George Chan, Winner of Spectroscopy’s Inaugural Emerging Leader in Atomic Spectroscopy Award, Focuses on ICP Matrix Effects and More

February 1, 2017

Spectroscopy is proud to have created a new award, the Emerging Leader in Atomic Spectroscopy Award. As its name implies, the award recognizes a young scientist, and it is designed to encourage the next generation of atomic spectroscopists. George Chan, the winner of the inaugural Emerging Leader in Atomic Spectroscopy Award, is a project scientist at Lawrence Berkeley National Laboratory (LBNL) in Berkeley, California. One of Chan’s most significant contributions to the field of atomic spectroscopy has been his work on matrix effects and excitation processes in the inductively coupled plasma (ICP).

 

Spectroscopy is proud to have created a new award, the Emerging Leader in Atomic Spectroscopy Award. As its name implies, the award recognizes a young scientist, and it is designed to encourage the next generation of atomic spectroscopists. George Chan, the winner of the inaugural Emerging Leader in Atomic Spectroscopy Award, is a project scientist at Lawrence Berkeley National Laboratory (LBNL) in Berkeley, California. One of Chan’s most significant contributions to the field of atomic spectroscopy has been his work on matrix effects and excitation processes in the inductively coupled plasma (ICP).  

The award will be presented to Chan at the 2017 European Winter Conference on Plasma Spectrochemistry (EWCPS), where he will give a plenary lecture. Chan recently spoke to Spectroscopy about his scientific background, interests, and recent work.

Where or how did your interest in analytical chemistry and atomic spectroscopy begin?

How I developed my interest in analytical chemistry and atomic spectrometry is a completely unplanned accidental journey. First, a little bit of background about me: I was born and raised in Hong Kong, and I knew that I wanted to become a chemist at the age of around 14 or 15. Naturally, I picked a chemistry major BSc program at the University of Hong Kong (HKU). Following the British system, it was a three-year program. During my first two years at the university as an undergraduate, what interested me most was neither analytical nor physical chemistry, but rather inorganic and organic chemistry. In fact, at that time, I wanted to become a synthetic chemist. This was not surprising and similar thinking was quite common among my classmates for several reasons. First, the research focus of the chemistry department at HKU at that time, in terms of the number of faculty and the sizes of their research groups, was on synthetic (inorganic and organic) chemistry, and analytical chemistry was underrepresented (and perhaps viewed by some as second tier). Second, analytical instrumentation was not that user friendly at that time (for example, chart recorders prone to paper jams). In the inorganic and organic laboratory classes, we readily followed the experiment as it proceeded (for example, color change during reflux, reaction product precipitated out from solvent). In contrast, the chemical instrumentation, in many cases, appeared as black boxes to us. We put the sample in and got a reading, but we did not really see how the instrument operated internally.

What caused a dramatic change was a summer internship in a chemical analysis laboratory outside the university. My job was to develop a method to analyze a suite of elements in stainless steel samples with inductively coupled plasma–atomic emission spectroscopy (ICP-AES)-a technique that was completely unknown to me at that time. I started to learn the technique through reading books and research articles as well as through hands-on experience with the various samples. I still remember a vivid learning experience: I observed that steel samples digested with different concentrations of sulfuric acid behaved differently; signals varied by more than 30% depending on the amount of sulfuric acid used in the digestion. I told my supervisor what I observed, and both of us were puzzled because we had the misconception that ICP analyses are relatively free from matrix effects. After researching this phenomenon in the literature, I realized that what I observed was a matrix effect from the sample introduction system. All of these experiences changed my view of what modern analytical chemistry and instrumentation look like. But the most important influence that made me look into specializing in analytical chemistry and atomic spectrometry was a casual chat with two colleagues after lunch one day. Looking retrospectively, that casual, 20-minute chat was in fact a life-changing conversation.

One of those two colleagues was my internship supervisor, who is a theoretical physical chemist by training, and the other is an organic chemist; both were then-recent PhD graduates from HKU. One afternoon, I casually asked their opinions on course selection. After learning about my strong intention to do a final-year undergraduate research project, they shared their research experience with me. The organic chemist advised me that synthetic chemistry in research might not be as much fun as I thought, whereas the theoretical chemist told me that research in physical and analytical chemistry are very different, both in scope and the equipment involved, than those encountered in teaching laboratories. The most important outcome of the conversation was that they both urged me to talk to an analytical professor before making a final decision. Because I was working with ICP during that summer, I made an appointment with Professor Wing-Tat Chan, whose research interests are in ICP and analytical atomic spectrometry in general. The research topic that he offered for undergraduate research involved capacitively coupled plasma-another technique completely unknown to me, but one that sounded challenging and interesting. I agreed to join and he welcomed me to his group as an undergraduate student. During the course of the project, there were problems with the instrument on several occasions , and Wing-Tat worked with me on repairing the instrument (more correctly, showed me how to do it). During those repair jobs, we took the instrument cover off and he explained to me the different parts inside the instrument and how the light paths go; he even opened the spectrometer cover and showed me what is meant by zero-order of a grating. Those are the experiences that would normally never be encountered in an undergraduate teaching laboratory, and they completely changed my view of analytical chemistry and instrumental analysis. In fact, I started to appreciate the beauty of the logic behind the designs of analytical instruments; they are definitely not black boxes, as I had naively thought. After the marvelous experience in my final year project, I decided to stay in Professor Wing-Tat Chan’s group to do a master’s degree in analytical atomic spectrometry, and never thought again of becoming a synthetic chemist.

What would you consider to be the greatest advances in inductively coupled plasma-mass spectrometry (ICP-MS) and ICP-AES over the past 10 years?

I can think of several. The design and commercialization of an array-ion detector ICP-MS system that allows simultaneous measurements of the full inorganic mass spectrum from 6Li to 238U is, in my view, the most significant advance in ICP-MS instrumentation in the last decade. The power of simultaneous multielement analysis with array detectors covering the whole ultraviolet–visible (UV–vis) wavelength range has long been realized in ICP-AES, and similar technology and capability finally arrived for ICP-MS. 

Another advance is the use of ICP-MS or ICP-AES for the analysis of individual single entities (such as nanoparticles, biological cells, and dust aerosols). In this analytical approach, termed single-particle ICP (sp-ICP), it becomes feasible to obtain sample information (for example, particle-size distributions and their stoichiometry) that is otherwise unavailable from simple bulk chemical analyses. 

Although still in its infant stage, the low-gas-flow ICP torch developed by the research group of Wolfgang Buscher at the University of Münster (1) is innovative; this novel ICP torch consumes only 1.5 L/min of argon but with detection capability approaching that attainable by conventional ICP torches.

 

 

 

You have done some significant work with matrix effects and excitation processes in ICP, where you discovered several new phenomena and mechanisms. What research led to the discovery that low second ionization potentials (IPs) produce the strongest matrix effect, even though it had been widely reported in the literature that elements with low first ionization potentials produced the strongest effects? What were the biggest challenges in that research? What benefits does it bring to the field?

That work on matrix elements with low second ionization potentials producing more severe matrix interferences in ICP started with some experiments that I performed at the Lawrence Berkeley National Laboratory (2) as an exchange student during the course of my master’s degree research. The initial planned objective of the experiments was to study another phenomenon-namely, a mass loading effect during laser ablation (LA)-ICP. Some pressed pellets of carbonates, chlorides, or oxides of Group I and II elements were individually introduced into the ICP by laser ablation. The objective was to understand how much introduced mass the ICP can handle before an observable change in plasma condition is induced. The results showed that the ICP can handle quite a large amount of Group I (lithium,  sodium, and potassium) salts without changing its conditions, but it is very sensitive even to a relatively small amount of Group II (magnesium, calcium, and strontium) salts. At that point, we believed that this was a mass loading effect and tried to correlate the effect with physical properties (for example, enthalpy of formation or enthalpy of atomization) of the solid. No single parameter really correlated well, except the second ionization potential. 

The second ionization potential (second IP) argument faced two difficulties: It is not a factor typically considered in the literature and we had not studied enough matrix elements (that is, we had insufficient data points) to be conclusive. With a very simple mind, I tested almost all the inorganic salts that I found in the chemical storage cabinet (the collection was not very large). Out of the 31 matrix elements studied, 11 elements exhibited low second ionization potentials and it was exactly these 11 samples that induced more severe matrix effects. It was a strong suggestion that the second IPs of the elements have an important role in defining the effect. The biggest challenge at that time was the comparatively low spectral resolution of the spectrometer that I used because many of these low second IP matrices are the rare-earth elements, whose emission spectra are very complex. Avoiding spectral interference from the matrix was a bit challenging. Also, although there was a strong suggestion that the severe matrix effect is related to the low second IP of the matrix, the relationship between the extents of matrix interference and the magnitudes of the second IP is not monotonic, indicating that there is at least one additional unknown factor. Understanding the mechanism was a challenge because there was still at least one unknown factor, and there was no clear hint about how the second IP would fit into the known models of plasma excitation and matrix effects. Before this work, there were reports in the literature stating that matrix effects from Ca are more severe than those of Na. This work “explained” the effect from the viewpoint of low second IP of Ca. In my view, the biggest benefit this work brings to the field is to inform the community that, despite the fact that ICP is being routinely used nowadays, our knowledge of fundamental ICP mechanisms is not complete and lots of work still needs to be done to understand this plasma source.

You found that the presence or absence of low-lying energy levels in the doubly charged matrix ion govern the severity of matrix effects. What research led to the discovery? How does this behavior relate to Penning ionization?

Let me start by saying that there is still controversy about the role and importance of Penning ionization in ICP. When I was a graduate student at Indiana University, the questions of how these low second ionization potential (second IP) elements produce more serious matrix interferences, and what is the identity of that unknown additional factor that governs the severity of matrix effects, were intermittently on my mind. There was a simultaneous multichannel echelle-grating ICP spectrometer, covering almost the entire UV–near infrared (NIR) range, sitting in the laboratory, and I had this idea that if I could follow a large pool of emission lines and see how each of them were affected by the matrix, and if the matrix pool were large enough, the results might help us to deduce the mechanism of this low second IP matrix effect. Because the focus of the study was to reveal the presence of any ‘‘hidden’’ secondary parameter and to deduce the matrix effect mechanism, I studied all the low second IP matrix elements that are practically feasible (that is, excluding the non-naturally occurring or radioactive ones) (3). With this expanded approach, it was very clear that seven low second IP matrix elements induce less severe matrix effects and all of them are with simple electronic structure. The most striking exception is ytterbium (a rare-earth element); when doubly ionized, all its electronic subshells are completely filled. With this hint, it was natural to correlate the matrix effects with electronic structure; a good correlation between magnitudes of matrix effect and presence of low-lying energy levels in the doubly charged matrix ions was found. In other words, the presence of low-lying energy levels in the doubly charged matrix ion is another critical parameter that defines the magnitude of the matrix effect.

With this additional piece of information, I started to solve the matrix-effect mechanism puzzle through matching with known analyte excitation mechanisms discussed in the literature (for example, electron impact, charge transfer, Penning ionization), and Penning ionization was found to fit the observations. In Penning ionization, the analyte is ionized and excited by argon-excited species (not necessarily limited to those that are metastable). A similar reaction also applies to the matrix (and thus uses up some Ar excited states that would otherwise be available to the ionization of the analyte). If the second IP of the matrix is low, second ionization of the matrix from its singly charged ion is also possible by Penning ionization. However, the net effect on Ar-quenching is not equivalent to simply doubling the matrix concentration because of ion-electron recombination. Ion-electron recombination has a square and cubic dependency on the ionic charge, depending on whether it is radiative or three-body recombination. So the recombination rate for doubly charged matrix ion is four to eight times faster than that of the singly charged matrix ion. The doubly charged matrix ion will be rapidly recombined to the singly charged ion and be available again for further quenching of Ar-species. Through such cycling, the excited Ar-species are significantly quenched and lead to the matrix effect for the ionic emission line of the analyte.  

There are two ways that the presence of the low-lying energy levels in the doubly charged matrix ion can affect the extent of matrix effects. First, the presence of more low-energy levels in the doubly charged matrix ion will increase the number of possible reaction routes for Penning ionization and may lead to a larger overall reaction cross-section. Second, the cross-sections for all feasible Penning ionization routes are not identical. With more feasible reaction routes, there is a higher likelihood to have one particular reaction route that has a very favorable reaction cross-section and increases the overall reaction rate.

What other work have you done to help with understanding fundamental mechanisms in the ICP?

There are two ideas that immediately come to my mind. The first one is on using matrix effects as a probe for the study of the charge-transfer mechanism in ICP (4). The concept is simple: We know that matrix effects from low second IP elements (like Ca or Ba) are more severe and affect mostly ionic emission lines. Under the charge-transfer mechanism, a high-energy ionic emission line of the analyte (M+*) is directly ionized and excited from its neutral (M) by Ar+ ion in a one-step reaction: M + Ar+ ® M+* + Ar. Because of this close linkage, these high-energy ionic lines (M+*) behave similarly to neutral-atomic lines. Hence, when you plot the extent of matrix effects (for example, relative emission intensities in the presence of Ca or Ba matrix) versus the total excitation potentials of the ionic lines, it is common to see an initially decreasing trend (that is, a more severe matrix effect as the excitation potential increases) but then a sudden decrease in matrix interference (that is, a jump in relative intensity). This jump signifies charge-transfer behavior for those emission lines. The use of matrix effects to probe charge-transfer reactions greatly simplifies the instrumentation and procedures, and hence, a large number of emission lines can be studied in a relatively short time.

 

 

 

Another study involved the development of an automated diagnostic tool for flagging matrix interferences in ICP-AES measurements. Very briefly, the spatial heterogeneity of the ICP is exploited and spatial emission maps are used to flag matrix interferences and system drift. Because of the plasma heterogeneity, the relative magnitude and even the direction of the changes in emission intensity caused by a matrix interference are also spatial-location dependent. Such a spatially dependent matrix effect will cause a corresponding alteration in the determined concentration-that is, the apparent analyte concentration will also vary, depending on the measurement location in the ICP, under the influence of matrix interference, and thus allowing the matrix interference be recognized. Interested readers can refer to the April 2015 issue of Spectroscopy magazine (5), in which a more detailed account of this methodological development was covered.

Can you tell us about how are you expanding your work with optical isotopic analysis, such as using multiple emission lines in the isotopic analysis of uranium by laser-induced breakdown spectroscopy (LIBS)?

The idea of using multiple emission lines in isotopic analysis of uranium is somewhat similar to the use of multiple emission lines in elemental determination by ICP, in which improvements in detection limit and precision and better immunity to withstand interferences have been demonstrated. It has been known for decades that isotopic shifts in some atomic emission lines of uranium are large enough and readily measurable even with an excitation source running under atmospheric pressure, in which several line-broadening mechanisms are operating. Isotopic analysis of uranium with atomic emission, in particular with LIBS, offers several advantages, such as no sample preparation, the possibility of remote analysis, and no radioactive contamination of the spectrometer. It is interesting to note that almost all optical isotopic analyses of U used only a very small set (one to several pairs) of emission lines, whereas U emits thousands of lines and many of these lines contain isotopic information. In other words, many information bandwidths are not utilized. It is true that many of these emission lines are overlapping; however, with the advancement in chemometric computer algorithms (such as partial least squares [PLS]), the extraction of useful information (in this particular case referring to the 235U/238U isotopic ratio of the sample) from a large pool of spectral data is quite straightforward in practice. As PLS calibration is a purely empirical approach, it functions properly on spectral features that are overlapped and only partially resolved, and requires neither the knowledge of the fundamental spectroscopic constants (for example, transition probabilities and isotopic shifts) for these emission lines nor any fundamental physical properties of the plasma (such as temperature and electron number density). Therefore, the use of multiple emission lines for U isotopic analysis with multivariate calibration is something worthy of exploring.

Our first work in this area was to use computer simulation to understand the potential gain in analytical performances for U isotopic analysis, if multiple emission lines are used. The model also helps us to understand, among net signal, signal-to-background ratio, and isotopic shift, which parameters are the most dominating factors defining the accuracy and precision of isotopic analysis by LIBS. The results showed that signal strength is a more important factor than the magnitude of isotopic shifts in influencing the analytical performance of U isotopic analysis by LIBS. This work is still on-going.  

Your work has also focused on laser-ablation molecular isotopic spectrometry (LAMIS) and ambient-ionization mass spectrometry. What specific developments are you hoping to achieve in these areas?

The fact that molecular rovibronic spectra depend on the masses of the isotopomers and exhibit magnified isotopic shifts (the underlying scientific principle of LAMIS) compared to those in atomic electronic transitions has been known for decades. However, molecular rovibronic spectra are often complex and not well-resolved; thus, their uses for isotopic analyses were previously limited to some special cases in which the bandheads were well separated. Modern computer power greatly changes the way data processing is done. What I am particularly interested in is the use of multivariable, nonlinear spectral fitting to extract the isotopic information from molecular spectra. I like this approach because it does not require the use of isotope-enriched calibration standards. In this method, we fit the experimental spectrum with a base model containing all the theoretical rovibronic line positions calculated from published molecular spectroscopic constants. I am currently developing a fitting model based on robust statistics (that is, fitting that can automatically identify outliers, such as spectral interferences). Currently, I am using a laser to generate the molecular emission, but other excitation sources definitely can also be used. I hope that with a better data-extraction algorithm and possible extension to other sources, the optical isotopic analysis can become a tool for those applications that do not require ultrahigh precision.

My work on ambient-ionization mass spectrometry is mostly on plasma diagnostics and identification of the processes for the generation of the key reagent ions in some of these plasma-based ambient-ionization sources. One specific area that I want to understand is how and through which species the energy flows from helium (the typical plasma gas) to the generation of reagent ions. I hope that a better fundamental understanding can help optimize these ambient-ionization sources and expand their capabilities, and guide the design of the next generation of sources.

What research are you most proud of thus far?

If I am allowed to choose only one, I will pick my work on using matrix effects as a probe for the study of the charge transfer mechanism in ICP for several reasons. Although charge transfer in ICP has been known, it has not been characterized extensively because the required instrumentation and methods are rather complex. With the matrix-effect approach, charge-transfer behavior of 22 elements, in which 12 were entirely new, were reported. What is more interesting and surprising is the observation of state-selective charge transfer in ICP (6). Although state-selective charge transfer is common in low-pressure discharges (such as glow discharge), it had been speculated that because of the high temperature and frequent collision environment in an atmospheric-pressure ICP, state-selective charge transfer should not be anticipated in the ICP. The successful observation of state-selective charge transfer in ICP was partly related to the simplified procedure, and a large pool of emission lines can be simultaneously studied. Matrix interference obviously is an undesirable phenomenon during an analysis, and I am proud of transforming it into a useful tool for another purpose and that valuable information can be obtained from it. 

You have published many articles (~51) and you have given or contributed to 55 presentations at national and international conferences. How do you balance working on new or cutting-edge research and giving lectures and writing papers to share with your peers? How important is it to you to give back to or share your knowledge with your peers?

The balance is not difficult to find, and I am very fortunate to have had incredibly supportive supervisors along my career path; they all allowed me to freely explore new and cutting-edge research and trusted my progress. My supervisors created very enjoyable research environments, and thus, it was easy to keep the enthusiasm for discovering something new. In my view, sharing knowledge with peers is very important; I have learned a lot from my peers through scientific discussions (either formal or casual), and I hope that they have also expanded their knowledge through such interactions with me. In fact, my work on plasma diagnostics on the ambient-ionization source described above started from an informal discussion with my peers in the laboratory.

References

  • T. Pfeifer, R. Janzen, T. Steingrobe, M. Sperling, B. Franze, C. Engelhard, and W. Buscher, Spectrochim. Acta Part B76, 48–55 (2012).

  • G.C.Y. Chan, W.T. Chan, X.L. Mao, and R.E. Russo, Spectrochim. Acta Part B56, 77–92 (2001).

  • G.C.Y. Chan and G.M. Hieftje, Spectrochim. Acta Part B61, 642–659 (2006).

  • G.C.Y. Chan and G.M. Hieftje, Spectrochim. Acta Part B59, 163–183 (2004).

  • G.C.Y. Chan, Spectroscopy30(4), 34–35 (2015).

  • G.C.Y. Chan and G.M. Hieftje, Spectrochim. Acta Part B59, 1007–1020 (2004).