Part 1 of this tutorial addressed how environmental contamination may be minimized by using proper personal analytical practices, sample preparation environmental suggestions, avoidance of glass and low purity quartz contact with samples and standards during sample collection and preparation, and appropriate choices of high purity solvents during sample collection and preparation. Part 2 continues the discussion in terms of minimizing sample contamination issues by using high purity polymer materials for sample collection and preparation, as well as discuss the differences between instrument and method limits of detection (LODs). Accepted and appropriate procedures for calculating method LODs are related to variability in instrument response and unavoidable environmental contamination during even the cleanest sample preparation. Adherence to such practices will help analysts avoid false positives and report results with a high level of confidence.

Investigation of materials for trapping aerosols with minimum metal contamination compared to glass and low purity quartz fiber filters led to investigation of membrane filters manufactured from polymers. Unfortunately, polymer membrane filters (47 mm diameter or 60 mm diameter, 1.0 μm stated pore size) either could not meet ISO standard 20768 requirements for puff profile nor pressure drop across the filters, or they initially met the puff profile specifications, but clogged too quickly to collect adequate aerosol for analysis due to the excessive air flow restriction (1). Acid hardened ashless filter papers (chosen for lower metals contamination) were inadequate because aerosol breakthrough occurred even when three filters were placed together. Also, after a few standard 3 sec 55 mL puffs, the filters became wet and limited proper air flow, introducing deviations from requisite puff profile specifications (1). Although fluoropolymers are of much higher purity than glass, similar problems arose during method development. Thus, we developed a simple aerosol condensation trapping method using simply sufficient lengths of 3.97 mm i.d. high metals purity fluorinated ethylene propylene (FEP) tubing (Figure 1) (2,3). Tubing with smaller i.d. had a greater surface area to length ratio for more efficient condensation, but did not pass puff profile specifications due to resistance to air flow.

Impingers compatible with smoking machines are available from most tobacco product instrument manufacturers. Most of the impingers are made of glass with 25 to 50 mL volume, and used for measuring organic analytes. The simple fact that they are predominantly glass indicates the “experienced inorganic analyst” that these traps are not suitable for analyses of metals in smoke particulate nor aerosols from electronic nicotine delivery systems (ENDS).

Also, the trap volumes are not suitable for trapping metals in smoke particulate nor in ENDS aerosols because impingers should not be filled more than half full of liquid. A fraction of the liquid is lost, as it is bubbled out of the exit tube at standard puff air flow rates if impingers are more than half filled. Even a full 50 mL volume of dilute acid is insufficient to quantitatively trap neither cigarette smoke using commonly accepted smoke regimens nor ENDS aerosols using the ISO Standard 27068 regimen nor more aggressive smoking or aerosol generation regimens that may be representative of human puffing behaviors (1,4,5). Further, aqueous solvents are not appropriate solvents for smoke particulate that contains tar. Much of the tar is insoluble in dilute aqueous acid, and sticks inside the connection tubing and sides of impinger vessels rather than dissolving in the aqueous impinger liquid. Because tar is a heterogeneous mixture including very hydrophobic substances, such as polycyclic aromatic hydrocarbons (PAHs), as well as metals, neither hydrophilic nor hydrophobic solvents are ideal for collecting total particulate matter (TPM). We have used high purity quartz distilled 

-dimethylformamide (DMF) for a few applications because it is a highly polar solvent that dissolves metal ions as well as very hydrophobic substances in a single phase (6). The analyst must then either digest the solvent with dissolved or well-suspended particulate matter or dilute it into aqueous solutions for inductively coupled plasma–mass spectrometry (ICP-MS). Thus, neither tobacco industry labs nor public health labs commonly use impingers for collection of smoke particulate analyses of metals.

Some perfluoroalkoxy alkanes (PFA) impingers (60 mL) are commercially available (Savillex, Figure 2) and are suitable for trapping ENDS aerosols. The high purity PFA can be soaked and rinsed with ultrahigh purity dilute acids before use. One 60 mL impinger containing 30 mL dilute acid, however, is still insufficient for collection of aerosols using 55 mL puff volumes with a 3 sec rectangular puff profile (1). The air flow during this puff is 18.33 mL/sec, slightly greater than the average air flow rate using 35 mL puff volume and a 2 sec puff with a bell-shaped puff profile, as specified for ISO standard 3308, but less than the average air flow rate during a WHO Method 1 55 mL puff in 2 sec (4,5). We have used dimethyl sulfoxide for collection of mainstream cigarette smoke TPM using the ISO Standard 3308 and WHO Method 1 regimens with PFA impingers to provide the U.S. Food and Drug Administration Toxicology Lab dissolved TPM samples for cell culture studies. Three 60 mL impingers connected in tandem containing 30 mL liquid in each impinger were required to quantitatively collect smoke particulate matter from cigarettes using either smoking regimen. It is apparent that the 18.33 mL/sec air flow rate specified in ISO Standard 20768 would require approximately the same volume of liquid in three 60 mL impingers to quantitatively trap aerosol for analysis of metals as was required to trap smoke aerosol using the slightly lower ISO Standard 3308 average air flow rate and the higher WHO Method 1 air flow rate. For these reasons, it is apparent that 25 mL or 50 mL glass impingers were not manufactured with collection of metals in smoke or aerosol samples for analysis of trace metals in mind.

may be defined as the persistence of a given elemental signal after the analysis of a standard or sample and a reasonable introduction system rinse time. Metals such as mercury are well known for memory effects, resulting in long introduction system rinse out times with nitric acid alone, since nitric acid is not a strong enough oxidizing acid to maintain mercury in oxidized form. The solution to this rinse out problem is the addition of an acid (such as hydrochloric acid), or a chelating agent that has a soft base anion ligand (such as sulfur) to the diluent and rinse solutions as previously discussed (7).

Modified polytetrafluoroethylene (TFM) vessel polymers are more rigid polymers than the PFA used to produce earlier generation microwave vessels. TFM vessels may be heated to higher temperatures without deformation, enabling more complete oxidation of organic substances. The TFM molecular polymer chains, like all molecules, have greater rotational motion at higher temperatures. It is possible that the more rapid motion of the PFA and TFM polymer chains at high temperatures permit metals to be captured in the surface polymer matrix. Therefore, PFA and even TFM vessels may superficially absorb metals whether the metals are known to exhibit memory effects when heated during a microwave digestion and may contaminate a subsequent digestion. The use of a hot nitric acid vapor cleaning system between analytical runs helps the problem, but does not eliminate it. Even after using two cycles with a hot acid vapor cleaning system to clean PFA vessels, we have observed trace metals carry over into digestion of a subsequent sample. This first became evident during digestion of urine samples for determination of uranium isotope ratios at low concentrations (8). It was noted that when urine samples containing spikes of reference materials or with known isotope ratios were digested in PFA vessels that previously contained a sample with a different isotope ratio, the resulting digested sample would have an isotope ratio intermediate between the previous sample and the subsequent sample, thus requiring blank digestions between samples to remove the metals absorbed by PFA vessels. We have subsequently observed similar effects with multiple metals from tobacco and other plant material digests when using higher temperatures in TFM microwave digestion vessels than were formerly used with PFA vessels. If a tobacco sample has a high concentration of any given metal, and a sample with a low concentration of the same metal follows in the same vessel, the resulting digestion yields a result that is slightly higher than expected. If the order is reversed, a tobacco sample that has a high concentration of any given metal following a sample that had a low concentration of the same metal in the same vessel will exhibit a less apparent effect. This demonstrates that even more rigid TFM vessels require blank digestions between samples that have significantly different metal concentrations before reutilization to avoid contamination from previous samples.

Memory effects are more troublesome if the metal is easily reduced in the presence of acid. For example, nitric acid catalyzes reduction of mercury as organic substances are oxidized. The reduced volatile elemental mercury is not quantitatively recovered. Some of the mercury is lost as vapor when the vessel is opened, and some is absorbed into the vessel walls, accumulating and causing increasing contamination of samples subsequently digested in the same vessel even after hot acid vapor cleaning. The addition of 1 mL hydrochloric acid to 9–10 mL nitric acid during microwave digestion will help remedy this problem, based on the same Pearson principle used to correct introduction system memory effects (4). Addition of sufficient soft base anion ligand such as chloride from hydrochloric acid chelates the Hg

. This complex ion is soluble in aqueous solvent. Since an equilibrium is formed between Hg

 complex, the equilibrium is shifted away from the Hg

 equilibrium (according to Le Chatelier’s principle), stabilizing mercury in soluble oxidized anionic complex form that does not have a memory effect, and minimizing contamination of subsequent samples in the same vessel.

The same considerations apply if a heating block is used. Figure 3 shows a heating block with a Teflon-coated surface for avoiding contamination and protecting the metal surface beneath. Note polypropylene tubes come from two sources, and PFA tubes come from another source. All need to be precleaned by heating the tubes with dilute nitric and hydrochloric acid, since heating causes the release of metals that were trapped in the polymer matrix. Note the discoloration of caps on the tubes on the left, demonstrating an attack of concentrated acids even on polymers.

Instrument manufacturer instrument detection limits (IDLs) are useful only to compare signal to background ratio with other instruments. They are typically calculated as three times the standard deviation of a single analysis of a dilute ultrahigh purity acid blank that is run on a well-optimized instrument that has never been contaminated with samples with more complex matrices. Some analysts have published data with what appears to be a list of instrument manufacturer IDLs in supplementary files rather than true method detection limits (LOD), or with no description of determination of an LOD (9). An IDL is not relevant for reporting analytical data obtained using a validated method. Method LODs take into account instrument response determined under identical analytical method conditions developed for the analytes and matrices used. A real sample matrix is likely more complex than diluted ultrahigh purity nitric acid, and is measured with instrumentation that is no longer not pristine. Method LODs take into account trace metal contamination from microwave digestion vessels (if used), devices and vessels used for sample dilution, imperfectly prepared calibration standards and internal standards, minor pipetting inaccuracies, differences in instrument response from day to day, and all other aspects of sample preparation and instrument variability and error over a multiday period. The ideal method LOD would be useful to statistically determine the uncertainty of analytical method results at low concentration, and establish a concentration below which it is recognized that analytical results are insufficiently certain to report. True analytical results can then be distinguished from false positives resulting from trace metal contaminants of samples from previous vessel and instrument utilization, sample preparation and analysis, analytical interferences, and day-to-day differences in instrument response as optimization parameters change. It is apparent that using good analytical practices described in Part 1, as well as maintaining vessels used for sample collection and preparation as free of metal contamination as possible, will result in lower method LODs.

Although polypropylene bottles and 15 or 50 mL “centrifuge tubes” sold as sample containers generally have very low extractable metal concentrations, it is a good analytical practice to verify this. We tested a new batch of polypropylene tubes from a source for which we had formerly observed very low extractable trace metals, and observed significantly higher extractable aluminum in the new batch. We inquired about this, and learned that the vendor had changed the manufacturer to save on costs. Since aluminum was one of our hydrophobic aerosol method analytes, we had to change the source of sample tubes and check tubes from different vendors to determine the best source for the method. This illustrates that it is a good practice to screen a few containers every time a new batch of sample containers is received to avoid false positives from samples, even when using relatively high metal purity polypropylene tubes.

To minimize contamination of samples from containers as described above, it is a good practice to fill the tubes with 1% v/v ultrahigh purity nitric acid + 1% v/v ultrahigh purity hydrochloric acid, cap them, invert them a few times, and let them sit overnight before using them. This procedure cleans out most sample vessel impurities, unless they will be heated. Since the method LOD should incorporate variability in trace element background from the rinsed containers that we use, any trace metals that may remain have already been taken into consideration. Although acid rinsed sample containers result in contamination well below our method LODs, the following will illustrate how a well-validated method LOD prevents reporting false positives due to contamination. Polypropylene tubes (15 mL) obtained from three sources were prepared in triplicate with acid rinse solution, capped, inverted three times, and soaked overnight. Rather than discarding the overnight rinses, the rinses were analyzed according to our ENDS aerosol method (3).

It is evident from the results in Table I that nearly all of the polypropylene tube rinses were below the method LODs even before the tube prerinse. These tube sources had very low extractable elemental contamination. Two of the three nickel results in the tubes from our second source were below the method LOD. The nickel concentration obtained from only one tube was shown in the table to demonstrate that contamination from one tube in a batch of otherwise clean tubes may cause a false positive result. Since metals from only 3 of 100 tubes from each manufacturer were checked, it is possible that other tubes in the batch may also have low levels of contamination. It is also possible that tubes from the other sources had contaminated tubes that were not randomly selected. This confirms that prerinsing sample containers as a precaution is a good inorganic analytical practice to prevent contamination and false positive results.

There are several widely accepted ways to calculate method LODs. The simplest way is to include a sample in each analytical run for validation that has the same matrix as samples that will be analyzed with the method, and run it through the same sample preparation procedures as the samples. An appropriate sample should have very low analyte concentrations though remain in the same matrix as samples, in order to represent the higher relative variability that is analytically observed for low concentrations than high concentrations in the sample matrix. The same sample matrix is prepared independently a minimum of 14 times for 14 analytical runs per formed on 14 or more days. The standard deviation of the results obtained over a minimum of 14 days is multiplied by 3 to obtain approximately 99% certainty that the resulting concentration is above a level of response that would cause result uncertainty (9). This method may be taken a step further by analyzing 10 or more replicates of the same or similar sample with low analyte concentrations in the same run. The within run and between run standard deviations are then combined ( [

 ), since long term variability and uncertainty is usually greater than short term standard deviations (9,10).

The most well-known method for calculation of a method LOD is the Taylor method (11). This method requires multiple analytical runs as described in the simpler methods above. A minimum of seven or more samples are spiked with known analyte concentrations, analyzed according to the method procedure on separate days, and the standard deviations of the analytical results are plotted on the ordinate (

axis) versus concentrations on the abscissa (

 axis). The regression line through the plotted data is extrapolated to the ordinate at “0” analyte (abscissa) concentration. The apparent standard deviation at “0” analyte concentration is multiplied by 3 to determine the method LOD. This technique is a very good approximation, although it was likely intended for analyses in which there is negligible analyte concentration in the matrix, whereas samples are unlikely to have “0” metal concentration. When used for calculation of a method LOD with external standards, it is possible that the standards will be in a cleaner matrix than the samples, so a quality control sample (QCS) or certified reference material (CRM) that has characterized analyte concentrations and the same or very similar matrix should be included to add matrix effect to determination of the method LOD (12).

An additional level of care requires adding a statistical adjustment to account for potential overlap between false positives from samples with false negatives. This method requires determination of the regression using the Taylor method and inclusion of a sample blank along with the spiked samples, but statistically adjusting the results using the regression slope and ordinate intercept (12,13). Procedural blank and standard deviation are determined with analyses from multiple analytical runs, 

is the slope of the regression determined with the Taylor method, and 

is the intercept obtained using the Taylor method, displayed in equation 1 (11–13).

By properly calculating method LODs with an appropriate method, the analyst considers not only day-to-day variability of instrument response, calibration, and sample preparation, but trace contamination in digestion vessels (if used) and sample containers and other ancillary materials that contact samples as well. In this way, the analyst avoids reporting false positive and false negative results. The analyst determines “…the smallest measure, 

, that can be detected with reasonable certainty for a given analytical procedure” (9). In other words, “A common misconception is that the LOD is the smallest concentration that can be measured [with an instrument]. Instead, it is the concentration at which we can decide whether an element is present or not…” (9). A method LOD does not tell us the lowest possible ion concentration that produces ion counts in a secondary electron multiplier detector; rather, it reveals the concentration at which we can distinguish an elemental concentration in a sample from artifacts of variability from instrumentation response, calibration, and sample preparation procedures.

Trace metal contamination is unavoidable during preparation of samples for inorganic analyses. In order to minimize sample contamination when using these techniques, the analyst must take care to learn appropriate sample preparation practices and wisely choose ancillary vessels made from appropriate materials that contribute as little trace metal contamination as possible.

Method LODs should be calculated to take into account not only day-to-day instrument response variability, but variability of trace contamination acquired during independent sample preparation procedures as well. This will aid in avoiding reporting false positive and false negative results.

The findings and conclusions in this report are those of the authors and do not necessarily represent the views of the Centers for Disease Control and Prevention.

(1) International Standard Organization. Vapour Products — Routine Analytical Vaping Machine — Definitions and Standard Conditions. ISO Standard 20768:2018 Geneva, Switzerland. 

(2) Gray, N.; Halstead, M.; Valentin-Blasini, L.; Watson, C., Pappas, R.S. Toxic Metals in Liquid and Aerosol from Pod-Type Electronic Cigarettes. 

(3) Halstead, M.; Gray, N.; Gonzalez-Jimenez, N.; Fresquez, M.; Valentin-Blasini, L..; Watson, C.; Pappas, R.S. Analysis of Toxic Metals in Electronic Cigarette Aerosols Using a Novel Trap Design. 

(4) International Standard Organization. Routine Analytical Cigarette-Smoking Machine — Definitions and Standard Conditions. ISO Standard 3308:2012 Geneva, Switzerland, 

https://www.en-standard.eu/iso-3308-routine-analytical-cigarette-smoking-machine-definitions-and-standard-conditions/#:~:text=ISO%203308%20ISO%203308%3A2012%3A%20defines%20smoking%20parameters%20and,analytical%20smoking%20machine%20complying%20with%20the%20standard%20conditions

(5) World Health Organization. WHO TobLabNet Official Method SOP 01. Standard Operating Procedure for Intense Smoking of Cigarettes. World Health Organization 2012 Brussels, Belgium. 

https://www.who.int/publications/i/item/standard-operating-procedure-for-intense-smoking-of-cigarettes

(6) Fresquez, M.R.; Gonzalez-Jimenez, N.; Gray, N.; Valentin-Blasini, L.; Watson, C.H., Pappas, R.S. Electrothermal Vaporization-QQQ-ICP-MS for Determination of Chromium in Mainstream Cigarette Smoke Particulate. 

(7) Pappas, R.S. Sample Preparation Problem Solving for Inductively Coupled Plasma-Mass Spectrometry with Liquid Introduction Systems I. Solubility, Chelation, and Memory Effects. 

(8) Pappas, R.S.; Ting, B.G.; Paschal, D.C. A Practical Approach to Determination of Low Concentration Uranium Isotope Ratios in Small Volumes of Urine. 

(9) Thomsen, V.; Schatzlein, D.; Mercuro, D. Limits of Detection in Spectroscopy. 

(10) Pappas, R.S.; Ting, B.G.; Jarrett, J.M.; Paschal, D.C.; Caudill, S.P.; Miller, D.T. Determination of Uranium-235, Uranium-238, and Thorium-232 in Urine by Magnetic Sector Inductively Coupled Plasma-Mass Spectrometry. 

Quality Assurance of Chemical Measurements

(12) Gonzalez-Jimenez, N.; Gray, N.; Pappas, R.S,; Halstead, M.; Lewis, E.; Valentin-Blasini, L.; Watson, C.; Blount, B. Analysis of Toxic Metals in Aerosols from Devices Associated with Electronic Cigarette, or Vaping, Product Use Associated Lung Injury. 

(13) National Committee for Clinical Laboratory Standards (NCCLS). Protocols for Determination of Limits of Detection and Limits of Quantitation. National Committee for Clinical Laboratory Standards 2004 Wayne, PA, 24, 9–35.

Articles

Columns: Atomic Perspectives

Minimizing sample preparation issues requires users to consider method LODs as well as high purity polymer materials for sample collection and preparation. Here, we examine proper personal analytical practices one can take to avoid environmental contamination.

Contamination Control During Sample Preparation for Trace Element Analysis of Electronic Cigarette Aerosol with Inductively Coupled Plasma-Mass Spectrometry, Part 2

 green phosphors were synthesized via hydrothermal method followed by a post-annealing process. The variation regularity of phase structure, morphology, particle size and concentration quenching mechanism were investigated. The results showed that the crystal structure of matrix materials was not affected when Tb

 ions were introduced. In addition, the micrometer-sized rods and excellent dispersion of the particles were obtained by adding urea. Under excitation at 377 nm, the phosphors exhibited the characteristic of Tb

 emissions with the peaks located at 488 nm (

), respectively. Moreover, based on optical analysis, it can be inferred that electric dipole-dipole (d-d) interaction plays an important role in concentration quenching process of Tb

 ions. The green light emitting phosphor MgAl

 can be realized by adjusting the concentrations of Tb

 was found to be 0.02. Finally, the average fluorescence decay lifetime and the CIE chromaticity coordinates of Mg

 phosphor were calculated to be 1.2005 ms and (0.3197 0.4109), respectively. The above results suggest that the developed phosphor could have potential to be used as a single-phase green emission phosphor for near-ultraviolet light-emitting diode (NUV-LED) applications.

Aluminate-based luminescent materials activated by rare earth ions are of interest due to their good water resistance, stable chemical properties, high efficiency, and lower cost (1–3). As the use of phosphors becomes more widely spread, the scope of aluminate applications includes light-storing luminescent pigments and other products, like LED phosphors (4–6). But with many similar products in the market at present, the best method for aluminate phosphors to remain competitive will be for their raw materials to be adjusted in the areas of processes, environmental protection, and applicability.

With the gradual decrease in mineral resources in the world, more and more attention has been paid to the green chemistry concept while considering the economic cost. Spinel (MgAl

) is an important branch of aluminate that can be widely used in ceramics (7), catalyst support (8), refractories (9), and other fields because of its unique structure. MgAl

 has obvious superiority to traditional oxide especially for the host material, which not only inherits the stability of the oxide, but also has lower phonon energy (about 670 cm

). A high-temperature, solid-state method is traditional to synthesize MgAl

 spinel, having the advantages of technical maturity, short duration, and simple operation, but there are some disadvantages, for example large particle size, poor dispersion, and easy agglomeration. Additionally, since it is difficult to mix trace additives evenly by ball milling, this method is not advisable for low concentrations. Moreover, carefully controlling the product morphology, size, and uniformity based on the advantages of the high-temperature solid-state method has become an urgent problem for researchers. To this end, scholars from various countries are further researching the synthesis and application of MgAl

. Mostafa Y. Nassar and coworkers described the preparation of MgAl

 particles by the sol-gel assisted solid phase method with oxalic acid as an additive (10). Guotian Ye and associates (11) prepared spherical MgAl

 particles via thermal decomposition method while an annealing process at 900–1000 

C for 5 h was needed to obtain the final product. Lu Yang and others (12) reported on a low-temperature hydrothermal method combined with a calcination process to prepare MgAl

 phosphors, and obtained micro-rod products through the induction of cationic surfactant CTAB. This research indicates that under mild conditions, liquid-phase reaction and solvent, pH, surfactant, and other controllable operations can help control the shape and size of the realized product powder. On this basis, the solid-phase reaction can further improve the purity of the target product.

Herein, we present a controlled synthesis of Tb

 microstructures with several well-defined morphologies via a urea-based hydrothermal method followed by a heat treatment process. The synthesis route developed in this work is environmentally benign without adding organic solvents except a small amount of urea. The size and morphology of the urea has been investigated in detail. In addition, the emission properties of the prepared MgAl

 green phosphors were prepared by the hydrothermal method followed by a post-annealing treatment, in which the starting chemicals included terbium nitrate pentahydrate [Tb(NO

O, 99.99%], magnesium nitrate hexahydrate [Mg(NO

O, ≥99%], aluminium nitrate nonahydrate [Al(NO

O were accurately weighed and dissolved in deionized water to prepare solutions. Then, according to Mg

 (x = 0.005, 0.010, 0.015, 0.020, and 0.030), the mixture containing Tb(NO

 solution drop by drop. After the mixture was well mixed, the urea was added and the pH was adjusted to about 11 with ammonia solution, and the stirring continued for more than 15 min. Subsequently, the mixed solution was transferred into a hydrothermal reactor at 220 

C for 6 h. After cooling the autoclave to room temperature, the precipitate was separated by centrifugation and washed with deionized water. Finally, MgAl

 phosphor was obtained by calcining the precursor at 1200 

The X-ray diffractometer (Empyrean X, Panalytical B.V.) was used to measure the phase structure of the samples. The micro-morphology of the phosphors was observed using a field emission scanning electron microscope (Supra 55). Meanwhile, the excitation and emission spectra and fluorescent decays were recorded by a fluorescence spectrometer (Fluoromax-4, HORIBA) equipped with a 150 W xenon lamp. For comparison, all excitation and emission spectra were measured at room temperature with the same instrumental parameters.

Figure 1a displays the XRD patterns of Tb

 phosphor for different concentrations of Tb

 = 0.005, 0.010, 0.015, 0.020, and 0.030). As shown in the figure, all the samples were readily indexed as cubic phase and identified as MgAl

, according to JCPDS Card No. 21-1152. The diffraction peaks were strong and sharp without any second phases traced, which implies that all of the products were of pure MgAl

 structure. Also verified was that the amount of Tb

To further confirm that the samples possessed cubic structure, a structural refinement by the Rietveld analysis was performed using the GSAS soft- ware presented in Figure 1b. As the refinement converged to 

wp(%) = 12.39, the results revealed that Tb

 samples had single-crystal cubic structure in space group Fd-3m. A schematic illustration of the crystal structure of MgAl

 is shown in the inset of Figure 1b. Here, O

 is arranged in the form of cubic close-packed, Mg

 is located in the tetrahedral gap composed of four oxygen ions, and Al

 is located in the octahedral gap composed of six oxygen ions. When Tb

 ions incorporated in either of two lattice positions, one occupied the tetrahedral gap (Mg

). Generally speaking, the substitution position of rare earth ions is related to the radius between doped ions and substituted ions, as well as charge. Based on ionic radius matching, the Tb

An SEM study was undertaken to investigate the surface morphology of Mg

 phosphors with and without urea, as clearly displayed in Figures 2a and 2b respectively. As presented in Figure 2a, most of the products are irregular in shape, and these particles are composed of aggregated flakes. Obviously, the shape and size of the sample greatly changed with the addition of urea, and a micrometer-sized, rod-like shape with uniform morphology and excellent dispersion appeared, as shown in Figure 2b. It can be clearly seen that the length of the rod is approximately between 3 μm and 4 μm. In addition, it is obvious that with excellent dispersion and almost no aggregation, the distribution of particle size was homogeneous.

Obviously, urea played an important role in the tuning of the morphology and dispersion, whereas the rate of the crystal growth and nucleation might be accelerated by the generation of hydroxide ions supplied by the decomposition of urea. When the same amount of Al

 ions were deposited into hydrotalcite, the surplus Al

 ions formed an ammonium aluminate carbonate with NH

 (equation 1) to avoid the formation of boehmite, which is easy to agglomerate under hydrothermal conditions. Therefore, the sample added with urea had better dispersion (13).

Figure 3a shows the excitation spectra of Mg

 phosphor recorded by monitoring emission wavelength at 543 nm. From 300 to 400 nm, there are a series of excitation peaks that can be observed, with excitation bands centered at 317, 339, 350, 367, and 377 nm originating from the 

, respectively. Based on these results, obvious absorption peaks appear at 377 nm, indicating that such series of phosphors can be effectively excited by NUV–LED.

 ions cover the visible spectral region, and the relative intensity of the emission lines of Tb

. Under 377 nm near-ultraviolet light excitation, the most intense emission concentrated at 544 nm was attributed to 

 transitions, respectively. Moreover, weak emission peaks were detected at 514, 619, and 686 nm. By contrasting the energy structure and spectrum structure of Tb

, it was found that none of these emission peaks belonged to characteristic transitions of Tb

 ions. Therefore, in order to determine the source of the above emission peaks, the emission spectra of undoped Tb

 ions were measured under the same conditions, as shown in Figure 3c. Obviously, the above emission peaks originated from the matrix. The emission intensity increased with Tb

 concentration, while it reached the maximum at 

 = 0.02 and was found to decrease with further increase in the Tb

 content due to a concentration quenching effect, originating from the greater probability of energy loss among the adjacent Tb

In order to explain the luminescence mechanism, Figure 4 presents the energy level scheme and the processes of photoluminescence excitation and emission of the Tb

 phosphors were irradiated with 377 nm NUV light, the Tb

 ions were excited from ground state to excitation state 

. The higher energy or excited quantum mechanical states quickly relaxed down to the lower energy excited state (

) via non-radiative relaxation. Subsequently, the radiative relaxation occurred between 

 ion, resulting in the characteristics emission peaks at 488, 544, and 586 nm, respectively.

Decay Spectral and Concentration Quenching Analysis 

Generally, the decay kinetics behavior of a phosphor relates to the luminescent centers, amount of defects, energy transfer, and impurities in the matrix material (15). The fluorescence decay curve is measured with excitation wavelength at 377 nm and emission wavelength at 544 nm, as depicted in Figure 5. These curves, which can fit via a double-exponential function, are defined below (16):

 is the luminescence intensities at time 

 stand for the luminescence lifetimes for exponential composition. As a function of these parameters, average luminescence lifetime (

According to equations 2 and 3, the calculation results of average fluorescence lifetime are displayed in Table I. Obviously, the average lifetime of the 

As an important parameter, the critical distance (

) related to the concentration quenching effect (17). Generally, the energy transfer process includes two interaction ways: one is exchange interaction, the other is electric multipole interaction. According to Blasse, if exchange interaction occurs, the critical distance between sensitized ions and activated ions should be less than 4 Å (18). As a function of the Blasse formula (19), the critical distance 

 is the critical doping concentration of Tb

 at the point that the quenching occurs, 

 refers to the number of formula units per unit cell. 

 = 8; the critical distance (Rc) of adjacent terbium ions was found to be 18.40 Å. Obviously, the energy transfer in the present case occurred through electric multipolar interaction.

In order to further study the concentration quenching mechanism of Tb

, the integral area of the emission peaks were calculated according to the Figure 3b spectrum data. The dependence of integrated emission intensities of 

 concentration are depicted in Figure 6. Using Van Uitert’s theoretical model, in which the dependence of the doping concentration of the luminescence center on luminescence intensity can be expressed by the following equation (20):

 is the concentration of the luminescence center, 

 are parameters related to the research system and measurement system parameters, and 

 is the interaction mechanism between luminescence centers where 

 = 6, 8, and 10 stand for the dipole-dipole (d-d) interaction, the dipole-quadrupole (d-q) interaction, and the electric quadrupole-quadrupole (q-q) interaction, respectively. Based on these, the concentration quenching experimental data of Tb

 are nonlinear fitted according to equation 5. A good agreement with the fit curve for the whole range of concentrations was found (see Figure 6), and the Q value was obtained from the fitting process to be 5.988±0.03922, which indicated that the fluorescence quenching behavior of 

Figure 7 exhibits the CIE chromaticity diagram for the emission spectra of MgAl

 doping concentration. The calculated values of color coordinate are presented in Table I. It can be found that the measured CIE coordinates of all the samples are located in the green area. Moreover, the corresponding luminescent photographs are presented in the inset of Figure 7, in which a dazzling green emission can be clearly observed in the Tb

 sample with naked eyes under the 377 nm laser radiation. The correlated color temperature (CCT) as one of the characteristics of phosphor has also been calculated by using the approximate formula reported by McCamy (21,22):

 = (X - 0.3320)/(Y - 0.1858). The results are also shown in Table I.

With water as a solvent, a series of MgAl

 phosphors were synthesized via a hydrothermal assisted solid-phase method adjusting the concentration of Tb

. The X-ray diffraction analysis revealed that the phosphors demonstrated cubic phase microcrystals; meanwhile, SEM results confirmed that particle size, morphology and dispersion of the product could be effectively controlled by adding urea. The phosphor emitted green luminescence centered at 586, 544, and 488 nm because of the 

 transition at an excitation wavelength of 377 nm. Furthermore, the emission intensity increased firstly and then decreased with the increasing of Tb

 concentration and reached a maximum value at 

, at which concentration quenching was attributed to the electric d-d interaction mechanism, and the critical energy transfer distance was estimated to be 18.40 Å.

This work was supported by the National Key Research and Development Program of China (2017YFB0310300), the Natural Science Foundation of Liaoning Province (2019-DZ-0090), and Petrochemical Research Institute, PetroChina (PRIKY21001).

(1) Zhu, Y.; Zeng, J.; Li, W.; Xu, L.; Guan, Q.; Liu, Y. Encapsulation of Strontium Aluminate Phosphors to Enhance Water Resistance and Luminescence. 

(2) Bartram, R. H.; Hamilton, D. S.; Kappers, L. A.; Lempicki, A. Electron Traps and Transfer Efficiency of Cerium-Doped Aluminate Scintillators. 

(3) Wako, A. H.; Dejene, F. B.; Swart, H. C. Combustion Synthesis, Characterization and Luminescence Properties of Barium Aluminate Phosphor. 

(4) Abdelhameed, M. M.; Attia, Y. A.; Abdelrahman, M. S.; Khattab, T. A. Photochromic and Fluorescent Ink Using Photoluminescent Strontium Aluminate Pigment and Screen Printing towards Anticounterfeiting Documents. 

(5) Lü, X.; Sun, M.; Zhang, J.; Wang, T. Effect of Mixing Process on the Luminescent Properties of SrAl

(6) Gao, Z. W.; Sun, P. F.; Zhong, Y. F.; Yu, R.; Deng, B. Eu

-Doped Highly Thermal-Stable Barium Yttrium Aluminate as a Red-Emitting Phosphor for UV Based White LED 

(7) Osipov, V. V.; Shitov, V. A.; Maksimov, R. N.; Lukyashin, K. E.; Solomonov, V. I.; Ishchenko, A.V. Fabrication and Characterization of IR-Transparent Fe

(8) Cho, E.; Yu, Y. J.; Kim, Y.; Phan, T. N.; Park, D.; Ko, C. H. Egg-Shell-Type Ni Supported on MgAl

 Pellets as Catalyst for Steam Methane Reforming: Enhanced Coke-Resistance and Pellet Stability 

(9) Gu, Q.; Liu, G. Q.; Li, H. X.; Jia, Q.; Zhao, F.; Liu, X. Synthesis of MgO–MgAl

 Refractory Aggregates for Application in MgO–C Slide Plate. 

(10) Nassar, M. Y.; Ahmed, I. S.; Samir, I. A Novel Synthetic Route for Magnesium Aluminate (MgAl

) Nanoparticles Using Sol–Gel Auto Combustion Method and Their Photocatalytic Properties. 

(11) Ye, G. T.; Oprea, G.; Troczynski, T. Synthesis of MgAl

 Spinel Powder by Combination of Sol–Gel and Precipitation Processes. 

(12) Lu, Y.; Wang, J.; Shi, Z. X.; Shi, J.; Liu, Y.; He, L. Site Occupation and Fluorescence Properties of MgAl

(13) Lu, Y.; Wang, J.; Shi, Z. X.; Dai, L. J.; Zhang, H. M.; Sun, Z. T.; Liu, Y. F.; Preparation and Upconversion Luminescence Properties of MgAl

(14) Dieke, G. H.; Crosswhite, H. M. The Spectra of the Doubly and Triply Ionized Rare Earths. 

(15) Yawalkar, M. M.; Zade, G. D.; Singh, V.; Dhoble, S. J. Investigation of Luminescence Processes in Li

(16) Zhang, X. J.; Wang, J.; Huang, L.; Pan, F.; Chen, Y.; Lei, B.; Peng, M.; Wu, M. Tunable Luminescent Properties and Concentration-Dependent, Site-Preferable Distribution of Eu

 Ions in Silicate Glass for White LEDs Applications. 

(17) Ren, Y. D.; Liu, Y. H.; Tan, S. M.; Cui, H. Y.; Wang, Y. L.; Li, X. M.; Yang, R.; Wei, X.; Zhang, H. W.; Sun, Y. D. Energy Transfer Rate and Electron–Phonon Coupling Properties in Eu

(18) Dexter, D. L.; Schulman, J. H. Theory of Concentration Quenching in Inorganic Phosphors. 

(19) Zhu, X. M.; Zhou, Z. F. Photoluminescence and Energy Transfer Mechanism of a Novel Tunable Color Phosphor Na

(20) Van Uitert, L. G. Characterization of Energy Transfer Interactions Between Rare Earth Ions. 

(21) Li, R. Q.; Yu, G. H.; Liang, Y. M.; Zhang, N.; Liu, Y.; Gan, S. Morphology-Controllable Synthesis of LaOF:Ln

 (Ln=Eu, Tb) ZCrystals with Multicolor Luminescence Properties. 

(22) McCamy, C. S. Correlated Color Temperature as an Explicit Function of Chromaticity Coordinates. 

 are with the Liaoning Key Laboratory for Fabrication and Application of Superfine Inorganic Powders, at Dalian Jiaotong University, in Dalian, China. 

 is also with the Department of Materials Science and Engineering at Yingkou Institute of Technology, in Yingkou, China. Direct correspondence to: 

Peer-Reviewed Research

A look at how the spectral properties single-phase green emission phosphor make it suitable for near-UV light-emitting diode (NUV-LED) applications.

Preparation and Spectral Properties of MgAl2O4: Tb3+ Phosphor

We continue our survey of the spectra of nitrogen containing polymers with a look at polyurethanes. Polyurethanes are widely made into foam rubber, elastic, and wood coatings, amongst other things. The spectra of polyurethanes are similar to polyamides since both polymer types contain C=O and N-H bonds. Therefore, we will begin with a review of the spectra of polyamides.

The structure and infrared spectra of polyurethanes and polymeric amides or polyamides are similar, hence a review of polyamides is in order. We studied polyamides in a couple of previous columns (1,2). We saw that most polyamides have a secondary amide group in their backbone. The molecular structure of a secondary amide is seen in Figure 1.

Recall that the primary, secondary, and tertiary nomenclature for amides counts the number of C-N bonds in the functional group. Also, as the number of C-N bonds goes up the number N-H bonds goes down (1). As you can see in Figure 1, a secondary amide has two C-N bonds and one N-H bond. Table I shows the group wavenumbers for secondary amides (1).

Since secondary amides have only one N-H bond they have only one N-H stretching vibration that falls from 3370 to 3170 cm

 (going forward, assume all peak positions are in cm

 units even if not so noted). The N-H in-plane bend for secondary amides is found from 1570 to 1515 and is unusually strong for an in-plane bending vibration peak (1). The N-H out-of-plane bend (wag) for secondary amides falls in the range 750 to 680. Recall that N-H stretching and bending peaks are broadened by hydrogen bonding, but are medium in width and intensity compared to O-H stretching and bending peaks because hydrogen bonds formed by N-H groups are weaker than hydrogen bonds formed by O-H groups (1). The C=O bond in all amides conjugates with the lone pair of electrons on the amide nitrogen, so all amides have their C=O stretching peak in the same region from 1680 to 1630 (1).

The structural framework of the secondary urethane functional group is seen in Figure 2.

Like amides, urethanes are characterized as primary, secondary, and tertiary, depending upon whether there are one, two, or three C-N bonds. Additionally, the number of N-H bonds in urethanes decrease from two, to one, to none as we go from primary to tertiary structures, which is, again, similar to amides.

Like polyamides, most polyurethanes contain secondary urethane groups in their backbones. Note in Figure 2 that the left hand side of the secondary urethane functional group contains an ester linkage circled in red. As we have seen, ester groups have C=O and C-O stretching vibrations. The right hand side of the secondary urethane functional group, denoted by the blue circle in Figure 2, contains an amide linkage (3). Note however that both the ester and amide linkages share the same carbonyl group. Thus, we might expect the spectrum of urethanes to be a combination of the spectra of an ester and an amide. This turns out to be the case for the most part, as seen in the spectrum of a polyurethane-polyether foam presented in Figure 3.

The single N-H stretch in Figure 3 is found at 3326 and is labeled A. For secondary urethanes, this peak generally falls from 3340 to 3250. Note the medium peak width typical of other N-H stretches we have studied (1,2). Both esters and amides have C=O stretching peaks, so where will the C=O stretch of a urethane fall? In Figure 3 it is at 1705 and is labeled B. The carbonyl stretch in Figure 3 labeled B is split, with peaks at 1720 and 1705. The cause of this split carbonyl peak is that some urethane C=O groups engage in hydrogen bonding, whereas others don’t. The 1720 peak is from non-hydrogen bonded or “unassociated” carbonyls, whereas the 1705 peak is from the hydrogen bonded or “associated” carbonyl groups (4). This pair of secondary urethane C=O stretching peaks is usually found from 1725 to 1705. Note from Table I that amide C=O stretches fall from 1680 to 1630. Amides and urethanes do have similar spectra, but for amides the C=O stretch is 1680 or lower, whereas for urethanes it is 1705 or higher. Thus, the position of the C=O stretching peak can be used to distinguish a polyamide from a polyurethane.

Recall that, for polyamides, there is an unusually intense N-H in-plane bending vibration (1,2). Urethanes also exhibit this peak. In Figure 2, it is found at 1531 and is labeled C. This peak is usually found from 1540–1520, overlapping the range for secondary amides. The main structural difference between amides and urethanes is that urethanes contain a C-O bond, whereas amides do not. We have studied the spectroscopy of functional groups containing C-O bonds extensively, and found that the spectral signature of a C-O bond is a big peak between 1300 and 1000 (3,5). Urethanes do not disappoint here, as the C-O stretch of the polyurethane in Figure 3 is labeled D, and falls at 1222. This peak is normally seen in urethanes from 1265 to 1200 (4). The big peak near 1000 in Figure 3 is from the ether linkages in this polyurethane/polyether foam rubber material. Table II summarizes the group wavenumbers for urethanes.

Polyurethanes are made by reacting a diol and a diisocyanate, as seen in Figure 4 (6).

is simply a molecule with two alcohol groups in it. Their infrared spectra are similar to those of mono-alcohols that we have studied previously (5).

 is a molecule with one -N=C=O linkage in it, while a 

 contains two of these groups. The synthesis seen in Figure 4 is called a 

 because the driving force is the splitting off of a water molecule. Polyurethane foam, also known as foam rubber, is made by copolymerizing urethanes with other functional groups such as ethers. The infrared spectrum of such a material is seen in Figure 3.

As mentioned above, diisocyanates contain two -N=C=O functional groups. This linkage is reminiscent of CO

 in that is mostly planar and linear (7). The infrared spectrum of a diisocyanate is shown in Figure 5.

Note that the spectrum is dominated by the peak labeled A at 2277 cm

. This peak is due to the asymmetric stretch of the -N=C=O group, which normally falls from 2280 to 2240, as noted in Table II. This peak is unusually intense and broad because of the large dipole moment of the -N=C=O group. The -N=C=O asymmetric stretching peak is the perfect example of a group wavenumber, because it is intense and shows up in a region where very few other functional groups absorb. Throughout the many installments of this column, you may have noticed that much of our attention is focused on the 3500 to 2800 and 2000 to 400 regions. This is because there are very few useful group wavenumbers between 2800 and 2000 (8).

The structure and infrared spectra of polyurethanes and polymeric amides are similar, but there are several key differences worth calling attention to.

July 2023

Contamination Control During Sample Preparation for Trace Element Analysis of Electronic Cigarette Aerosol with Inductively Coupled Plasma-Mass Spectrometry, Part 2

Preparation and Spectral Properties of MgAl2O4: Tb3+ Phosphor

Infrared Spectroscopy of Polymers XIII: Polyurethanes

Technological Analysis of Glazed Tiles Unearthed from Bao’ensi in Nanjing

Differentiating Glycans by Surface-Enhanced Raman Spectroscopy