Ed McCurdy is ICP-MS Product Specialist with Agilent Technologies, Manchester, UK.

Ed McCurdy

Since the launch of the first commercial triple-quadrupole ICP-MS (ICP-MS/MS) instrument in 2012, the technique has become well established in both research and academic institutes, as well as in industrial and commercial laboratories. Users in routine applications wish to benefit from the improved interference removal performance of ICP-MS/MS, without facing the more complex method development inherent in the use of ion-molecule reaction chemistry. In this article, we follow method setup and optimization steps to illustrate how an ICP-MS/MS method can be defined and tested to ensure consistent performance.

	Single-quadrupole inductively coupled plasma–mass spectrometry (ICP-MS) provides performance that meets the needs of most inorganic (metals) analysis laboratories, and exceeds the method requirements for many. The technique offers a unique combination of fast multielement analysis, good elemental coverage, very low detection limits, and the ability to measure major and trace elements in a single acquisition. The development of collision-reaction cells (CRCs) operating in helium collision mode addressed many of the polyatomic interferences that previously hampered accurate analysis of several critical trace elements (1). However, helium mode cannot resolve interferences such as isobaric and doubly charged overlaps. Also, some combinations of analyte and interference require more effective attenuation of the overlap than can be achieved using collision mode and kinetic energy discrimination (KED). For these cases, ion-molecule reaction chemistry using a reactive cell gas may offer a better solution.

	Reactive cell gas methods have been investigated since CRCs were first developed for single-quadrupole ICP-MS in the late 1990s (2). Early work presented the benefits of reaction chemistry for historically problematic analytes that suffer intense spectral overlap (such as selenium, arsenic, iron, calcium, and so on). However, these early applications were often based on the theoretical efficiency of the reaction chemistry to resolve one specific interference, and were illustrated using simple, consistent samples or synthetic solutions. Often, when these methods were applied to more complex and variable natural samples, it became apparent that the reaction processes could be adversely affected by other components in the sample (3,4). An unexpected matrix element or co-existing analyte could react with the cell gas, forming unwanted overlaps on the analyte ion (or product ion). Consequently, a reaction gas method developed for one sample matrix or interference might give erroneous results for a different sample type or combination of analytes.

	The development of triple quadrupole ICP-MS (ICP-MS/MS) in 2012 offered a solution to the inaccurate results observed when using reactive cell gases on single quadrupole instruments (5). ICP-MS/MS has an additional quadrupole mass filter (Q1) before the CRC, in a tandem mass spectrometer (MS/MS) configuration. When Q1 is operated as a true mass filter (1 u mass window), it controls the ions passed to the cell, rejecting all nontarget masses. This means the reaction processes and product ions formed are predictable and consistent, and are unaffected by other analytes or matrix elements in the sample (6).

	In addition to improving data quality, the control of reaction chemistry provided by ICP-MS/MS means that method settings can be applied consistently, even for variable samples. This makes reaction mode method setup easier than on a single quadrupole instrument or a configuration with a bandpass filter (mass window >1 u) before the cell. In this article, we illustrate the information and tools available for ICP-MS/MS to support development and performance testing of a new reaction mode method.

	Reference Methods for Reactive Cell Gas Operation with ICP-MS/MS

	Since the first commercial ICP-MS/MS instrument was released in 2012, many applications have been developed, optimized, and published. In addition to manufacturers' application notes (7), many hundreds of peer-reviewed journal articles detail method settings for both typical and unusual applications.

	The ICP-MS/MS instrument's built-in methods can predefine the typical cell gas mode, analyte and product ion masses, and integration times for common analytes. Where an analyte does not have an established setting predefined in a preset method, the user can easily set up one or more mass pairs for the chosen reaction gas. The choice of reaction gas used to resolve a specific interference can be informed by the literature. Theoretical ion-molecule reaction thermodynamics show which reactions are favored (8,9), while ICP-MS/MS manufacturer's publications show experimental results for various product ion formation rates. Hydrogen (H

) are among the most widely-used reaction gases. For each gas, several common reaction transitions are defined in the literature, and have been confirmed experimentally (10).

	With oxygen cell gas, for example, analytes may be measured on mass (Q2 is set to the same mass as Q1) or mass shifted as the oxide ion (Q2 is set to Q1 + 16 u) or dioxide ion (Q2 is set to Q1 + 32 u). These mass shift transitions, together with equivalent transitions for other reaction gases, are available as preset values in the method setup software, allowing easy selection of the preferred mass shift. Multiple mass shifts can be selected if the user wants to compare results for several product ions or cell gases, and custom mass shifts can be defined, if appropriate.

cell gas, reactions are usually fast and are often sequential, so multiple cluster ions can form from both the analyte ions and any other ions present in the cell. Typical reactions and product ions formed with NH

 cell gas are shown in Table I. MS/MS mode controls the precursor ions that can enter the cell and react, so product ions can only form from ions at the mass of the target analyte ion. However, optimized methods must still be verified to confirm that the chosen reaction product ion is free from overlap by product ions from other sample components present at the same mass as the analyte precursor ion.

	Product ion scanning, which is unique to MS/MS, is a useful method development tool for unfamiliar sample types and unusual combinations of analyte interferences. This mode enables users to check and verify the best product ion(s) to use by identifying product ions that are free from overlap in a given sample type or matrix. This is particularly useful for complex and variable sample matrices, and for the large number of product ions produced when a highly reactive cell gas such as NH

	In a product ion scan, Q1 is set to the mass of the target analyte isotope, and Q2 is scanned across the mass range where useful product ions could appear. It is essential that Q1 is operated at 1 u resolution, so only ions at the target analyte isotope mass enter the cell. A product ion scan is acquired while aspirating a single-element standard of the target analyte, allowing product ions formed from the analyte isotope alone to be identified. A second product ion scan is then acquired while aspirating the unknown sample or synthetic matrix to identify product ions formed from other (nontarget) precursor ions at the target analyte mass. By comparing the spectra from the two scans, analyte product ions that are free from overlap by matrix product ions can easily be identified (11).

	An Agilent 8900 ICP-MS/MS instrument (model #100, Advanced Applications configuration) was used for all measurements. The instrument was fitted with the standard glass concentric nebulizer, quartz spray chamber, quartz torch with 2.5 mm injector, and standard nickel sampling and skimmer cones. Samples were introduced via a peristaltic pump with 1.02-mm i.d. pump tubing. Operating conditions are shown in Table II.

	The hafnium standard was prepared from a single element stock solution, and the rare earth matrix was prepared from a mixed 14-element stock standard (Agilent Technologies). Ultrapure water was prepared in the laboratory using a Milli-Q Element system, and nitric and hydrochloric acids used for sample stabilization were Ultrapure NORMATOM grade (VWR Chemicals).

	Hf and mixed rare earth elements (REE) standards were prepared at 10 ppb and 1 ppm, respectively, in a final diluent of 2% HNO

 and 1% HCl. A spike solution of 10 ppb Hf in 1 ppm REE mix was also prepared.

Choosing Reaction Gas Modes to Resolve Known Interferences

	Theoretical ion-molecule reaction rates can be used to identify potential approaches to resolve a given interference, based on differences in the reactions for the analyte and interfering ion. Figure 1 shows the reactions of all elemental ions with ammonia cell gas (10), displayed in a color-coded periodic table layout. To illustrate how this information can be used, the different reactions of Hg

 are highlighted in the table. Mercury (a Type 3 element) reacts with NH

 cell gas via a charge-transfer reaction leading to the neutralization of the Hg

 ion. In contrast, Pb (a Type 1 element) does not react with NH

, so it can be measured on mass. The difference in reactions of Hg and Pb ions indicates that NH

 could be a suitable cell gas to resolve the 


	Figure 1: Reaction pathways and rates for all analyte ions with NH

 cell gas derived from reference 10. Hg and Pb are highlighted, showing how reaction data can be used to identify the potential to resolve the isobaric overlap from 

	In the case of hafnium analysis in a rare earth element matrix, the reaction data presented in Figure 1 suggests that NH

 cell gas might be able to resolve the Yb and Lu overlaps on Hf. The 

Hf isotope is the preferred isotope for quantitative analysis, but 

Hf is required for some applications such as Hf/Lu and Hf/Hf isotope geochronology. The 

Hf isotope could be considered the worst case in terms of interferences in sample matrices that contain REEs. As well as direct isobaric interferences from 

Dy. Figure 2 shows that Hf (a Type 2b element) reacts with NH

 cell gas, whereas Yb (a Type 1 element) does not. Hafnium will therefore form Hf-NH

 product ions that cannot be overlapped by corresponding NH

 may be a suitable cell gas to resolve Yb and Lu overlaps on Hf (refer to Figure 1 for key to color-coding of NH

	The same principle does not necessarily apply to resolving the 

Hf. Lu is a Type 2a element, so it does react with NH

, and may form product ions that can overlap the Hf-NH

product ions. Also, ion-molecule reaction data does not give specific information about the potential for a given cell gas to resolve polyatomic overlaps, such as the 

Hf. The solution to confirming the suitability of a given reaction gas to resolve such interferences is to verify the performance experimentally by using a product ion scan.

 Product Ions in REE Matrix Using Product Ion Scanning 

	For analytes, interferences, and sample types that are not defined in preset methods or available from reference publications, product ion scanning can identify the best product ion(s) for the application. Product ion spectra were acquired for a 10 ppb Hf standard and a 1 ppm 14-element REE mix. An overlay of the two product ion spectra covering the mass range from 

 175 to 265 is shown in Figure 3. The spectrum comparison facilitates identification of product ions formed from 

 176 precursor ions in the Hf standard but not the REE mix. The most intense interference-free 

 258. This Hf product ion is free from any REE-based overlap-that is, no overlapping NH

 product ions are formed from Lu and GdO/DyO precursor ions present at 

 product ions formed from m/z 176 in Hf standard (gray) and REE mix (pink), showing overlapped and interference-free Hf-NH

	In this example, the REE-based overlaps being investigated were known in advance, so a synthetic standard solution could be prepared, ensuring the creation of the interfering REE-NH

 product ions. For complex sample matrices, which might give rise to unknown product ions, the product ion scan can be performed for a real sample matrix. In this way, potential overlaps from all sample components can be checked and accounted for during method development.

Controlling Product Ion Formation with MS/MS vs. Single Quadrupole or Bandpass

	Product ion scanning-and the MS/MS methods developed using it - rely on Q1 operating at unit mass resolution (1 u mass window). The necessity for this is illustrated in Figure 4, where normal Agilent 8900 MS/MS operation is compared to deliberately "detuned" operation where Q1 has been set to pass a 3 u mass range. This simulates ICP-MS operation with a bandpass filter rather than a unit mass filter before the cell.

 product ions in the mass range around the preferred 

Hf product ion at m/z 258. In (a) MS-MS mode, 

 is free from any REE product ion overlap. (b) In simulated 3 u bandpass mode, 

	The top spectrum in Figure 4 shows the mass range from 

250 to 270 extracted from the full mass product ion spectra in MS/MS mode shown in Figure 3. This includes the preferred, interference-free 

 258. The lower spectrum shows the same mass range for the same product ion scans for the 10 ppb Hf and 1 ppm REE solutions, acquired with 3 u bandpass mode. With a bandpass filter before the cell, adjacent precursor ions are not rejected, so they enter the cell and react. In this example, with Q1 passing a 3 u range of masses around the set mass, 

Lu can enter the cell when the precursor ion mass is set to 

	The example of interferences on Hf product ions in a REE matrix illustrates a general limitation of incomplete rejection of adjacent masses before the cell when using highly reactive cell gases. Any nontarget ions that enter the cell may react to form overlapping product ions, giving errors on the target analyte product ion mass.

	Triple-quadrupole ICP-MS is often perceived as being more complicated and difficult to setup than single-quadrupole ICP-MS. However, the main benefit of the MS/MS tandem mass spectrometer configuration is the superior control it offers with reactive cell gas methods for interference removal. For these reaction gas methods, the ability of MS/MS to select the specific mass that is passed to the cell provides much greater control of the reaction chemistry. Rejecting nontarget, off-mass elements and isotopes using Q1 means these ions cannot enter the cell and affect the reaction processes. As a result, reaction gas methods on ICP-MS/MS are more consistent, more reliable, less matrix dependent, and therefore easier to setup than on single quadrupole ICP-MS.

	(3) L.J. Moens, F.F. Vanhaecke, D.R. Bandura, V.I. Baranov, and S.D. Tanner, 

	(4) F. Vanhaecke, L. Balcaen, I. Deconinck, I. De Schrijver, C. Marisa Almeida, and L. Moens, 

	(5) S.D. Fernandez, N Sugishama, J.R. Encinar, and A. Sanz-Medel, 

	(6) N. Sugiyama, "The accurate measurement of selenium in twelve diverse reference materials using on-line isotope dilution with the 8800 Triple Quadrupole ICP-MS in MS/MS mode" (Agilent publication 5991-0259EN, 2012).

(7) Handbook of ICP-QQQ Applications using the Agilent 8800 and 8900, 

	(9) V.V. Lavrov, V. Blagojevic, G.K. Koyanagi, G. Orlova, and D. K. Bohme, 

	(10) N. Sugiyama and K. Nakano, "Reaction data for 70 elements using O

gases with the Agilent 8800 Triple Quadrupole ICP-MS" (Agilent publication 5991-4585EN, 2014).

	(11) L. Balcaen, E. Bolea-Fernandez, M. Resano, and F. Vanhaecke, 

 are with Agilent Technologies in Stockport, United Kingdom. 

 is with Agilent Technologies International, in Tokyo, Japan. Direct correspondence to: 

Articles

Method setup and optimization steps are explored to illustrate how an ICP-MS/MS method can be defined and tested to ensure consistent performance. Users can benefit from improved interference removal performance without the complex method development inherent in the use of ion-molecule reaction chemistry

Method Development with ICP-MS/MS: Tools and Techniques to Ensure Accurate Results in Reaction Mode

	Inductively coupled plasma–mass spectrometry (ICP-MS) has good elemental coverage, low detection limits, and wide dynamic range, making it the technique of choice in many inorganic analytical laboratories. However, many sample types contain high levels of dissolved solids, which may be problematic for routine analysis using ICP-MS. Also, many new users focus on maximizing sensitivity when they optimize their systems, which can lead to degraded matrix tolerance. In this tutorial, we discuss the factors that affect the matrix tolerance of an ICP-MS instrument, and recommend steps that can be taken to optimize the key parameters.

	The maximum level of total dissolved solids (TDS) in samples to be analyzed using ICP-MS has long been accepted as 0.2%, or 2000 ppm. Above this level, drift is accelerated due to matrix deposits on the interface, analyte signals are reduced due to ionization suppression, and sensitivity is lower due to space charge effects during ion extraction and focusing.

	An example of a typical sample dilution approach might involve a soil sample prepared using 0.2 g of dried material, digested, and diluted to a final volume of 100 mL (0.2 g/100 mL = 0.2% TDS). Note that the TDS limit is usually based on the inorganic dissolved solids, since the organic content will usually be decomposed during the sample preparation. Samples or digests that contain higher than 0.2% TDS would typically be diluted to reach this level.

	As an alternative to liquid dilution, most modern ICP-MS instruments offer a facility to apply "aerosol dilution." Aerosol dilution uses an additional argon gas flow to dilute the aerosol after it emerges from the spray chamber. Together with a reduction in the nebulizer gas flow rate to reduce the amount of aerosol generated, this has the effect of diluting both the total amount of sample matrix passed to the plasma and the amount of water vapor present in the gas stream.

	Aerosol dilution is much simpler and more cost effective than an automated liquid dilution system, which typically includes multiple pumps and switching valves, all linked using tubing and connectors that can clog or leak. Aerosol dilution also offers the following benefits compared to liquid dilution:

Aerosol dilution uses argon gas, rather than a liquid, as the diluent. This eliminates the possibility of adding contaminants in the diluent, and avoids dilution errors.

The reduction in the amount of water vapor passed to the plasma means that the plasma does not expend as much energy on drying the aerosol, so the effective temperature is higher. Together with the lower level of oxygen from the reduced amount of water vapor, this leads to a reduction in the level of oxides and other polyatomic interferences.

Reduces maintenance and improves stability.

 The lower aerosol density entering the plasma reduces the localized cooling around each droplet, so the droplets are dried and decomposed more efficiently. This improved decomposition reduces the amount of undissociated matrix that reaches the interface, so interface cleaning is less frequent and less drift occurs.

Improves ionization of poorly ionized elements. 

Compared to the sensitivity obtained in a sample diluted using conventional liquid dilution, the higher temperature of the drier plasma under aerosol dilution conditions increases in the degree of ionization, particularly for poorly ionized analytes. Some of these analytes (beryllium, arsenic, selenium, cadmium, and mercury) are critical trace elements with low detection limit requirements. The effect of plasma temperature on the ionizationÂ­-and therefore sensitivityÂ­-of poorly ionized elements is illustrated in Figure 1.


	Figure 1: Approximate degree of ionization (%) for several elements at different plasma temperatures. Illustrates how changes in plasma temperature have a greater impact on sensitivity for poorly ionized elements. First ionization potential shown (eV).

Select Appropriate Sample Introduction System and Settings 

	As a rule, sample introduction hardware and operating conditions that increase sensitivity will reduce matrix tolerance. Optimizing for maximum sensitivity can therefore lead to operating conditions that do not provide sufficient robustness for routine analysis of many sample types. Sample introduction options and settings that affect matrix tolerance include:

A low-flow nebulizer operating at 200 µL/min solution flow rate gives lower sensitivity, but better matrix tolerance, compared to a nebulizer operating at 1 mL/min. If samples contain particulates, an alternative nebulizer type, such as a v-groove or parallel path, is preferred.

Spray chamber design and operating conditions.

 The spray chamber filters the primary aerosol, removing the larger droplets that would not be fully decomposed in the plasma. A double-pass or baffled spray chamber provides better filtering of the aerosol, so the droplet sizes are smaller and more uniform, improving matrix tolerance. Alternative spray chamber designs or gas flow settings may allow larger droplet sizes to pass through, which increases sensitivity, but leads to poorer matrix tolerance. Regardless of the design, cooling the walls of the spray chamber is recommended, as this condenses excess water vapor. Reducing the water vapor loading helps to maintain a higher plasma temperature.

Plasma torch injector internal diameter. 

A wider diameter injector reduces the density of the aerosol in the plasma, and therefore improves the plasma's capacity to dry the aerosol droplets and decompose the sample matrix.

 The lower matrix/vapor loading and higher plasma temperature provided by aerosol dilution improve matrix tolerance, allowing more variable, higher matrix samples to be measured accurately against simple aqueous calibrations.

Reduce Interferences and Maximize Plasma Robustness by Optimizing Plasma Conditions for Lower Cerium Oxide Ratio Formation (CeO/Ce)

	Having addressed the sample preparation and nebulizer/spray chamber/torch configuration, the plasma conditions should be optimized to maximize robustness. Plasma parameters are interrelated, with the different settings interacting to produce a trade-off between sensitivity and robustness. Plasma robustness is typically monitored using the cerium oxide (CeO/Ce) ratio. Cerium oxide is used as it is among the most strongly-bound metal-oxide ions, so a low CeO/Ce ratio indicates that the plasma has enough energy to dry the aerosol droplets, decompose the matrix, dissociate molecular species, and ionize the analyte atoms. These processes are illustrated schematically in Figure 2. Parameters that can be used to optimize the matrix tolerance of the ICP-MS plasma are:

A lower central channel gas flow rate reduces cooling at the back of the plasma, so the aerosol enters the hottest part of the plasma sooner, allowing more time for sample decomposition. For a given carrier gas flow, a wider torch injector also effectively increases the time the aerosol droplets spend in the hottest part of the plasma.

Increasing the distance between the load coil (plasma) and interface sampling cone allows more time for matrix decomposition and analyte ionization.

Higher RF power means higher plasma temperature, and therefore better matrix decomposition. The temperature of the plasma central channel (where the aerosol droplets are) is also affected by the RF frequency. Lower RF frequency (27 MHz used on most ICP-MS systems) gives a higher central channel temperature than higher RF frequency (40 MHz as commonly used for inductively coupled plasma–optical emission spectrometry (ICP-OES).


	Figure 2: Schematic illustration of the processes in the ICP-MS plasma, from aerosol droplet drying to analyte ionization.

Verify that the Selected Configuration and Settings Allow the Required Method Detection Limits (DLs) to Be Achieved for Critical Trace Elements 

	Optimizing the hardware configuration and operating parameters for robustness leads to reduced sensitivity. Consequently, the underlying detection limit performance must be checked under the selected robust method conditions.

	Modern ICP-MS systems are capable of extremely high sensitivity when optimized for sensitivity rather than matrix tolerance. Optimizing for matrix tolerance as described above will reduce this sensitivity but should still allow the method requirements to be met. This needs to be confirmed using performance tests such as low level spike recoveries and assessment of method detection limit, as appropriate for the application.

Confirm Chosen Settings Allow High-level Analytes and Major Elements to Be Measured 

	An important consideration for multielement analysis is the capability of the ICP-MS to measure the major elements together with the trace analytes. For many commercial laboratories, this is a key requirement for maximizing productivity by allowing all the required elements to be determined in a single measurement. For this to be achieved routinely, the dynamic range of the ICP-MS detector must be as wide as possible.

	Modern ICP-MS systems use a detector with a dynamic range of up to about 10 or 11 orders of magnitude. This ensures that the highest concentration major elements can be measured without requiring customized settings to attenuate the most intense signals. User-defined signal attenuation is not easy to apply when the sample types being measured are of variable and unknown composition, and where the major elements are not known in advance.

Monitor Long-term Stability Using Periodic Check Standards or Internal Standard Signal Responses 

	Once the ICP-MS has been optimized for robustness, and detection limits and dynamic range have been confirmed, method performance should be verified by analyzing a representative batch of samples.

	Most regulated methods will include analysis of periodic check standards or similar in-run quality control (QC) to check that the calibration remains valid, and the signal has not drifted out of control. A periodic check standard is useful to confirm the ongoing validity of the calibration, but a better indication of the true sensitivity during a sample sequence can be obtained from plotting the internal standard (ISTD) signals for each sample. The ISTD signals are not corrected, so indicate any changes in sensitivity of the instrument due to drift or suppression, for every solution run throughout the entire batch.

	For ICP-MS analyses, ISTDs are usually added automatically using an online mixing T-connector. If several ISTD elements are added, the ISTD plot can also indicate any mass-dependency in the signal change as well as enabling contamination of an ISTD element in a sample to be identified.

 is an ICP-MS Product Specialist for Agilent Technologies in Cheshire, United Kingdom. Direct correspondence to: 

ICP-MS: Essential Steps to Optimize Matrix Tolerance

) nanoparticles (NPs) are widely used in industry, manufactured goods, and consumer products, such as food additives. However, there are several significant challenges associated with the measurement of silicon by inductively coupled plasma–mass spectrometry (ICP-MS). First, silicon has a moderately high first ionization potential (8.152 eV), which means it is less than 70% ionized in a typical argon plasma, so the sensitivity is significantly lower than for a fully ionized element. In addition, background polyatomic ions affect all three silicon isotopes, 

Si, with the most intense interferences (

Si isotope. Finally, silicon-based materials are widely used in manufactured goods and industrial materials ranging from glass, alloys, and air and water filters to organic compounds such as silicone tubing, sealants, adhesives, lubricants, and coatings. Consequently, maintaining a low background for silicon in the analytical laboratory can be highly problematic. When characterizing silica in nanoparticle form, the analysis is especially difficult. In this article, we describe a method to measure ultrafine silica nanoparticles using ICP–tandem MS (ICP-MS/MS) to control the elemental and polyatomic ion backgrounds. The method provides high sensitivity to ensure accurate analysis at the small (~50 nm) particle sizes of interest in environmental and biological applications.

	Nanomaterials are increasingly used in industrial processes, manufactured goods, medicines, and as ingredients in consumer products such as cosmetics, sunscreen, and food. The measurement of nanoparticles (NPs) is a subject of interest because the fate of NPs in the environment and the potential for toxic effects once absorbed into biological systems are not yet well understood.

	Some of the most common types of NPs are difficult to measure accurately using conventional single-quadrupole inductively coupled plasma–mass spectrometry (ICP-MS). Such instruments often cannot achieve the combination of high sensitivity and low background that is required to allow the signal generated from a small particle to be distinguished from the background. Much of the work reported thus far has concentrated on particles that contain elements such as zirconium, silver, and gold (1,2), which are relatively easy to measure by ICP-MS. However, many of the natural and engineered NPs of interest are composed of elements that are much more difficult to measure by ICP-MS-for example, iron, sulfur, titanium, and silicon.

	Nanostructured synthetic amorphous silica (SAS) is one of the most commonly used nanomaterials. It is an authorized food additive (E-551) and is also found in a wide variety of industrial and consumer products from polymers, composites, paints, and textiles to toothpastes, detergents, and cosmetics. Silica NPs also have great potential for a variety of other uses, including silica aerogels as thermal and electrical insulators, and imaging and therapeutic applications in medicine (3). Reliable analytical methods are needed to allow monitoring of compliance with upcoming regulations (4) and for the assessment of nano-safety (5). However, Si measurement by ICP-MS is not easy because of the combination of relatively low sensitivity and high spectral background. Silicon is only about 70% ionized and the major isotope (

Si – 92.23% abundance) is interfered by the intense background polyatomic ions CO

. These limitations mean that conventional single-quadrupole ICP-MS is unable to detect silica particles with diameters smaller than about 200 nm (6). This capability is not sufficient for characterization of the nanoscale particle sizes of interest in emerging applications and for regulatory requirements; the accepted definition of a nanomaterial is one with major dimensions of 100 nm or less (7).

 can be resolved using a reactive cell gas such as hydrogen (H

) in the collision–reaction cell (CRC) of an ICP-MS system. However, ion–molecule reaction chemistry can only be controlled effectively using a tandem mass spectrometer, where an additional mass filter is situated before the CRC of the ICP-MS system. This extra mass filter enables MS/MS operation, where nontarget ions are prevented from entering the cell. This approach ensures that reaction processes and the resulting reaction product ions are consistent, even when the sample composition is complex or variable. The general layout and performance capability of a triple-quadrupole configuration of tandem ICP-MS is described in the literature (8).

 NPs were measured by single-particle ICP-MS (spICP-MS), using MS/MS mode and H

 reaction gas on a triple-quadrupole ICP-MS system. The instrument software included an optional single nanoparticle application module for method development, signal processing, and data interpretation. The practical applications capabilities of the instrument for the measurement of a range of SiO

	An Agilent 8900 ICP-QQQ system (model #100, Advanced Applications Configuration) was used for the analysis of silica nanoparticles. The instrument was equipped with a glass concentric nebulizer, a quartz spray chamber, a quartz torch with a 1-mm injector, and standard nickel sampling and skimmer cones. Samples were introduced with a standard peristaltic pump and pump tubing (1.02 mm i.d.). The configuration of the instrument used has a novel argon gas delivery path that uses special inert materials to minimize contamination of the gas supply through the instrument, reducing the elemental background for common contaminant elements Si and S. 

Si was measured in continuous fast time-resolved analysis (fast TRA) mode, using a dwell time of 0.1 ms (100 μs) per point with no settling time between measurements. The 

Si signal was measured on-mass in MS/MS mode, with both quadrupoles (Q1 and Q2) set to 

 28. Hydrogen cell gas was used to resolve the on-mass polyatomic interferences, N

 and CO. The main operating conditions of the instrument are shown in Table sI online; please visit 

www.spectroscopyonline.com/supplementary-information-accurate-measurement-ultrafine-silica-nanoparticles-using-icp-msms

Reference Materials and Sample Preparation 

 NP reference materials with nominal diameters of 50, 60, 100, and 200 nm determined by transmission electron microscopy (TEM) were purchased from nanoComposix. The reference materials were diluted with deionized (DI) water to give particle concentrations of 40, 40, 100, and 1000 ng/L (ppt) for the 50-, 60-, 100-, and 200-nm particles, respectively. The solutions were sonicated for 5 min before analysis to ensure sample homogeneity. To allow the elemental response to be calculated, a Si ionic standard of 5 μg/L (ppb) was prepared in DI water. MS/MS with H

 cell gas should be effective in removing the CO

 polyatomic overlap. To check this interference removal capability, a SiO

 NP solution containing a matrix of 1% ethanol was also prepared.

	NPs measured using TRA with ICP-MS give a narrow peak or signal plume lasting around 0.5 ms for each particle that passes through the plasma; the peak signal area of each particle plume is proportional to the particle mass. The signals measured in DI water (without particles) and in solutions containing SiO

 particles with diameters of 50, 60, and 100 nm are shown in Figure 1. Fast TRA mode allows multiple measurements to be made during the signal peak from each individual particle. The shape and duration of the ion plume from each NP can be measured, and the total ion count can be converted to a particle mass and therefore diameter (assuming spherical particles). The Single Nanoparticle Application module of the ICP-MS MassHunter software includes the calculations to convert signal intensity to particle diameter, along with other key parameters, such as minimum detectable particle size and background equivalent diameter. Because of the high sensitivity, low elemental background, and effective interference removal of the ICP-MS instrument used, the smallest (50-nm) SiO

 particles provided clear signal peaks (Figure 1b) that were easily distinguished above the low background signal seen in DI water (Figure 1a). As expected, the 60-nm and 100-nm particle solutions gave proportionally higher signal peaks as shown in Figures 1c and 1d, respectively.

	Frequency distributions were calculated automatically by the instrument software from the signals measured for the four SiO

 nanoparticle solutions shown in Figure 1; the frequency distribution plots are shown in Figure 2. The particle signals were clearly separated from the background caused by the dissolved  ionic component, although the background and particle signals were partially overlapped for the 50-nm particles (shown in Figure 2b). From these results, we can estimate that the practical detection limit for the particle diameter was below 50 nm. The background equivalent diameter (BED) calculated by the instrument software from the 50-nm particle solution was 22 nm.

	The particle size distribution results for the different SiO

 NP solutions are shown in Figure 3. The plots confirm that all the SiO

 particle sizes could be determined accurately, including the 50 nm NPs. The high sensitivity and low background of triple-quadrupole ICP-MS combined with the effective removal of polyatomic interferences in MS/MS mode with H

 cell gas, enabled the smaller NP sizes to be clearly distinguished from the background signal. All the particle size plots demonstrated a Gaussian distribution, which agrees with the information provided by the supplier of the NP solutions. The triple-quadrupole ICP-MS results obtained for median, mode, and mean particle sizes also agreed well with the reference sizes obtained by TEM (Table I). Also, the measured particle concentrations agreed well with the theoretical particle concentrations, although the concentration result for the 50-nm solution was slightly high (49 ng/L compared to an expected value of 40 ng/L) because of the overlap with the background signal.

	Figure 4 shows the mean signal intensity measured by spICP-MS/MS plotted against the reference particle diameter determined by TEM. The logarithmic plot has a slope of 2.84, which is in good agreement with the theoretical cube relationship between diameter and mass indicated in equation 1: 

	Equation 1 is used to calculate particle mass (

Resolving Carbon Interference Using MS/MS and H

	Real samples of interest-such as biological materials, food matrices, polymers, pharmaceutical ingredients, and organic solvents-may contain a high level of carbon, which causes a 

Si. Triple-quadrupole ICP-MS in MS/MS mode with H

 cell gas can eliminate this interference effectively. A mixed solution of the 100-nm and 200-nm SiO

 NPs was prepared in a sample containing 1% ethanol. The particle size distribution for this sample is shown in Figure 5. Despite the high concentration of carbon, the size distribution for each group of particle sizes was consistent with the result obtained by TEM, and the different particle sizes were separated with excellent resolution. This result suggests that the spICP-MS/MS technique can resolve the carbon interference successfully, allowing accurate determination of particle sizes for SiO

 nanoparticles in natural sample matrices.

	Triple-quadrupole ICP-MS operating in MS/MS mode with H

 cell gas was used for the successful determination of SiO

 nanoparticles at sizes smaller than 100 nm. The triple-quadrupole ICP-MS method delivered exceptionally low background signals and high sensitivity, allowing the smallest (50-nm) SiO

 particles to be distinguished from the background signal. MS/MS operation gave effective control of the H

 reaction chemistry, eliminating polyatomic interferences from N

 and CO, even in the presence of a high level of carbon matrix. Dedicated software was used to calculate the particle sizes, providing accurate results for individual NP reference materials and mixed solutions. The spICP-MS/MS method provided fast analysis times, excellent detection limits for particle size and concentration, and accurate results for SiO

		J. Tuoriniemi, G. Cornelis, and M. Hassellöv, 

		C. Degueldre, P.-Y. Favarger, and C. Bitea, 

		D. Napierska, L.C.J. Thomassen, D. Lison, J.A. Martens, and P.H. Hoet, 

		L. Calzolai, D. Gilliland, and F. Rossi, 

		S. Dekkers, A.G. Oomen, E.A. Bleeker, R.J. Vandebriel, C. Micheletti, J. Cabellos, G. Janer, N. Fuentes, S. Vázquez-Campos, T. Borges, and M.J. Silva, 

		M.D. Montaño, B.J Majestic, Å.K. Jämting, P. Westerhoff, and J.F. Ranville, 

		International Organization for Standardization, 

ISO TS 80004-1:2010: Nanotechnologies - Vocabulary - Part 1: Core terms,

		L. Balcaen, E. Bolea-Fernandez, M. Resano, and F. Vanhaecke, 

 is with Agilent Technologies in Stockport, UK. 

 is with Agilent Technologies in Tokyo, Japan. 

 is with Agilent Technologies in Bellevue, Washington. Direct correspondence to: 

A method to measure ultrafine silica nanoparticles using ICP-MS/MS to control the elemental and polyatomic ion backgrounds is described here.

Accurate Measurement of Ultrafine Silica Nanoparticles Using ICP-MS/MS

Titanium (Ti) is used in lightweight alloys, catalysts, cutting and grinding tools, and medical implants. Most Ti is refined into the white pigment titanium dioxide (TiO

), used in paint, paper, plastics, food, and consumer products such as toothpaste and sunscreens. TiO

nanoparticles are much more reactive and bioavailable, so they may have adverse effects on human health and the environment. There is a need for analytical techniques and methods for accurate, low level analysis of Ti in environmental samples, biological tissues, and other samples. This application is difficult for conventional quadrupole inductively coupled plasma-mass spectrometry (ICP-MS) because of the presence of matrix-based spectral interferences on the isotopes of Ti. In this article, we show how the interferences can be removed using tandem ICP-MS (ICP-MS-MS). This approach allows sub-nanogram-per-liter (part-per-trillion) detection limits to be achieved and ensures that accuracy is maintained in the presence of complex sample matrices.

	Titanium (Ti) is an important element for high-performance alloys used in automotive, consumer, and medical products, and especially in aerospace applications, where light weight, high strength, and tolerance of extremely high temperatures are important factors. However, most titanium is refined into titanium dioxide powder (TiO

), which is used as a whitener or anticaking agent in industrial and consumer products such as paints, paper, plastics, food, cosmetics, and sunscreen. Fine TiO

 particles (>0.1 mm in diameter) are known as pigment grade, and smaller (<0.1 mm) ultrafine TiO

 particles are classified as nanoparticles. Although TiO

 powder is considered a poorly soluble, low-toxicity (PSLT) material, there is some evidence of it causing lung damage including tumor growth from respiratory exposure, with damaging effects being greater for smaller particles (1). Ti is also used in medical devices and implants such as joint replacements, so methods are needed to monitor the possible degradation of these implants.

	Quadrupole inductively coupled plasma–mass spectrometry (ICP-QMS) is a standard technique for analyzing trace metals in many sample types including environmental, food, and biological materials. ICP-MS can measure Ti at low levels, although three of its five isotopes (

Ti) are overlapped by minor isotopes of other elements (

Ti (5.41% abundance) are therefore typically used for Ti quantitation by ICP-MS, so sensitivity and detection limits are poorer than they would be if a major isotope was available. In sample matrices that contain high levels of Ca, P, C, Cl, or S, common matrix elements in many natural materials, all of the Ti isotopes can also be affected by polyatomic interferences including CaH

	ICP-QMS with collision-reaction cell (CRC) technology operating in helium (He) mode is able to reduce most common polyatomic overlaps to background or negligible levels (3). But small residual interferences may remain, affecting the measurement of certain analytes at trace levels in some matrices. The presence of residual interferences has led to interest in using reactive cell gases to improve interference removal and provide lower detection limits for specific elements.

	Reaction gas methods can use either on-mass or mass-shift measurement. On-mass measurement is used when the interfering ions are more reactive than the analyte ion; the interferences are neutralized or react to form new product ions at a different mass, leaving the analyte ion free to be measured at its original mass. Mass shift mode is used when the analyte ion is more reactive than the interferences; the analyte is measured as a new, cell-formed product ion, free from the original overlap. Either of these reaction cell modes can be used with conventional ICP-QMS, but the single mass analyzer configuration does not provide any control of the ions that enter the cell. As a result, all ions are able to react with the cell gas, leading to the formation of many new product ions, some of which will overlap the analyte ions of interest. Even the presence of low mass ions such as Ar

, which are ubiquitous in the ICP-MS ion beam in solution analysis, will lead to reaction product ions that may overlap the higher masses where analyte product ions are measured, especially when a highly reactive cell gas such as ammonia (NH

) is used (4). In practice, this means that the analyte product ions measured in reaction mode with ICP-QMS may in fact consist of several different product ions derived from different analyte isotopes, and other analyte, matrix and solvent or plasma gas elements. The errors caused may not be apparent during method development, since the product ions will vary depending on the matrix. Also, reaction chemistry with ICP-QMS is usually specific to one isotope of the analyte, so there is often no independent confirmation of the result based on a secondary or qualifier ion, as commonly used in organic mass spectrometry, and ICP-MS with He cell gas.

	An ICP-QMS system with a quadrupole ion guide in the cell can reject masses above and below a user-set mass. This approach, known as a 

, can be used to reject low (and high) mass ions that might react in the cell to form new interfering ions, but the mass window limits cannot be set very close to the mass of the analyte precursor ion, or severe loss of analyte sensitivity occurs (5). Thus, bandpass operation cannot reject existing ions at the intended analyte product ion mass, nor can it differentiate between reaction product ions that occur at the same mass, but are formed from different isotopes or elements. An example might be the overlap of the NH3 product ions [

. This effect is apparent in the poor isotopic template fit observed when the reaction product ions from several isotopes of the same element are measured in ICP-QMS (6).

	MS-MS mode is an alternative mode of operation available on a triple-quadrupole ICP-MS. With MS-MS, the ions that enter the cell are controlled by an additional quadrupole mass spectrometer (Q1), situated before the CRC. Q1 can be operated as a mass filter, allowing only the target analyte ions and any on-mass interferences to enter the cell. This gives far better control of reaction processes, and allows the use of highly reactive cell gases without the matrix-dependent variability that occurs with ICP-QMS. A further benefit of MS-MS mode is that, since each analyte ion enters the cell in isolation (without any other isotopes of the same element), and the mass shift transition is specific, as defined by the mass difference between Q1 and Q2 (for example, + 16 amu for an O-atom addition reaction), the analyte’s original isotopic abundance pattern is preserved in the product ion spectrum. This allows results to be verified from multiple analyte isotopes, and also enables isotope-based measurements such as isotope ratio and isotope dilution to be used (7).

	In this article, we compare the performance of MS-MS mode and single-quadrupole (SQ) mode (comparable to conventional quadrupole ICP-MS) for Ti analysis. Using both O

 as reaction gases, we investigate Ti measurement in the presence of potential interferences from simple, single-element matrices, and complex matrices.

	A triple-quadrupole ICP-MS system (Agilent Technologies model 8800 ICP-QQQ) was used for all measurements. The instrument configuration and operating modes have been described previously (8). The standard configuration with Micromist glass concentric nebulizer, quartz sample introduction system, and Ni interface cones was used. Operating conditions are shown in Tables I and II.

Calibration Standards and Sample Preparation

	The standards used for this work were prepared using single-element stock solutions (Spex CertiPrep, Claritas grade).  For the TiO

 experiment, the acids were UpA ultrapure grade (Merck).

	For the biological matrix experiment, a 0.5-mL sample was diluted with 2.5 mL of diluent mix (0.01% EDTA, Sigma Aldrich 99.999%); 0.005% Triton X-100 (Romil SpA); 4% BuOH (Merck Uvasolv); 1% NH

Titanium (Ti) is used in lightweight alloys, catalysts, cutting and grinding tools, and medical implants. Most Ti is refined into the white pigment titanium dioxide (TiO2), used in paint, paper, plastics and food, and in consumer products such as toothpaste and sunscreens.
TiO2 is considered inert and safe, but TiO2 nanoparticles are much more reactive and bioavailable, so they may have adverse effects on human health and the environment. There is a need for analytical techniques and methods for accurate, low level analysis of Ti in environmental samples, biological tissues, and other samples. This application is difficult for conventional quadrupole ICP-MS because of the presence of matrix-based spectral interferences on the isotopes of Ti. In this paper, we show how the interferences can be removed using tandem ICP-MS (ICP-MS-MS). This allows sub-ng/L (ppt) detection limits to be achieved, and ensures that accuracy is maintained in the presence of complex sample matrices

Author | Ed McCurdy

Articles