In addition, Hall et al.¹⁸⁴ have reported a non-linear response for D₅ and L₅ contrary to Sparham et al. researches.¹⁵²

4.2.3 GC with atomic detection. Contrary to GC-MS when it is performed in SIM mode, the combination between GC and atomic detector such as AED, ICP-OES and ICP-MS are relatively scarce even if providing sub μg L⁻¹ detection limits and structural information (Fig. 2). According to our own knowledge, GC-AED is the most reported coupling with atomic detection^32,51,53 and only two papers have been published in Si speciation by GC-ICP-OES in biogas¹⁷⁵ and by GC-ICP-MS in medical field.¹⁸⁵ However, this hyphenation, mainly GC-HR-ICP-MS appear as a versatile analytical tool for silicon speciation. Indeed, the high resolving capacity of GC and the high sensitivity, selectivity and multi-elements of ICP-MS have made the combination more efficient and attractive for speciation analysis of different elements in complex matrices, such as environmental, biological and petroleum and derived products.^186–189 For silicon speciation, only one application of GC-ICP-MS was carried out for silylated alcohol standard solutions in medical field¹⁸⁵ and its use has never been extended to real matrices.

In 1997, a novel and highly sensitive method for detection, quantification and characterization of LMW siloxanes in biological tissues by GC-AED and GC-MS for further identification has been reported.⁵³ After an extraction by ethyl acetate, detection limit of 100 μg of Si per kg of tissue was calculated by GC-AED at 251.6 nm using a DB-1 column. Similarly, Lykissa et al. identified cyclic compounds (D₃-D₇) and L₇ in silicon gel implants⁵¹ with extra low bleed column in order to minimize Si contamination (see chapter 3).

The development of canister sampling and GC-MS-AED was performed by Schweigkofler and Niessner for the determination of siloxanes (D_n and L_n) and silanols.³² Detector signals obtained are linear over more than 4 orders of magnitude (R²>0.99) and detection limit of 9.5 μg L⁻¹ of Si was determined in pentane. After thermodesorption and analysis, L₂-L₄, D₃-D₆ and TMS are the major compounds found in biogas with detection limit of 342 ng m⁻³.³² As mentioned in section 1.3, LOD depend on sample extraction and concentration before analysis. Consequently, the comparison of sensitivity between two different techniques is accurate in biogas analysis especially considering collected sampling volume.

Grumping et al. identified TMS, L₂ and cyclic oligomers (D₃-D₅) in biogas sample by LT-GC-ICP-OES.¹⁷⁵ For example, VMS (L₂, D₃-D₅) were measured at concentrations ranging between 0.1 and 1.1 μg of Si per litre of waste water samples. In addition, resolution between L₂ and TMS was non-optimal using a non-polar column (100% PDMS). Different responses between compounds of interest have also been observed due to molecule volatility or condensation phenomena.¹⁷⁵ According to Grumping et al., the choice of ICP-OES avoids mass interferences (m/z 28) occurring in ICP-MS when collision/reaction chamber is not available (previously mentioned in section 2.1.3). The second study concerns the application of GC-HR-ICP-MS to silylated alcohols quantification.¹⁸⁵ Sylilation was achieved by reaction between four alcohols (C₄–C₇) with N-methyl-N-trimethylsilyltrifluoroacetamide (MSTFA) in pyridine. Medium resolution (m/Δm = 4000) gives access to the resolving of mass interferences between ²⁸Si⁺, ¹²C¹⁶O⁺ and ¹⁴N₂^+121,154,185 in order to obtain detection limit of 0.1 μmol L⁻¹, which correspond to 3 μg L⁻¹ of Si for C₄ silylated alcohol.¹⁸⁵ Nevertheless, column bleeding can limit these detection limits and contributes to the background pollution.¹⁸⁵

4.3 Supercritical Fluid Chromatography (SFC)

Silicon is dispersed in very low levels for several fields such as environment,²² biology⁵⁰ and petroleum and derived products.^70,80 SFC method is not very adapted to trace analysis and the combination between SFC and detectors like ICP-OES, FID or MS don't lead to sufficient sensitivity. Indeed, SFC coupled to an atomic detector (ICP-OES) was achieved to separate various siloxanes with detection limit of 57.9 mg kg⁻¹ of Si.¹⁹⁰ Excellent separation of PDMS with a molecular mass of 2000 g mol⁻¹ has been reported by SFC-FID.¹⁹¹ This method was applied with SFC/MS for the determination and the identification of cyclic siloxanes in technical silicon oils and rubbers.¹⁹²

4.4 Other mass spectrometry techniques

Without chromatographic separation, mass spectrometry avoids contamination due to the GC parts (see chapter 3) and is applied both for LMW and HMW silicon molecules (Fig. 6). These techniques were mainly performed in environment (soil¹⁹³ and biogas^194,195) with variable sensitivity depending on the sample preparation (2 μg m⁻³¹⁹⁵ and 50 mg kg⁻¹¹⁹³) and in polymer characterization (FT-ICR/MS and TOF/MS) usually as qualitative analysis for industrial applications.

The determination of D₄ and D₅ in biogas by APCI-MS/MS without extraction or prior chromatographic separation was a primer in Si speciation.¹⁹⁴ Badjagbo et al. developed and validated a sensitive (4 < LOD (μg m⁻³ of air) < 6) and selective method for direct analysis of siloxanes in gaseous matrices.¹⁹⁴ Indeed, direct MS detection completely avoided background contamination from GC systems (see chapter 3) and allows the direct distinction between VMS and siloxanediols previously discussed (section 4.2.2) due to the soft APCI ionization [M + H]⁺.¹⁹⁴ More recently, Badjagbo et al.¹⁹⁵ quantified D₄ and D₅ with an internal deuterated standard hexamethyldisiloxane (HMDS-d18) using the previous analytical method.¹⁹⁴ The use of HMDS-d18 provided effective signal compensation of D₄ and D₅ and improved the sensitivity and reliability of these compounds in biogas.¹⁹⁵ According to Badjagbo et al., a detection limit around 2 μg m⁻³ has been obtained,¹⁹⁵ which is 500 times better than that reported for a GC-FID¹⁷⁶ and much more sensitive than that reported in a recent study by a microcantilever array sensor.⁴⁶ The latter method, recently developed by Long et al., provided linear and cyclic siloxanes analysis in biogas with detection limits of 17 μg L⁻¹ for D₅, corresponding to 257 μg m⁻³.⁴⁶ After possible adsorption or deposition of silicon in the environment (section 1.2), a method for isotopic determination of ³⁰Si by MS in plants and soils was proposed with detection limit of 50 mg kg⁻¹ for the soil samples.¹⁹³

MS is applied to have a better control of silicon compounds present in industrial applications such as semiconductors and packaging industries¹⁹⁶ and also in some processes.¹⁹⁷ For instance, in situ MS was carried out to characterize silicon compounds (tetraethoxysilane and hexamethyldisiloxane) enabling the production of protective layers for semiconductors.¹⁹⁶ In addition, Apicella et al. observed siloxane series after an extraction by dichloromethane during the analysis of soot recovered in fuel-rich flames using several burners and different fuels by MALDI/MS.¹⁹⁷ In addition, Apicella et al. observed siloxane series after an extraction by dichloromethane during the analysis of soot formed in fuel-rich flames by MALDI/MS.¹⁹⁷

For HMW silicon compounds (Fig. 6) that do not elute in GC, mass spectrometry methods without separation such as MALDI-TOF/MS (time of flight mass spectrometry) or ESI-FT-ICR/MS (Fourier transform-ion cyclotron resonance mass spectrometry) were applied to silicon molecules.¹⁷ In polymer synthesis, the degree of intramolecular condensation, defined as the number of residual silanol (Si–OH) groups per oligomer, for a variety of silsesquioxane polymer (thermal stability and chemical resistance), was measured by MALDI-TOF/MS.^198,199 Indeed, condensation of the Si–OH groups leads to the formation of intramolecular group Si–O–Si bridges accompanied by the loss of water.¹⁹⁹ This phenomenon was easily characterized by high resolution MS.¹⁹⁸ The use of high resolution MS, like FT-ICR/MS,²⁰⁰ offers a sub-ppm mass measurement accuracy and allows successful identification of polymers which generally provide nearly identical mass spectra at low m/z.²⁰¹ Murthy et al. studied the kinetics and pathways of the ion-molecule reactions for a mixture of silanes and chlorosilanes (SiH_nCl_4-n) by FT-ICR/MS using a 1 Tesla magnet and EI ionization.²⁰² By coupling electrospray ionization (ESI) to FT-ICR/MS, the first work on PDMS pointed out several fragmentation patterns by hydrogen bond rupture or by methyl transfer.²⁰¹ Another study by ESI-FT-ICR/MS in positive ion mode demonstrated the benefit of high resolution in order to characterize resulting molecules, but also to determine fragmentation pathways.²⁰³ Note that polymer containing labile hydrogen in terminal groups can easily undergo a fragmentation in gas phase and generate species with silanol groups.²⁰³ Tecklenburg et al. compared ESI-FT-ICR/MS and MALDI/MS for the characterization of silsesquioxane polymers.²⁰⁴ Samples are prepared in a mixture chloroform–methanol with addition of ammonium acetate to improve silicone ionization. Mass accuracy down to 5 ppm was reported for each fraction and Si isotopes (²⁸Si, ²⁹Si and ³⁰Si) were identified in one structure.²⁰⁴ In conclusion, MS methods with or without chromatographic separation allow identification and quantitative analysis of Si compounds.

5 Conclusions

Due to their wide use in many applications, silicones are spread in several matrices, mainly in environmental and biological, usually at trace concentrations where they have negative impacts. Many analytical methods improve the identification of silicon compounds in environmental, biological and industrial fields. Two different analytical strategies were investigated for silicon analysis: direct methods for total silicon determination and hyphenated techniques for silicon speciation.

However, the unravelling of silicon speciation is relatively scarce, particularly in petroleum and derived samples where complex reactions can occur. Indeed, the matrix complexity associated to contamination, instability and trace level presence makes silicon analysis very hard to achieve. Besides, contamination must be minimized using analyzer specific parts and with care concerning sample conservation and storage.

Atomic spectroscopic methods were usually employed for total Si determination. AAS is one of the first technique used for Si determination with detection limits near equal to 1 mg L⁻¹ for environmental and biological samples. The application of ICP-OES and ICP-MS in environmental, biological and industrial matrices progressed over the decades due to their sensitivity (sub μg L⁻¹ and sub ng L⁻¹ respectively), selectivity and robustness performance. Nevertheless, these analytical methods do not allow molecular separation and identification (speciation) unless they are coupled to a separation technique. On the opposite, NMR methods have the potential to achieve silicon determination but also silicon speciation at high levels (50 mg kg⁻¹) in various matrices.

Hyphenated techniques based on the coupling of a chromatographic separation (GC, LC) giving access to retention time, and a sensitive detection (SIM, AED, ICP) are an established versatile analytical tool for Si speciation. GC-MS in SIM mode has proved to be very effective for silicon speciation with detection levels of nearly 1 μg kg⁻¹ although possible contamination can occur by column and septum bleeding. MS detection (MS/MS, FT-ICR/MS) without previous separation avoids contamination by GC parts and offers bright perspectives both for LMW and HMW molecules respectively in environmental and polymer field respectively. The coupling between SEC and appropriate detection also allows a better understanding of HMW silicon molecules. RP-HPLC-ICP appears as a good alternative for sensitivity with detection limits ranging between 0.1 and 500 μg L⁻¹ for LMW molecules in environmental and biological samples but chromatographic resolution is lower than GC. Thus, GC separation combined to atomic detectors such as AED, ICP-OES and ICP-MS seems to be the more convenient solution with LOD ranging between 0.1 and 10 μg L⁻¹ for Si speciation.

Considering that detection limits greatly depend on sample extraction and concentration, sensitivity comparison is a hard task. However, the more sensitive detection (sub μg L⁻¹) by ICP-MS using as chromatographic detector has been demonstrated. Consequently, the high resolving capacity of GC and the high sensitivity capability of ICP-MS have made this combination the most efficient and attractive for speciation analysis of silicon in the main fields of interest.

This state of art shows different techniques for silicon speciation. Despite the performances of the classical analytical tools, there remains limitations concerning sensitivity and selectivity in complex matrices. For that reason, the coupling of GC-ICP-MS allowing mass interferences resolution offers promising perspectives and deserves further developments to unravel silicon structures.

6 Glossary of abbreviations

AAS	Atomic Absorption Spectrometry
ACN	Acetonitrile
AED	Atomic Emission Detector
APCI	Atmospheric Pressure Chemical Ionization
B.P	Boiling Point
BSTFA	Bis(trimethylSilyl)-TriFluoroAcetamide
BSA	Bis(trimethylSilyl)Acetamide
C18	Octadecylsilica
CDCl3	Chloroform
CI	Chemical Ionization
CRI	Collision Reaction Interface
c VMS	cyclic Volatile Methyl Siloxanes
d-DIHEN	demountable Direct Injection High Efficiency Nebulizer
DRC	Dynamic Reaction Cell
EI	Electron Ionization
ESI	Electro Spray Ionization
ETV	Electrothermal Vaporization
FID	Flame Ionization Detector
FS	Full Scan
FT-ICR/MS	Fourier Transform-Ion Cyclotron Resonance Mass Spectrometry
FTIR	Fourier Transform InfraRed
GC	Gas Chromatography
GFAAS	Graphite Furnace Atomic Absorption Spectrometry
GPC	Gel Permeation Chromatography
HB-PDMS	α,ω-bis(4-HydroxyButyl)Poly(DiMethylSiloxane)
h-TISIS	heated Torch Integrated Sample Introduction System
HPLC	High Pressure Liquid Chromatography
HPV	High Production Volume
HR-ICP-MS	High Resolution Inductively Coupled Plasma Mass Spectrometry
HS	Head Space
HMW	High Molecular Weight
ICP	Inductively Coupled Plasma
ICP-MS	Inductively Coupled Plasma Mass Spectrometry
ICP-QMS	Inductively Coupled Plasma Quadruple Mass Spectrometry
ICP-OES	Inductively Coupled Plasma Optical Emission Spectroscopy
IR	Infrared Spectroscopy
IUPAC	International Union of Pure and Applied Chemistry
LA-ICP-MS	Laser Ablation Inductively Coupled Plasma Mass Spectrometry
LC	Liquid Chromatography
LMW	Low Molecular Weight
LOD	Limit Of Detection
LT	Low Temperature
M4Q	Tetrakis(trimethylsiloxy)silane
MALDI	Matrix Assisted Laser Desorption Ionization
MC-ICP-MS	Multi Collector Inductively Coupled Plasma Mass Spectrometry
MIBK	Methyl Isobutyl Ketone
M M	Molecular Mass
MPT-OES	Microwave Plasma Torch Atomic Optical Spectrometry
MS	Mass Spectrometry
MS/MS	Tandem Mass Spectrometry
MTBE	Methyl Tertiary Butyl Ether
MSTFA	N-methyl-N-trimethylsilyltrifluoracetamide
NMR	Nuclear Magnetic Resonance
NP-HPLC	Normal Phase-High Pressure Liquid Chromatography
OECD	Organisation for Economic Co-operation and Development
PDMS	Polydimethylsiloxanes
ppm	part per million
PTFE	Polytetrafluoroethylene
RI	Refractive Index detector
RP-HPLC	Reverse Phase-High Pressure Liquid Chromatography
RSD	Relative Standard Deviation
SEC	Size Exclusion Chromatography
SFC	Supercritical Fluid Chromatography
SIM	Single Ion Monitoring
TGA	Thermogravimetric analysis
THF	Tetrahydrofuran
TOF/MS	Time of Flight Mass Spectrometry
USEPA	US Environmental Protection Agency
USN	Ultrasonic Nebulization
UV	Ultra-Violet
VMS	Volatile Methyl Siloxanes
VOC	Volatile Organic Compounds
XRF	X-ray Fluorescence

7 Acknowledgments

The authors would like to thank J. Castrogeorgi and B. Omais for their useful comments. Special acknowledgments go to the reviewers for their constructive reports, which considerably improved the manuscript.

8 References

1. G. Chandra., in Organosilicon Materials; The Handbook of Environmental Chemistry, ed. Chandra G, Springer, Berlin, 1997, 1–324p.
2. R. Dewil, L. Appels and J. Baeyens, Energy Convers. Manage., 2006, 47(13–14), 1711–1722.
3. N. J. Fendinger, R. G. Lehmann, and E. M. Mihaich, in Organosilicon Materials; The Handbook of Environmental Chemistry, ed. Chandra G, Springer, New York, 1997, 1, pp.1–25.
4. A. L. Smith, in The Analytical chemistry of Silicones, ed. Smith A. L., John Wiley & Sons, New York, 1991, 1, pp.1–16.
5. USEPA (United States Environmental Protection Agency), High Production Volume (HPV) Challenge Program, Sponsored Chemicals, 2007, 18-5-2010.
6. OECD (Organisation for Economic Co-operation and Development), The 2007 OECD list of high production volume chemicals, oecd secretariat, 2007, 18-1-2010.
7. E. F. C. Griessbach and R. G. Lehmann, Chemosphere, 1999, 38(6), 1461–1468.
8. L. Kaj, M. Schlabach, J. Andersson, A. P. Cousins, N.Schmidbauer, and E. Brorström-Lundén, Siloxanes in the Nordic Environment, Report 2005 : 593, Copenhagen, 2005.
9. A. L. Quinn, J. M. Regan, J. M. Tobin, B. J. Marinik, J. M. McMahon, D. A. Mcnett, C. M. Sushynski, S. D. Crofoot, P. A. Jean and K. P. Plotzke, J. Toxicol. Sci., 2007, 96(1), 145–153.
10. A. L. Quinn, A. Dalu, L. S. Meeker, P. A. Jean, R. G. Meeks, J. W. Crissman, R. H. Gallavan and K. P. Plotzke, Reprod. Toxicol., 2007, 23(4), 532–540.
11. M. J. Utell, R. Gelein, C. P. Yu, C. Kenaga, E. Geigel, A. Torres, D. Chalupa, F. R. Gibb, D. M. Speers, R. W. Mast and P. E. Morrow, J. Toxicol. Sci., 1998, 44(2), 206–213.
12. Y. Horii and K. Kannan, Arch. Environ. Contam. Toxicol., 2008, 55(4), 701–710.
13. R. Wang, R. P. Moody, D. Koniecki and J. P. Zhu, Environ. Int., 2009, 35(6), 900–904.
14. D. M. Templeton, F. Ariese, R. Cornelis, L. G. Danielsson, H. Muntau, H. P. Van Leeuwen and R. Lobinski, Pure Appl. Chem., 2000, 72(8), 1453–1470.
15. E. H. Evans, Anal. Bioanal. Chem., 2003, 376(3), 311–312.
16. A. L. Smith, in The Analytical Chemistry of Silicones, ed. Smith A. L., John Wiley & Sons, New York, 1991, 1–551p.
17. S. Varaprath, D. Stutts and G. Kozerski, Silicon Chem., 2006, 79, 79–102.
18. A. L. Smith and R. D. Parker, in The Analytical chemistry of Silicones, ed. Smith A. L., John Wiley & Sons, New York, 1991, 4, pp.71–91.
19. J. C. Carpenter and R. Gehards, in Organosilicon Materials; The Handbook of Environmental Chemistry, ed. Chandra G, Springer, New York, 1997, 2, pp.29–50.
20. S. J. Lugowski, D. C. Smith, H. Bonek, J. Lugowski, W. Peters and J. Semple, J. Trace Elem. Med. Biol., 2000, 14(1), 31–42.
21. C. L. Frye, Environ. Toxicol. Chem., 1987, 6(5), 329–330.
22. S. Varaprath and R. G. Lehmann, Journal of Environmental Polymer Degradation, 1997, 5(1), 17–31.
23. R. B. Allen and P. Kochs. G. Chandra, in Organosilicon Materials; The Handbook of Environmental Chemistry, ed. Chandra G, Springer, New York, 1997, 1, pp.1–25.
24. C. Hurd, J. Am. Chem. Soc., 1946, 68, 364.
25. J. F. Hobson, R. Atkinson, and W. P. L. Carter, in Organosilicon Materials; The Handbook of Environmental Chemistry, ed. Chandra G, Springer, New York, 1997, 6, pp.138–177.
26. J. T. James, T. F. Limero, J. H. Leano, J. F. Boyd and B. A. Covington, Aviation, Space, and Environmental Medicine, 1994, 851–857.
27. H. C. Shields, D. M. Fleischer and C. J. Weschler, Indoor Air-International Journal of Indoor Air Quality and Climate, 1996, 6(1), 2–17.
28. S. Xu, Environ. Sci. Technol., 1999, 33(4), 603–608.
29. R. Dewil, L. Appels, J. Baeyens, A. Buczynska and L. Vaeck, Talanta, 2007, 74(1), 14–19.
30. R. Huppmann, H. W. Lohoff and H. F. Schroder, Fresenius J. Anal. Chem., 1996, 354(1), 66–71.
31. R. Grumping and A. V. Hirner, Fresenius J. Anal. Chem., 1999, 363(4), 347–352.
32. M. Schweigkofler and R. Niessner, Environ. Sci. Technol., 1999, 33(20), 3680–3685.
33. R. G. Lehmann, S. Varaprath and C. L. Frye, Environ. Toxicol. Chem., 1994, 13(11), 1753–1759.
34. R. G. Lehmann, S. Varaprath and C. L. Frye, Environ. Toxicol. Chem., 1994, 13(7), 1061–1064.
35. R. G. Lehmann, S. Varaprath, R. B. Annelin and J. L. Arndt, Environ. Toxicol. Chem., 1995, 14(8), 1299–1305.
36. R. E. Pellenbarg, Environ. Sci. Technol., 1979, 13(5), 565–569.
37. R. G. Lehmann, J. R. Miller, S. Xu, U. B. Singh and C. F. Reece, Environ. Sci. Technol., 1998, 32(9), 1260–1264.
38. R. Atkinson, Environ. Sci. Technol., 1991, 25(5), 863–866.
39. R. Sommerlade, H. Parlar, D. Wrobel and P. Kochs, Environ. Sci. Technol., 1993, 27(12), 2435–2440.
40. M. J. Whelan, E. Estrada and R. van Egmond, Chemosphere, 2004, 57(10), 1427–1437.
41. R. R. Buch and D. N. Ingebrigtson, Environ. Sci. Technol., 1979, 13(6), 676–679.
42. J. C. Carpenter, J. A. Cella and S. B. Dorn, Environ. Sci. Technol., 1995, 29(4), 864–868.
43. M. Arnold and T. Kajolinna, Waste Manage., 2010, 30(6), 1011–1017.
44. L. Appels, J. Baeyens, J. Degreve and R. Dewil, Prog. Energy Combust. Sci., 2008, 34(6), 755–781.
45. M. Rahmani and M. Sohrabi, Kinet. Catal., 2006, 47(6), 891–900.
46. Z. Long, J. Storey, S. Lewis and M. J. Sepaniak, Anal. Chem., 2009, 81(7), 2575–2580.
47. T. Matsui and S. Imamura, Bioresour. Technol., 2010, 101, S29–S32.
48. M. Schweigkofler and R. Niessner, J. Hazard. Mater., 2001, 83(3), 183–196.
49. J. J. Barnard, E. L. Todd, W. G. Wilson, R. Mielcarek and R. J. Rohrich, Plast. Reconstr. Surg., 1997, 100(1), 197–203.
50. D. Flassbeck, B. Pfleiderer, R. Grumping and A. V. Hirner, Anal. Chem., 2001, 73(3), 606–611.
51. E. D. Lykissa, S. V. Kala, J. B. Hurley and R. M. Lebovitz, Anal. Chem., 1997, 69(23), 4912–4916.
52. D. Flassbeck, B. Pfleiderer, P. Klemens, K. G. Heumann, E. Eltze and A. V. Hirner, Anal. Bioanal. Chem., 2003, 375(5), 356–362.
53. S. V. Kala, E. D. Lykissa and R. M. Lebovitz, Anal. Chem., 1997, 69(7), 1267–1272.
54. L. Lipworth, R. E. Tarone and J. K. McLaughlin, Plast. Reconstr. Surg., 2009, 123(3), 790–793.
55. L. A. Brinton, Plast. Reconstr. Surg., 2007, 120(7), 94S–102S.
56. J. Maittala, S. Pennanen and J. Liesivuori, Analyst, 1999, 124(5), 665–668.
57. Dow Corning Corporation, Decamethylcyclopentasiloxane:A 24-month combinedchronic toxicity and oncogenicity whole body vapor inhalation study in Fischer-344 rats, Report 2004-10000-54953, 2005.
58. T. H. Thomas and T. C. Kendrick, J. Polym. Sci., Part A-2, 1969, 7, 537.
59. N. Grassie and I. G. Macfarlane, Eur. Polym. J., 1978, 14(11), 875–884.
60. G. Camino, S. M. Lomakin and M. Lageard, Polymer, 2002, 43(7), 2011–2015.
61. G. Camino, S. M. Lomakin and M. Lazzari, Polymer, 2001, 42(6), 2395–2402.
62. V. V. Rode, M. A. Verkhotin and S. R. Rafikov, Vysokomol soyed, 1968, 7, 1529–1538.
63. G. Deshpande and M. E. Rezac, Polym. Degrad. Stab., 2001, 74(2), 363–370.
64. N. Grassie and K. F. Francey, Polym. Degrad. Stab., 1980, 2(1), 53–66.
65. P. Dufresne, Appl. Catal., A, 2007, 322, 67–75.
66. K. R. Carduner, R. O. Carter and L. C. Westwood, Appl. Spectrosc., 1988, 42(7), 1265–1267.
67. L. N. Kremer and T. G. Hueston, Petrol Tech Q, 2002, 65–69.
68. L. Kellberg, P. Zeuthen and H. J. Jakobsen, J. Catal., 1993, 143(1), 45–51.
69. C. P. Lienemann, Oil Gas Sci. Technol., 2005, 60(6), 951–965.
70. C. P. Lienemann, S. Dreyfus, C. Pecheyran and O. F. X. Donard, Oil Gas Sci. Technol., 2007, 62(1), 69–77.
71. B. Didillon, J. Cosyns, C. Cameron, D. Uzio, P. Sarrazin and J. P. Boitiaux, Catalyst Desactivation, 1997, 111, 447–454.
72. J. W. Da, J Porus Matter, 2008, 15(199), 204.
73. M. O. G. Souza, P. Reyes and M. C. Rangel, Stud. Surf. Sci. Catal., 1999, 126, 469–472.
74. M. D. Phillips and E. L. Sughrue, Fuel Sci. Technol. Int., 1991, 9(3), 305–319.
75. A. Molnar, I. Bucsi, M. Bartok, F. Notheisz and G. V. Smith, J. Catal., 1986, 98(2), 386–391.
76. R. Breivik and R. Egebjerg, Petrol Tech Q, 2008, Q1, 69–75.
77. K. Arnby, M. Rahmani, M. Sanati, N. Cruise, A. A. Carlsson and M. Skoglundh, Appl. Catal., B, 2004, 54(1), 1–7.
78. A. C. Larsson, M. Rahmani, K. Arnby, M. Sohrabi, M. Skoglundh, N. Cruise and M. Sanati, Top. Catal., 2007, 45(1–4), 121–124.
79. M. Rahmani and M. Sohrabi, React. Kinet. Catal. Lett., 2005, 86(2), 397–405.
80. F. McElroy, A. Mennito, E. Debrah and R. Thomas, Spectroscopy, 1998, 13(2), 42.
81. J. A. D. Amaro and S. L. C. Ferreira, J. Anal. At. Spectrom., 2004, 19(2), 246–249.
82. R. I. Botto, J. Anal. At. Spectrom., 1993, 8(1), 51–57.
83. H. M. Dong and V. Krivan, J. Anal. At. Spectrom., 2003, 18(4), 367–371.
84. S. Hauptkorn, J. Pavel and H. Seltner, Fresenius J. Anal. Chem., 2001, 370(2–3), 246–250.
85. J. P. McConnell, T. P. Moyer, D. E. Nixon, P. L. Schnur, D. R. Salomao, T. B. Crotty, J. Weinzweig, J. B. Harris and P. M. Petty, American Journal of Clinical Pathology, 1997, 107(2), 236–246.
86. H. T. Liu and S. J. Jiang, Spectrochim. Acta, Part B, 2003, 58(1), 153–157.
87. C. H. Yang and S. J. Jiang, Spectrochim. Acta, Part B, 2004, 59(9), 1389–1394.
88. A. Mukhtar and A. Limbeck, Anal. Chim. Acta, 2009, 646(1–2), 17–22.
89. C. M. Malata, S. Varma, M. Scott, J. C. Liston and D. T. Sharpe, Medical Progress Through Technology, 1994, 20(3–4), 251–260.
90. W. Peters, D. Smith and S. Lugowski, Ann. Plast. Surg., 1995, 35(4), 442–443.
91. S. Lugowski, D. Smith and J. Bzdega, Chemia Analityczna, 1998, 43(6), 1011–1019.
92. S. Felby, Forensic Sci. Int., 1986, 32(1), 61–65.
93. M. D. Huang and V. Krivan, Spectrochim. Acta, Part B, 2007, 62(3), 297–303.
94. F. Y. Leung and P. Edmond, Clin. Biochem., 1997, 30(5), 399–403.
95. M. Hornung and V. Krivan, J. Anal. At. Spectrom., 1997, 12(10), 1123–1130.
96. H. J. Gitelman and F. B. Alderman, J. Anal. At. Spectrom., 1990, 5(8), 687–689.
97. Z. E. Huang, Spectrochim. Acta, Part B, 1995, 50(11), 1383–1393.
98. W. Frech and A. Cedergren, Anal. Chim. Acta, 1980, 113(2), 227–235.
99. C. J. Rademeyer and I. Vermaak, J. Anal. At. Spectrom., 1992, 7(2), 347–351.
100. G. N. Brown, D. L. Styris and M. W. Hinds, J. Anal. At. Spectrom., 1995, 10(8), 527–531.
101. A. F. Shoukry, Y. M. Issa, R. A. Farghaly, M. Grasserbauer, H. Puxbaum and J. Rendl, Fresenius J. Anal. Chem., 1998, 360(6), 650–653.
102. P. Bermejo-Barrera, M. C. Barciela-Alonso, R. Dominguez-Gonzalez, A. Bermejo-Barrera, J. A. C. de Juan and J. M. Fraga-Bermudez, Anal. Bioanal. Chem., 2002, 374(7–8), 1290–1293.
103. R. D. Parker, Journal of The Association of Official Analytical Chemists, 1990, 73(5), 721–723.
104. D. A. Mccamey, D. P. Iannelli, L. J. Bryson and T. M. Thorpe, Anal. Chim. Acta, 1986, 188, 119–126.
105. P. Masson, M. Dauthieu, F. Trolard and L. Denaix, Spectrochim. Acta, Part B, 2007, 62(3), 224–230.
106. Y. C. Qin and H. X. Weng, Estuarine, Coastal Shelf Sci., 2006, 67(3), 433–440.
107. A. L. Molinero, L. Martinez, A. Villareal and J. R. Castillo, Talanta, 1998, 45(6), 1211–1217.
108. J. M. T. Carneiro, A. L. R. M. Rossete, G. S. Oliveira and J. A. Bendassolli, Commun. Soil Sci. Plant Anal., 2007, 38(11–12), 1411–1423.
109. N. Watanabe, T. Nakamura, E. Watanabe, E. Sato and Y. Ose, Sci. Total Environ., 1984, 35(1), 91–97.
110. N. Watanabe, H. Nagase and Y. Ose, Sci. Total Environ., 1988, 73(1–2), 1–9.
111. J. F. Belliveau, W. C. Griffiths, C. G. Wright and J. R. Tucci, Annals of Clinical and Laboratory Science, 1991, 21(5), 328–334.
112. G. M. Bercowy, H. Vo and F. Rieders, Journal of Analytical Toxicology, 1994, 18(1), 46–48.
113. X. J. Jia, T. B. Wang, X. D. Bu and J. Wu, Microchem. J., 2003, 75(2), 103–107.
114. A. P. Krushevska and R. M. Barnes, J. Anal. At. Spectrom., 1994, 9(9), 981–984.
115. J. J. Powell, S. A. McNaughton, R. Jugdaohsingh, S. H. C. Anderson, J. Dear, F. Khot, L. Mowatt, K. L. Gleason, M. Sykes, R. P. H. Thompson, C. Bolton-Smith and M. J. Hodson, Br. J. Nutr., 2005, 94(5), 804–812.
116. R. Sanchez, J. L. Todoli, C. P. Lienemann and J. M. Mermet, J. Anal. At. Spectrom., 2009, 24(4), 391–401.
117. R. Sanchez, J. L. Todoli, C. P. Lienemann and J. M. Mermet, J. Anal. At. Spectrom., 2009, 24(10), 1382–1388.
118. R. Sanchez, J. L. Todoli, C. P. Lienemann and J. M. Mermet, J. Anal. At. Spectrom., 2010, 25(2), 178–185.
119. J. Zhang, L. Li, J. Zhang, Q. Zhang, Y. Yang and Q. Jin, Pet. Sci. Technol., 2007, 25(3–4), 443–451.
120. R. I. Botto and J. Talbott, 2009 Winter Conference on Plasma Spectrochemistry, Graz, Austria, 2009.
121. Y. Takaku, K. Masuda, T. Takahashi and T. Shimamura, J. Anal. At. Spectrom., 1994, 9(12), 1385–1387.
122. E. Engstrom, I. Rodushkin, D. C. Baxter and B. Ohlander, Anal. Chem., 2006, 78(1), 250–257.
123. L. Y. Alleman, D. Cardinal, C. Cocquyt, P. D. Plisnier, J. P. Descy, I. Kimirei, D. Sinyinza and L. Andre, J. Great Lakes Res., 2005, 31(4), 509–519.
124. C. F. Wang, F. H. Tu, S. L. Jeng and C. J. Chin, J. Radioanal. Nucl. Chem., 1999, 242(1), 97–103.
125. X. B. Feng, S. L. Wu, A. Wharmby and A. Wittmeier, J. Anal. At. Spectrom., 1999, 14(6), 939–946.
126. P. Klemens and K. G. Heumann, Fresenius J. Anal. Chem., 2001, 371(6), 758–763.
127. K. Van Dyck, H. Robberecht, R. Van Cauwenbergh, H. Deelstra, J. Arnaud, L. Willemyns, F. Benijts, J. A. Centeno, H. Taylor, M. E. Soares, M. L. Bastos, M. A. Ferreira, P. C. D'Haese, L. V. Lamberts, M. Hoenig, G. Knapp, S. J. Lugowski, L. Moens, J. Riondato, R. Van Grieken, M. Claes, R. Verheyen, L. Clement and M. Uytterhoeven, J. Anal. At. Spectrom., 2000, 15(6), 735–741.
128. J. Riondato, F. Vanhaecke, L. Moens and R. Dams, J. Anal. At. Spectrom., 1997, 12(9), 933–937.
129. J. Talbott and R. I. Botto, Silicon in Naphtas: A near cold plasma approach for sub-ppb detection limits on an ICP/MS, 2008 Winter Conference on Plasma Spectrochemistry, Temecula, California, 2008.
130. H. M. Kuss, D. Bossmann and M. Muller, At. Spectrosc., 1994, 15(4), 148–150.
131. M. Resano, M. Verstraete, F. Vanhaecke and L. Moens, J. Anal. At. Spectrom., 2002, 17(8), 897–903.
132. I. De Schrijver, M. Aramendia, M. Resano, A. Dumoulin and F. Vanhaecke, J. Anal. At. Spectrom., 2008, 23(4), 500–507.
133. R. Hamilton, University of Lakehead, 2001.
134. R. B. Taylor, B. Parbhoo, and D. M. Fillmore, in The Analytical chemistry of Silicones, ed. Smith A. L., John Wilery & Sons, New York, 1991, 12, pp.347–414.
135. P. Fux, Analyst, 1990, 115(2), 179–183.
136. L. Garrido, B. Pfleiderer, B. G. Jenkins, C. A. Hulka and D. B. Kopans, Magn. Reson. Med., 1994, 31(3), 328–330.
137. L. Garrido, A. Bogdanova, L. L. Cheng, B. Pfleiderer, E. Tokareva, J. L. Ackerman and T. J. Brady, Immunology of Silicones, 1996, 210, 49–58.
138. B. Pfleiderer, A. Moore, E. Tokareva, J. L. Ackerman and L. Garrido, Biomaterials, 1999, 20(6), 561–571.
139. W. E. Hull, Magn. Reson. Med., 1999, 42(5), 984–995.
140. P. Macdonald, N. Plavac, W. Peters, S. Lugowski and D. Smith, Anal. Chem., 1995, 67(20), 3799–3801.
141. J. M. Bellama, S. R. Meyer and R. E. Pellenbarg, Appl. Organomet. Chem., 1991, 5(2), 107–109.
142. D. R. Peterson, R. D. Parker, and M. J. Owen, in The Analytical chemistry of Silicones, ed. Smith A. L., John & Wiley Sons, New York, 1991, pp.485–515.
143. A. D. Lipp and A. L. Smith, in The Analytical chemistry of Silicones, ed. Interscience Publication, New York, 1991, pp.305–337.
144. M. Hernandez, P. Diaz and M. Carrillo, UNITAR 5th International Conference, 1991, 57–63.
145. E. D. Lipp, P. S. Rzyrkowski and R. F. Geiger, Tappi J., 1987, 95–98.
146. J. A. Emery, G. Kasnic, N. Hardt and S. S. Spanier, Modern pathology, 1994, 7(7), 728–733.
147. N. S. Hardt, L. T. Yu, G. Latorre and B. Steinbach, Modern pathology, 1994, 7(6), 669–676.
148. L. H. Kidder, V. F. Kalasinsky, J. L. Luke, I. W. Levin and E. N. Lewis, Nat. Med., 1997, 3(2), 235–237.
149. M. Ende and G. Spiteller, Mass Spectrom. Rev., 1982, 1, 29–62.
150. J. De Zeeuw, Am. Lab., 2005, 37, 18–19.
151. S. Varaprath and P. S. Larson, J. Polym. Environ., 2002, 10(4), 119–131.
152. C. Sparham, R. van Egmond, S. O'Connor, C. Hastie, M. Whelan, R. Kanda and O. Franklin, J. Chromatogr., A, 2008, 1212, 124–129.
153. S. Varaprath, M. Seaton, D. McNett, L. Cao and K. P. Plotzke, Int. J. Environ. Anal. Chem., 2000, 77(3), 203–219.
154. J. Carter, L. Ebdon and E. H. Evans, Microchem. J., 2004, 76(1–2), 35–41.
155. D. M. Chambers, D. O. McElprang, J. P. Mauldin, T. M. Hughes and B. C. Blount, Anal. Chem., 2005, 77(9), 2912–2919.
156. S. J. Pattinson and J. P. G. Wilkins, Analyst, 1989, 114(4), 429–434.
157. Y. X. Wang, Am. Lab., 2006, 38(2), 10–12.
158. S. B. Dorn and E. M. S. Frame, Analyst, 1994, 119(8), 1687–1694.
159. C. Eaborn, in Organosilicon Compounds, ed. Butterworth Scientific, London, 1960, 8, pp.227–288.
160. S. Varaprath, K. L. Salyers, K. P. Plotzke and S. Nanavati, Drug Metab. Dispos., 1999, 27(11), 1267–1273.
161. S. Varaprath, J. M. McMahon and K. P. Plotzke, Drug Metab. Dispos., 2003, 31(2), 206–214.
162. V. Blechta, M. Kurfurst, J. Sykora and J. Schraml, J. Chromatogr., A, 2007, 1145(1–2), 175–182.
163. W. R. Biggs, J. C. Fetzer and R. J. Brown, Anal. Chem., 1987, 59(23), 2798–2802.
164. W. R. Biggs and J. C. Fetzer, Anal. Chem., 1989, 61(3), 236–240.
165. D. R. Heine, M. B. Denton and T. D. Schlabach, J. Chrom. Sci., 1985, 23(10), 454–458.
166. L. Ebdon, M. Foulkes, K. Fredeen, C. Hanna and K. Sutton, Spectrochim. Acta, Part B, 1998, 53(6–8), 859–865.
167. R. G. Lehmann and J. R. Miller, Environ. Toxicol. Chem., 1996, 15(9), 1455–1460.
168. J. C. Ivaldi and J. F. Tyson, Spectrochim. Acta, Part B, 1995, 50(10), 1207–1226.
169. R. D. Steinmeyer and M. A. Becker, in The Analytical chemistry of Silicones, ed. Smith A. L., John Wiley & Sons, New York, 1991, 10, pp.255–303.
170. D. M. Smith, R. G. Lehmann, R. Narayan, G. E. Kozerski and J. R. Miller, Compost Science & Utilization, 1998, 6(2), 6–12.
171. D. W. Hausler and L. T. Taylor, Anal. Chem., 1981, 53(8), 1223–1227.
172. D. W. Hausler and L. T. Taylor, Anal. Chem., 1981, 53(8), 1227–1231.
173. E. P. Maziarz, X. M. Liu, E. T. Quinn, Y. C. Lai, D. M. Ammon and G. L. Grobe, J. Am. Soc. Mass Spectrom., 2002, 13(2), 170–176.
174. X. M. Liu, E. P. Maziarz, D. J. Heiler and G. L. Grobe, J. Am. Soc. Mass Spectrom., 2003, 14(3), 195–202.
175. R. Grumping, D. Mikolajczak and A. V. Hirner, Fresenius J. Anal. Chem., 1998, 361(2), 133–139.
176. S. C. Popat and M. A. Deshusses, Environ. Sci. Technol., 2008, 42(22), 8510–8515.
177. S. Wacholz, F. Keidel, U. Just, H. Geissler and K. Kappler, J. Chromatogr., A, 1995, 693, 89–99.
178. E. D. Pellizzari, J. E. Bunch, R. E. Berkley and J. Mc Rae, Anal. Chem., 1976, 48(6), 1–4.
179. X. M. Wang, S. C. Lee, G. Y. Sheng, L. Y. Chan, J. M. Fu, X. D. Li, Y. S. Min and C. Y. Chan, Appl. Geochem., 2001, 16(11–12), 1447–1454.
180. J. J. Kennan, L. L. M. Breen, T. H. Lane and R. B. Taylor, Anal. Chem., 1999, 71(15), 3054–3060.
181. H. Homma, T. Kuroyagi, K. Izumi, C. L. Mirley, J. Ronzello and S. A. Boggs, IEEE Trans. Nucl. Sci., 2000, 15(2), 796–803.
182. K. Kappler, U. Scheim, F. Keidel and U. Just, Fresenius J. Anal. Chem., 1996, 354(1), 21–26.
183. H. Hillborg, S. Karlsson and U. W. Gedde, Polymer, 2001, 42(21), 8883–8889.
184. A. D. Hall and M. Patel, Polym. Degrad. Stab., 2006, 91(10), 2532–2539.
185. M. Edler, D. Metze, N. Jakubowski and M. Linscheid, J. Anal. At. Spectrom., 2002, 17(10), 1209–1212.
186. J. A. Caruso and M. Montes-Bayon, Ecotoxicol. Environ. Saf., 2003, 56(1), 148–163.
187. C. A. P. de Leon, M. Montes-Bayon and J. A. Caruso, J. Chromatogr., A, 2002, 974(1–2), 1–21.
188. J. C. A. Wuilloud, R. G. Wuilloud, A. P. Vonderheide and J. A. Caruso, Spectrochim. Acta, Part B, 2004, 59(6), 755–792.
189. B. Bouyssiere, J. Szpunar and R. Lobinski, Spectrochim. Acta, Part B, 2002, 57(5), 805–828.
190. K. A. Forbes, J. F. Vecchiarelli, P. C. Uden and R. M. Barnes, Anal. Chem., 1990, 62(18), 2033–2037.
191. D. W. Later, E. R. Campbell and B. E. Richter, J. High Resolut. Chromatogr., 1988, 11(1), 65–69.
192. U. Just, F. Mellor and F. Keidel, J. Chromatogr., A, 1994, 683(1), 105–113.
193. J. M. T. Carneiro, A. L. R. M. Rossete and J. A. Bendassolli, Anal. Lett., 2008, 41(9), 1640–1647.
194. K. Badjagbo, A. Furtos, M. Alaee, S. Moore and S. Sauve, Anal. Chem., 2009, 81(17), 7288–7293.
195. K. Badjagbo, M. Heroux, M. Alaee, S. Moore and S. Sauve, Environ. Sci. Technol., 2010, 44(2), 600–605.
196. V. Rouessac, S. Roualdes and J. Durand, Chem. Vap. Deposition, 2002, 8(4), 155–161.
197. B. Apicella, A. Ciajolo, M. Milian, C. Galmes, A. A. Herod and R. Kandiyoti, Rapid Commun. Mass Spectrom., 2004, 18(3), 331–338.
198. W. E. Wallace, C. M. Guttman and J. M. Antonucci, J. Am. Soc. Mass Spectrom., 1999, 10(3), 224–230.
199. W. E. Wallace, C. M. Guttman and J. M. Antonucci, Polymer, 2000, 41(6), 2219–2226.
200. A. G. Marshall, C. L. Hendrickson and G. S. Jackson, Mass Spectrom. Rev., 1998, 17(1), 1–35.
201. E. P. Maziarz, G. A. Baker and T. D. Wood, Macromolecules, 1999, 32(13), 4411–4418.
202. S. Murthy and J. L. Beauchamp, J. Phys. Chem., 1992, 96(3), 1247–1257.
203. H. P. Chen, J. Am. Soc. Mass Spectrom., 2003, 14(9), 1039–1048.
204. R. E. Tecklenburg, W. E. Wallace and H. P. Chen, Rapid Commun. Mass Spectrom., 2001, 15(22), 2176–2185.

---

Footnote

† This article is part of a themed issue highlighting outstanding and emerging work in the area of speciation.

---