- Email: [email protected]
Reproducibility of high-performance thin-layer chromatography (HPTLC) in textile dye analysis
Accepted Manuscript Reproducibility of High-Performance Thin-Layer Chromatography (HPTLC) in Textile Dye Analysis Ethan Groves, Skip Palenik, Christopher S. Palenik PII: DOI: Reference:
S2468-1709(17)30135-2 https://doi.org/10.1016/j.forc.2018.03.004 FORC 98
To appear in:
Forensic Chemistry
Received Date: Revised Date: Accepted Date:
31 October 2017 6 March 2018 10 March 2018
Please cite this article as: E. Groves, S. Palenik, C.S. Palenik, Reproducibility of High-Performance Thin-Layer Chromatography (HPTLC) in Textile Dye Analysis, Forensic Chemistry (2018), doi: https://doi.org/10.1016/j.forc. 2018.03.004
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Reproducibility of High-Performance Thin-Layer Chromatography (HPTLC) in Textile Dye Analysis
Ethan Groves, B.S., Skip Palenik, B.S., and Christopher S. Palenik, Ph.D. Microtrace LLC, 790 Fletcher Dr., Suite 106, Elgin, IL 60123
Abstract In the forensic comparison of dyed fibers, thin-layer chromatography (TLC) is one of the few analytical techniques that have proven sensitive enough to detect and separate the minute quantities of dyestuffs present in single fibers. The method has become well-established by trace evidence examiners as a means of further distinguishing fibers whose colors appear to be the same by both comparison microscopy and microspectrophotometry. As practiced at present, the forensic analysis of dyes is limited to a comparison of the separated dye bands from known and questioned fibers performed on the same TLC plate. It is recognized, however, that retardation factor (Rf) alone is not sufficient proof of identify. The limited use of investigative TLC analysis in forensic fiber examinations is due in part to the range of uncontrolled or poorly defined variables that affect the reproducibility of the developed plate. Through the study of a six component test dye mixture developed on over 50 high-performance thin-layer chromatography plates, the effects of several critical variables that affect reproducibility and resolution, including: plate selection, pre-elution, tank saturation, developing distance, and eluent stability have been evaluated. When considered collectively, the results of this research provide a means for acquiring and archiving repeatable data, both from casework and known reference samples collected on different plates and at different times. These provide a pathway for the development and utilization of reference databases for the identification of dyes. The empirical uncertainty established for the generalized separation procedure used in this research provides objective guidance for evaluating the significance of associations (or eliminations) made on the basis Rf. Ultimately, this research also opens a pathway to the use of forensic dye analysis as an investigative tool, rather than one exclusively restricted to comparative analyses.
Keywords Forensic science, dye identification, textile dyes, high performance thin-layer chromatography (HPTLC), reproducibility.
1. Introduction Fiber comparisons represent a significant proportion of all forensic trace evidence analyses [1]. Color is perhaps the most important optical property exhibited by evidential fibers, since along with thickness and shape these are the easiest characteristics to recognize when searching for target fibers through a stereomicroscope. In addition to polymer identification by polarized light microscopy (PLM) and Fourier-transform infrared microspectroscopy (micro-FTIR), the principal property relied upon in most fiber comparison is color, which is evaluated most commonly by comparison microscopy and microspectrophotometry (MSP). It may come as a surprise to some that the colors of most fibers are only rarely, if ever, due to absorptions from a single dyestuff. Since colors are commonly matched to shades specified by their customers, most dye houses are experts at color matching. They characterize these colors using the same spectrophotometric techniques as those employed by forensic scientists. From a spectrum obtained from an exemplar fabric, a dye house will formulate a dye mixture, which can change over time due to changes in dye lots, dyestuff availability (as a function of price and/or quality), and a master dyer’s preferences for specific dyestuffs. One consequence of this method of coloring fibers, by accurately matching a spectrophotometrically specified color produced from different dyes, is an additional level of discrimination in forensic fiber comparisons that is achieved by dye identification. Planar chromatography is a cost effective, routine analytical tool that can be conducted in almost any forensic laboratory. When properly performed, the results provide a direct visual comparison of the dyes extracted from both known (K) and questioned (Q) fibers. The thin-layer chromatography (TLC) plates or high quality color images of the developed plate make easy-tounderstand exhibits that can be displayed to the court and jury when providing testimony regarding the significance of a fiber comparison. If a reproducible technique is employed in performing this analysis, not only do the developed plates demonstrate the correspondence in the dye bands with respect to both color and retardation factor (Rf), they also permit the semiquantitative estimation of the relative proportions of the individual dyestuffs. Furthermore, the development of a database of the chromatographic, physical, and chemical properties of the most commonly used textile fiber dyestuffs make it possible for a laboratory using the database to identify the dyes by their Colour Index (C.I.) designations [2]. Neumann et al (2011) reports on search algorithms that can be used to screen standardized HPTLC data of the Secret Service Digital Ink Library allowing unknown samples to be compared and identified against a database of known inks [3]. There is a wealth of information in the literature concerning the characterization and identification of a wide range of compounds using TLC [4-7], including a standard guide for the analysis of textile dyestuffs [8]. The development of high-performance thin-layer chromatography (HPTLC) plates, with higher resolution and smaller sample requirements than conventional plates [9], makes this approach more amenable to the limitations of sample size imposed by typical forensic casework samples. Despite the versatility of this method, HPTLC (and TLC, in general) has typically been limited to a strictly comparative role, when utilized at all, in forensic casework. However, HPTLC need not be limited to a strictly comparative role. Investigative HPTLC in forensic casework can provide detailed information about a sample, which cannot be obtained through alternative analytical methods (e.g., MSP, FTIR, Raman, etc.) alone. For example, Palenik et al. (2015) show a
pair of fibers in which different combinations of dyes have resulted in fiber colors that are indistinguishable by microscopy and MSP alone [10]. In addition, the actual identification of the dyes present in an unknown sample may yield additional investigative information, such as application, timescale, or location constraints, depending on the dye, or dye combinations identified. This work explores several analytical variables inherent in basic HPTLC technique that are anticipated to have the largest potential for impact on the quality of the resulting data. Specifically, these topics include: plate selection, pre-elution, tank saturation, developing distance, and eluent stability. Based on the results of these studies, an optimized experimental path has been developed for the separation and analysis of dyes extracted from textile fibers. This controlled approach has been developed to provide an optimized balance between the chromatographic separation of components as well as time considerations while taking into account the ever-present size limitations associated with typical forensic samples. By fixing experimental constraints, the data collected in this work provides a means by which to objectively evaluate the uncertainty associated with this approach. The ability to obtain reproducible HPTLC separations provides a basis from which a) dyes on separate plates, collected at different times can be evaluated for quality, b) tolerances (uncertainty) that can be established for the forensic comparison of questioned and known HPTLC data, and c) a database of HPTLC data from a collection of reference dyes that has been developed under a set of fixed conditions and evaluated for quality [11]. This, in turn provides a means by which the identification of dyes in casework samples can be more reliably and more practically accomplished [12].
2. Experimental Variables and Evaluations The following sections address several aspects of HPTLC analysis, which impact repeatability and resolution. Each sub-section: plate selection, pre-elution of plates, tank saturation, developing distance, and eluent stability, was explored both through a survey of the literature on the topic and actual experimentation to maximize the repeatability and define the anticipated variation of the method when configured for forensic dye separations.
2.1 Plate Selection Chromatography plates are available in a variety of configurations, which include variations in plate types (preparatory, analytical, etc.), sizes, stationary phase thicknesses, adsorbent chemistries, support substrates, with and without visualization additives (among others). Therefore, plate selection is a critical first variable to consider when developing a planar chromatography application. The plates which had been used reliably for forensic TLC dye analyses in the authors’ laboratory for decades were the Whatman HPTLC plate (Cat. #: 05-713-255); however, production was discontinued in approximately 2011. The manufacturer’s specifications for these plates are provided in Table 1. Analogous products from four suppliers (Analtech, EMD Millipore, Machery-Nagel, and Sorbtech) were sourced, and a survey of plate performance of these products was conducted; manufacturers’ specifications for these plates are provided in Table 1. Analtech, at the time, did not have a stock of
HPTLC plates without a fluorescent brightener, and could only supply plates containing a fluorescent dye which is typically added for locating colorless UV absorbing compounds. Communications with chromatography suppliers indicated that the majority of chromatography plates are produced with a fluorescent indicator. The presence of such an indicator is unnecessary for the examination of dyestuffs, which are themselves are highly colored. Furthermore, the presence of a fluorescent indicator presents an added complication to the subsequent evaluation and analysis of developed plate. Each plate was spotted with an equivalent amount of a reference dye mixture (Analtech, test dye I), consisting of four dyes: Sudan IV, Bismarck Brown Y, Rhodamine B, and Fast Green FCF. This dye mixture was used to compare the developing characteristics of the different HPTLC plates. The plates were then dried and developed using an eluent proposed by Wiggins [13]:n-Butanol, acetone, water, and ammonia (5:5:1:2), which was selected for evaluation because it was reported by Wiggins [13] to be applicable to the widest range of dye application classes. The evaluation was completed in triplicate, developing a single plate from each manufacturer on separate days using a fresh batch of eluent. In each evaluation the plates show similar separations of the dye mixture, but also show variations in Rf and band densities. Figure 1 shows a comparison of the developed HPTLC plates from one of the evaluations. The images have been cropped to show only the developed lanes (no post processing of the images was conducted). Since the Whatman plates are no longer commercially manufactured, the data has been included here only to represent a baseline measurement, typical of results that were historically obtained by our laboratory for dye analyses. The stationary phase of the Analtech plate is gray compared to the other surveyed plates, making visualization of the yellow band of the test mixture more difficult. Yellow bands, separate from that of the dye in the test mixture, are present on the upper third of the developed plate from Sorbtech. The presence of these bands also makes visualization difficult, particularly for the yellow dyestuff, but also for minor components that may develop into this region when analyzing unknown samples. The Machery-Nagel and Sorbtech plates each show a loss of resolution of the components at the highest Rf values as well as general lane broadening. The EMD Millipore plate retains the best band density compared to the other evaluated plates and also exhibits the least lane broadening. These characteristics are likely due to the more tightly constrained particle size distribution of this plate, reported to be 5-7 µm, compared to the others which are in the range of 2-10 µm. In comparison to the discontinued Whatman plates, the EMD plate shows a comparable level of performance for the test dye that was studied. Based upon this performance comparison, the EMD Millipore plate was selected for use in the remainder of experiments reported in this paper.
2.2 Pre-Elution Historically, TLC analysis begins with “activating” a plate [4]. TLC plate activation involved predeveloping a plate with an elution solvent to remove contaminants and impurities, followed by a period of heating to drive off residual solvent and water to maximize adsorptive capacity of the stationary phase [14]. This pre-analysis step was designed to optimize development conditions and ensure consistent results from plates, which were typically made in-house.
Commercial production of TLC plates minimizes plate to plate differences with better constraints on stationary phase thickness and purity. Additionally, the micronization of the stationary phase particles for HPTLC dramatically increases the surface area for adsorption, providing a greater level of sorbent activity. Commercially produced TLC plates are typically activated prior to packaging; however, the storage conditions of the plates within a laboratory can greatly affect sorbent activity [15]. Storage of plates in areas free of moisture and solvents will minimize the need for (re-)activation in the laboratory. With the availability of uniform, pre-activated plates, the authors have explored the relevance of plate activation in everyday casework. HPTLC plates stored under ambient laboratory conditions (~21 °C, 45% RH) were scored and broken in half. One half was developed without any sample and dried in an oven (~65 °C) for 30 minutes. The second half was left in the oven to avoid exposure to additional moisture. The two halves of the plate (one “activated”, one untreated) were each spotted with the same quantity and combination of textile dyes. The plates were simultaneously developed and evaluated. In general, the pre-developed (i.e., activated) plates show broadened bands, tailing, and variation in Rf compared to neat plates (Figure 2). The loss of resolution and lane broadening observed after inhouse pre-elution suggests that the commercial production of HPTLC plates produces a reliable product that can be used without the need for additional activation. Given the potential effect that storage conditions can have upon sorbent activity [15], the performance of the test dye mixture over time was evaluated. Test dye samples were run on plates from the same lot over a period of ~3.5 years. The plates were stored in their factory packaging in a cabinet away from solvents and other chemicals. Figure 3 shows an example of plates from the same lot developed with the same test dye mixture over the course of ~3.5 years. An evaluation of the four major dyes in the test dye shows that the Rf values vary by less than 4.9% relative standard deviation (RSD) over this period, which is within the uncertainty of the method (See §2.6 Evaluating Reproducibility). Therefore, even over a period of ~3.5 years, with controlled storage conditions, no significant broadening or Rf variations were noted.
2.3 Tank Saturation Tank saturation is an important variable associated with planar chromatography. The developing chamber (interior volume ~825 cm3) is loaded with the elution solvent (~40 mL for this size tank) and allowed to equilibrate to generate a uniform environment for analysis. The time period needed to obtain saturation, and therefore equilibrium, in the tank is dependent on the composition of the mobile phase, temperature, and the volume of the developing chamber [16]. It is commonly advised that the developing tank should be lined with filter paper to encourage and retain this state, as the equilibrium can be disturbed by changing solvents, conducting multiple runs with the same solvent, or leaving the chamber open for prolonged periods. While this work has not attempted to determine the point at which equilibrium is reached, these empirical studies provide insight into whether a selected equilibrium period is appropriate for a given configuration. In this work, the mobile phase was added to a conventional rectangular glass container with a removable lid tank and fitted with filter paper along the wide, vertical wall of the chamber (interior dimensions ~12.5 x 6 x 11 cm). In this way, the plate could still be observed during development. This setup represents the simplest of the various developing setups that are available and have
been utilized for dye separations [17]; yet, it is amenable to the limitations of casework sized samples. The chamber was covered and allowed to equilibrate overnight. Plates were loaded into the tank while minimizing the amount of time that the lid was opened. Serial development of plates was performed with a two hour recovery period between runs. While this duration was initially selected because it provided sufficient time to document a developed plate, calculate Rf values, and spot and dry a new plate for analysis, the results (discussed in eluent stability) illustrate that this period is sufficient to obtain consistent data. Further work would be needed to justify shorter periods of equilibration and recovery.
2.4 Developing Distance Optimization of developing distance is the determination of the height to which the solvent front is permitted to advance on a plate. Bearing in mind that capillary flow rate of the mobile phase is not constant, and distance varies as a square root function of time, developing distance also has an impact on the speed of an analysis (i.e., greater developing distances require increasingly longer separation times). A review of the literature shows that an “optimal” range of developing distances, varying from 2 to 18 cm, have been proposed. These distances are related to either specific methods or general guidelines proposed for TLC analysis [4, 13, 18]. The selected distance directly impacts the separation and density of bands on a plate (i.e., the ability to separate closely spaced bands) and is affected by a several variables. For instance, smaller particle size of the stationary phase of HPTLC plates results in better resolution over a comparatively shorter developing distance relative to traditional TLC plates [9]. From a simplistic viewpoint, an optimal developing distance might be considered as the point when adequate separation of the components has occurred. While this is a practical approach, it does not necessarily ensure optimal resolution. From a theoretical standpoint, defining optimal resolution in planar chromatography will in turn specify the developing distance. In column chromatography, resolution is described through the van Deemter equation, which describes resolution as a number of theoretical “disks” into which the column is divided; the higher the number, the greater the resolution. Optimizing flow rate to the smallest disk height yields the greatest number of plates, within a fixed distance, and consequently the best resolution. However, in planar chromatography the flow rate of the mobile phase is not constant. Modifying the van Deemter equation to relate developing distance (instead of flow rate) to theoretical disk height allows for a calculation of optimal resolution [19]. Based on an adsorbent particle size of 6 µm (the average value of the EMD Millipore plate selected for use in this work), the optimal resolution corresponds to a developing distance of approximately 4 cm [20]. For the topic of interest presented in this work, systematic forensic dye analysis, the subject of developing distance was also explored empirically. Four HPTLC plates were developed to varying distances: 2, 2.5, 3.75, and 5 cm. The resulting plates, which were otherwise developed under the same conditions, are presented in Figure 4. The Rf of each component was measured (Table 2) and examination of the values shows a maximum variation in Rf of 0.02, which is within the standard
deviation for each band (for further discussion of this point, see the latter section on evaluating reproducibility). Short developing distances result in faster analysis times with more concentrated dye bands, but provide less separation between them, which can ultimately result in a loss of resolution if reduced beyond a certain threshold. The converse is true of longer developing distances, which come with the drawback of poorer visualization (i.e., loss of band density). Of the four distances surveyed, a developing distance of 3.75 cm(~1.5 in.) was selected as the optimal compromise between band density, resolution, and analysis time. This distance is very close to the theoretical maximum resolution for this plate, which was discovered after the empirical evaluation had been completed.
2.5 Eluent Stability The stability of a developing solution is directly related to the repeatability of an analysis. Developing solutions are typically not shelf stable for extended periods of time, and it has been suggested in the literature that developing solutions be prepared frequently, especially if highly volatile components are used [21]. With use (i.e., opening a saturated chamber, or developing multiple plates), the composition of the mobile phase can change, which, in turn, can impact the repeatability of plate development [15]. It has been suggested that any developing solution composed of more than a single component should be discarded after a single use [15]; however, this becomes costly and time consuming when the analysis of multiple plates is required. One way to evaluate the stability of an eluent is to monitor the developing characteristics of the plates by means of a standard dye mixture [17]. This practice is common in drug chemistry laboratories whereby plates are evaluated against an archived reference. To evaluate the potential impact of eluent age on plate repeatability, test dye data was examined from 39 plates for which the age of the eluent was known, both in terms of the time since eluent preparation, and the number of plates run since the eluent was changed. Figure 5a shows the average Rf for each of the six bands in the test dye, relative to the number of days since the eluent was prepared. Figure 5b shows analogous data for test dye bands relative to the number of plates run since the eluent was changed. The dashed lines in Figure 5a represent a range of one standard deviation from the mean Rf value for each band from plates developed the same day that a fresh batch of eluent was prepared. The dashed lines in Figure 5b represent a range of one standard deviation from the mean Rf value for each test dye band from the first developed plate in a fresh eluent. Examination of Figure 5a shows that on days one through three, after the preparation of a new eluent, there is a decrease in the average value of bands 3, 4, 5 and 6. However, the average Rf for bands 1 and 2 suggest an increasing trend. Therefore, no consistent trend in Rf could be identified as a function of eluent age. As expected, some variation occurs outside of the calculated one standard deviation range; however, the largest outliers are from bands 3 and 4, which are weak and diffuse, respectively, making them more prone to measurement errors. Ultimately, all variation observed over the eluent age of six days fell within 2 standard deviations, and most (77 %) fell within 1 standard deviation.
Similarly, Figure 5b provides insight into possible exhaustion of the eluent due to overuse (e.g., changes in composition). There are no consistent trends in the average band positions. While some bands show variation exceeding one standard deviation from the fresh eluent average, particularly bands 3 and 4 (which are weak and diffuse, respectively), all variation is again within 2 standard deviations of the fresh eluent average, and most (~80 %) is within 1 standard deviation, even after six plates were run in the same eluent. Together, the eluent aging and use data both demonstrate the stability of the parameters used for the analysis of multiple plates over a period of several days. Therefore, as a conservative rule of thumb, the authors proceeded with a guideline that eluent should be replaced after no more than 5 days, which is consistent with the shelf life reported by Wiggins [13], or 5 plate runs.
2.6 Evaluating Reproducibility Application of the factors that were studied up to this point provides the basis for a reproducible approach to the development of TLC plates, which in turn, provides a means by which data obtained from multiple plates can be relied upon to provide Rf values of sufficient validity for database searching. In terms of our goal of identifying the separated dyes, this offers a practical means by which to build a reference dye database. However, to ensure that the results obtained under such conditions are consistent, and to assist with the development of reasonable search tolerances, it was also necessary to develop an understanding of the expected variation in Rf values. Lederer [5] provides some guidance on the topic of Rf value tolerances and suggests that variations in the Rf of a developed band should be within 0.02, when the operating temperature is well controlled and the system is at equilibrium. When this criterion is applied to the dataset collected here, consisting of the test dye separated on 54 plates, 43% of the data would be rejected. Examination of the data reveals that this tolerance is exceeded more often for bands with a low Rf values relative to those with a high Rf values (56% for bands 1 and 2 vs. 12% for bands 5 and 6). This is consistent with the data presented in Figure 5, which indicates a general increase in the standard deviation range (dotted lines) as a function of decreasing Rf. Given the empirical evidence that band position uncertainty is not accurately represented by a fixed value, we set out to establish a more appropriate means of assigning uncertainty by evaluating test dye data. The average, range, and standard deviation for each of the six bands in the test dye mixture are presented in Table 3. A plot of the standard deviation calculated from the Rf data for each band, as a function of Rf value of each band is plotted in Figure 6. With the exception of Band 4, the data show a strong inverse linear correlation (R2=0.995). Band 4 is a comparatively broad band, relative to the other five bands of the test dye, from which it is more difficult to manually extract an Rf value. The inverse relationship between retardation factor and uncertainty provides, for the plate conditions evaluated, a means by which one can ensure the quality of a developed plate and define reproducibility of a measurement.
3. Discussions Variables of plate selection, plate pre-elution, tank saturation, developing distance, and eluent stability, were explored both through a survey of the literature and experimentation to optimize the
experimental conditions and data repeatability. The results from these evaluations provide an objective means by which to evaluate the variation of the operating conditions studied throughout this paper. While the results presented here are directly applicable to this specific set of conditions, various general lessons arose from this study, which can be applied to other systems:
Numerous planar chromatography plate configurations are available (e.g., HPTLC/TLC; size; stationary phase composition; and the addition of visualization aids). Appropriate selection of these factors for a specified analytical goal can impact the resolution of the method, the ability to recover separated analyte, and cost. An evaluation of comparable plates from several suppliers demonstrated the benefits of a specific combination of these factors to the goal of forensic dye separation.
The commercial production of HPTLC plates provides a highly uniform product that can be used without additional stationary phase activation. In fact, in this small study, a laboratory-based, pre-development process contributed undesirable artifacts.
Saturation of the developing chamber for planar chromatography is critical for achieving and maintaining repeatability. Ultimately, the time required to reach equilibrium is dependent on many factors, including chamber volume and mobile phase composition. A two hour equilibrating period was used in this research.
The multi-phase Wiggins #1 eluent [12] has a reported shelf life of one week. This work confirms that the eluent may be stable for up to five days or the development five HPTLC plates (whichever comes first). Ultimately, the stability of an eluent should be monitored through the use of a test dye mixture and defined assessment criteria, such as that discussed in this paper.
The use of a test dye mixture on each plate is a simple and beneficial way to assess the uncertainty of the method. It also provides a standard that permits the quality of inter-plate and archival data to be evaluated.
The distance to which the mobile phase is eluted is tied to developing time and band resolution. For the system evaluated here, a developing distance of 3.75 cm represents a good compromise between band density and analysis time. This distance is also very close to the theoretical optimal developing distance of 4 cm, which represents the maximum resolution for a plate with a 6 µm silica particle size.
Variation in Rf value is dependent on the complexity of the mobile phase. The multicomponent mobile phase used in this work, Wiggins [13] #1, results in an uncertainty that scales with an inverse linear relationship to Rf.
4. Conclusions Planar chromatography remains a low-cost, highly sensitive analytical tool for the isolation, separation, and purification of textile dyestuffs. When performed in the context of a forensic dye analysis, HPTLC has been generally limited to the realm of a comparative analytical technique that requires the simultaneous analysis of known and questioned samples. However, when the
variables discussed in this work are defined and controlled, the results between different plates and even between different laboratories can be reproducibly developed and objectively evaluated. While variation is inevitable in any analytical technique, an understanding of the impact of experimental parameters on the resulting data permits results to be evaluated for quality and compared over time. In forensic dye analysis, this opens the possibility of developing collections of reference data. The results of this research provide practical insight that is specific to a particular HPTLC separation configuration that is optimized for general forensic dye analysis but which may be utilized as a guideline to optimize and evaluate other experimental TLC separation configurations.
Acknowledgements This work was partially supported by the National Institute of Justice (Grant number 2012-DN-BXK042). The authors are grateful for the efforts of Stephanie Friewald, who spotted, developed, and documented the several dozen HPTLC plates for this project.
References [1] M. Grieve, The role of fibers in forensic science examinations, J. For. Sci. 28 (4) (1983) 877–887. [2] E. Groves, S. Palenik, C.S. Palenik, A Generalized Approach to Forensic Dye Identification: Development and Utility of Reference Libraries. Journal of the AOAC International. In press. [3] C. Neumann, R. Ramotowski, T. Genessay, Forensic Examination of ink by high-performance thinlayer chromatography – The United State Secret Service Digital Ink Library. Journal of Chromatography A. 1218 (2011) 2793-2811. [4] E. Stahl, Thin-Layer Chromatography: A Laboratory Handbook. Second Edition. New York: SpringerVerlag, 1969. [5] E. Lederer, M. Lederer, Chromatography. Second Edition. New York: Elsevier. 1957. [6] F. Fadil, W. McSharry, Extraction and TLC separation of food, drug, and cosmetic dyes from tabletcoating formulations, J. Pharm. Sci. 68 (1) (1979) 97–98. [7] L.C. Lee, Feasibility of High Performance Thin Layer Chromatography for the Forensic Analysis of Ballpoint Pen Inks, Problems of Forensic Science. Vol 97 (2014) 14-22. [8] ASTM E2227-13, Standard Guide for Forensic Examination of Non-Reactive Dyes in Textile Fibers by Thin-Layer Chromatography, ASTM International, West Conshohocken, PA, 2013 www.astm.org. [9] A. Zlatkis, R.E. Kaiser. HPTLC, High Performance Thin-layer Chromatography. Amsterdam: Elsevier Scientific Pub., 1977. [10] C.S. Palenik, J.C. Beckert, S.J. Palenik, Microspectrophotometry of Fibers: Advances in Analysis and Interpretation. NIJ Report 2012-DN-BX-K040 (2015). [11] E. Groves, The Discriminating Power of High Performance Thin Layer Chromatography (HPTLC) for Commercial Textile Dyestuffs. Inter/Micro; 2014 Jun 2-4; Chicago, IL. Microscope 62(3) (2014) 103. [12] C.S. Palenik, E. Groves, S. Palenik, Dye Identification in Casework: How Far Can You Go? Inter/Micro; 2015 Jun 6-8; Chicago, IL. Microscope 63(3) (2015) 115. [13] K. Wiggins, Thin Layer Chromatographic Analysis for Fibre Dyes. Forensic Examination of Fibres, Second Edition. London: Taylor & Francis, (1999) 291-310. [14] R.J. Hamilton, S. Hamilton, D. Kealey. Thin Layer Chromatography. ACOL, London, Wiley, 1987. [15] J. Touchstone, Practice of Thin Layer Chromatography. Third Edition. New York: Wiley, 1992. [16] K. Randerath, Thin-Layer Chromatography. Second Edition. New York: Academic Press. 1966.
[17] D. Laing, The standardization of TLC for comparison of fibre dyes, Journal of the Forensic Science Society. Vol. 30, No. 5 (1991) 299-307. [18] J. Dean, Chemical Separation Methods. New York: Can Nostrand Reinhold, 1969. [19] C. Poole, Multidimensionality in Planar Chromatography. Journal of Chromatography A. Vol. 703, No. 1-2 (1995) 573-612. [20] B. Spangenberg, Theoretical Basis of Thin Layer Chromatography (TLC). Quantitative Thin-Layer Chromatography: A Practical Survey. Berlin: Springer-Verlag, 2011. [21] J. Bobbit, Thin-Layer Chromatography. Third Edition. New York: Reinhold, 1963.
Tables Table 1. Specifications of the HPTLC plates examined in this study. Manufacturer Analtech EMD Millipore Machery-Nagel Sorbtech Whatman
Adsorbent Fluor. Thickness Particle size Chemistry indicator 61077 150 µm 8-10 µm Yes 13748 150-200 µm 5-7 µm No Unmodified 811032 200 µm 2-10 µm No Silica 4214056 200 µm 2-10 µm No 05-713-255 200 µm 4.5 µm No
Pre-conc. Zone size 1.5 cm 2.5 cm 2.9 cm 2.8 cm 2.0 cm
Catalog #
Plate size
Plate substrate
10 x 10 cm
Glass
Table 2. Retardation factor (Rf) measurements for the test dye bands developed varying distances. The Rf value of each band was manually measured in triplicate and the average value is provided. Band # Band 6 Band 5 Band 4 Band 3 Band 2 Band 1
Developing Distance (cm) 2.0 2.5 3.75 5.0 0.99 0.99 0.98 0.97 0.90 0.90 0.90 0.89 0.76 0.78 0.78 0.77 0.66 0.67 0.68 0.66 0.55 0.55 0.53 0.53 0.52 0.52 0.50 0.51
Min
Max
Range
0.97 0.89 0.76 0.66 0.53 0.50
0.99 0.90 0.78 0.67 0.55 0.52
0.02 0.01 0.02 0.01 0.02 0.02
Table 3. Average, standard deviation (σ), and range of Rf values for the test dye mixture (n=54). Band #
Avg. Rf
3σ
Band 6 Band 5 Band 4 Band 3 Band 2 Band 1
0.98 0.88 0.68 0.62 0.49 0.46
0.03 0.04 0.12 0.09 0.10 0.11
Range Min Max 0.96 0.99 0.85 0.91 0.60 0.78 0.56 0.68 0.38 0.55 0.34 0.53
Groves, Palenik (2018) – Reproducibility of High-Performance Thin-Layer Chromatography (HPTLC) in Textile Dye Analysis.
Figure captions: Figure 1. Comparison of the five surveyed plates. Left to right: Analtech, Sorbtech, Machery-Nagel, EMD Millipore, and Whatman. The same dye mixture has been developed on each plate under comparable conditions. The labels on the left identify the developed bands by number. Figure 2. Comparison of developed dyes on neat (left) and pre-developed (right) HPTLC plates. Figure 3. Comparison of four HPTLC plates developed over a period of several years. The plates were stored under the same conditions within our laboratory. Approximate age (months): 10, 23, 35, 40 (left to right). Figure 4. Set of four HPTLC plates developed varying distances. Left to right: 2.0, 2.5, 3.75, 5.0 cm. Figure 5. A plot of retention factors measured from 39 separate HPTLC plate runs of a test dye mixture that consists of six bands. The age and number of plates run between eluent changes was monitored. These two plots and show the variation in retention factor as a function of eluent age in terms of (A) time since preparation and (B) plates run since the eluent was changed. The dashed lines represent a range of one standard deviation from the mean for all bands from either day 1 (A) or plate 1 (B). Note that day 1 represents the day that the eluent was changed. Figure 6. Standard deviation as a function of retention using Wiggins # 1 eluent, on an EMD Millipore HPTLC plate developed a distance of 3.75 cm. The linear trendline has been calculated without Band 4 (bright pink data point), which shows a higher deviation due to the diffuse shape of this band.
Groves, Palenik (2018) – Reproducibility of High-Performance Thin-Layer Chromatography (HPTLC) in Textile Dye Analysis.
Highlights:
Forensic HPTLC analysis is not strictly a comparative tool. A test dye mixture can be used to monitor and assess HPTLC data quality. Libraries of HPTLC data of dyes are possible with controlled analysis conditions. Uncertainty across the HPTLC plate is not uniform for multi-component eluents. Libraries of HPTLC data may lead to more investigative exploration of dyed evidence.
6 5
4 3
2 1
Figure 1
Test Dye Mixture
Acid Red 92
Figure 2
Acid Orange 116
Figure 3
Figure 4
B 1.00
1.00
0.90
0.90
Retardation Factor
Retardation Factor
A
0.80 0.70 0.60 0.50
0.70 0.60 0.50
0.40
0.40 0
1
2
3
4
Days Since Eluent Change
Figure 5
0.80
5
6
7
0
1
2
3
4
5
Plates Run Between Eluent Change
6
7
0.12
Standard Deviation (3σ)
0.10
0.08 y = -0.1475x + 0.1737 R² = 0.9951 0.06
0.04
0.02
0.00 0.40
Figure 6
0.50
0.60
0.70 Retardation Factor (Rf)
0.80
0.90
1.00
Graphical abstract
S2468-1709(17)30135-2 https://doi.org/10.1016/j.forc.2018.03.004 FORC 98
To appear in:
Forensic Chemistry
Received Date: Revised Date: Accepted Date:
31 October 2017 6 March 2018 10 March 2018
Please cite this article as: E. Groves, S. Palenik, C.S. Palenik, Reproducibility of High-Performance Thin-Layer Chromatography (HPTLC) in Textile Dye Analysis, Forensic Chemistry (2018), doi: https://doi.org/10.1016/j.forc. 2018.03.004
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Reproducibility of High-Performance Thin-Layer Chromatography (HPTLC) in Textile Dye Analysis
Ethan Groves, B.S., Skip Palenik, B.S., and Christopher S. Palenik, Ph.D. Microtrace LLC, 790 Fletcher Dr., Suite 106, Elgin, IL 60123
Abstract In the forensic comparison of dyed fibers, thin-layer chromatography (TLC) is one of the few analytical techniques that have proven sensitive enough to detect and separate the minute quantities of dyestuffs present in single fibers. The method has become well-established by trace evidence examiners as a means of further distinguishing fibers whose colors appear to be the same by both comparison microscopy and microspectrophotometry. As practiced at present, the forensic analysis of dyes is limited to a comparison of the separated dye bands from known and questioned fibers performed on the same TLC plate. It is recognized, however, that retardation factor (Rf) alone is not sufficient proof of identify. The limited use of investigative TLC analysis in forensic fiber examinations is due in part to the range of uncontrolled or poorly defined variables that affect the reproducibility of the developed plate. Through the study of a six component test dye mixture developed on over 50 high-performance thin-layer chromatography plates, the effects of several critical variables that affect reproducibility and resolution, including: plate selection, pre-elution, tank saturation, developing distance, and eluent stability have been evaluated. When considered collectively, the results of this research provide a means for acquiring and archiving repeatable data, both from casework and known reference samples collected on different plates and at different times. These provide a pathway for the development and utilization of reference databases for the identification of dyes. The empirical uncertainty established for the generalized separation procedure used in this research provides objective guidance for evaluating the significance of associations (or eliminations) made on the basis Rf. Ultimately, this research also opens a pathway to the use of forensic dye analysis as an investigative tool, rather than one exclusively restricted to comparative analyses.
Keywords Forensic science, dye identification, textile dyes, high performance thin-layer chromatography (HPTLC), reproducibility.
1. Introduction Fiber comparisons represent a significant proportion of all forensic trace evidence analyses [1]. Color is perhaps the most important optical property exhibited by evidential fibers, since along with thickness and shape these are the easiest characteristics to recognize when searching for target fibers through a stereomicroscope. In addition to polymer identification by polarized light microscopy (PLM) and Fourier-transform infrared microspectroscopy (micro-FTIR), the principal property relied upon in most fiber comparison is color, which is evaluated most commonly by comparison microscopy and microspectrophotometry (MSP). It may come as a surprise to some that the colors of most fibers are only rarely, if ever, due to absorptions from a single dyestuff. Since colors are commonly matched to shades specified by their customers, most dye houses are experts at color matching. They characterize these colors using the same spectrophotometric techniques as those employed by forensic scientists. From a spectrum obtained from an exemplar fabric, a dye house will formulate a dye mixture, which can change over time due to changes in dye lots, dyestuff availability (as a function of price and/or quality), and a master dyer’s preferences for specific dyestuffs. One consequence of this method of coloring fibers, by accurately matching a spectrophotometrically specified color produced from different dyes, is an additional level of discrimination in forensic fiber comparisons that is achieved by dye identification. Planar chromatography is a cost effective, routine analytical tool that can be conducted in almost any forensic laboratory. When properly performed, the results provide a direct visual comparison of the dyes extracted from both known (K) and questioned (Q) fibers. The thin-layer chromatography (TLC) plates or high quality color images of the developed plate make easy-tounderstand exhibits that can be displayed to the court and jury when providing testimony regarding the significance of a fiber comparison. If a reproducible technique is employed in performing this analysis, not only do the developed plates demonstrate the correspondence in the dye bands with respect to both color and retardation factor (Rf), they also permit the semiquantitative estimation of the relative proportions of the individual dyestuffs. Furthermore, the development of a database of the chromatographic, physical, and chemical properties of the most commonly used textile fiber dyestuffs make it possible for a laboratory using the database to identify the dyes by their Colour Index (C.I.) designations [2]. Neumann et al (2011) reports on search algorithms that can be used to screen standardized HPTLC data of the Secret Service Digital Ink Library allowing unknown samples to be compared and identified against a database of known inks [3]. There is a wealth of information in the literature concerning the characterization and identification of a wide range of compounds using TLC [4-7], including a standard guide for the analysis of textile dyestuffs [8]. The development of high-performance thin-layer chromatography (HPTLC) plates, with higher resolution and smaller sample requirements than conventional plates [9], makes this approach more amenable to the limitations of sample size imposed by typical forensic casework samples. Despite the versatility of this method, HPTLC (and TLC, in general) has typically been limited to a strictly comparative role, when utilized at all, in forensic casework. However, HPTLC need not be limited to a strictly comparative role. Investigative HPTLC in forensic casework can provide detailed information about a sample, which cannot be obtained through alternative analytical methods (e.g., MSP, FTIR, Raman, etc.) alone. For example, Palenik et al. (2015) show a
pair of fibers in which different combinations of dyes have resulted in fiber colors that are indistinguishable by microscopy and MSP alone [10]. In addition, the actual identification of the dyes present in an unknown sample may yield additional investigative information, such as application, timescale, or location constraints, depending on the dye, or dye combinations identified. This work explores several analytical variables inherent in basic HPTLC technique that are anticipated to have the largest potential for impact on the quality of the resulting data. Specifically, these topics include: plate selection, pre-elution, tank saturation, developing distance, and eluent stability. Based on the results of these studies, an optimized experimental path has been developed for the separation and analysis of dyes extracted from textile fibers. This controlled approach has been developed to provide an optimized balance between the chromatographic separation of components as well as time considerations while taking into account the ever-present size limitations associated with typical forensic samples. By fixing experimental constraints, the data collected in this work provides a means by which to objectively evaluate the uncertainty associated with this approach. The ability to obtain reproducible HPTLC separations provides a basis from which a) dyes on separate plates, collected at different times can be evaluated for quality, b) tolerances (uncertainty) that can be established for the forensic comparison of questioned and known HPTLC data, and c) a database of HPTLC data from a collection of reference dyes that has been developed under a set of fixed conditions and evaluated for quality [11]. This, in turn provides a means by which the identification of dyes in casework samples can be more reliably and more practically accomplished [12].
2. Experimental Variables and Evaluations The following sections address several aspects of HPTLC analysis, which impact repeatability and resolution. Each sub-section: plate selection, pre-elution of plates, tank saturation, developing distance, and eluent stability, was explored both through a survey of the literature on the topic and actual experimentation to maximize the repeatability and define the anticipated variation of the method when configured for forensic dye separations.
2.1 Plate Selection Chromatography plates are available in a variety of configurations, which include variations in plate types (preparatory, analytical, etc.), sizes, stationary phase thicknesses, adsorbent chemistries, support substrates, with and without visualization additives (among others). Therefore, plate selection is a critical first variable to consider when developing a planar chromatography application. The plates which had been used reliably for forensic TLC dye analyses in the authors’ laboratory for decades were the Whatman HPTLC plate (Cat. #: 05-713-255); however, production was discontinued in approximately 2011. The manufacturer’s specifications for these plates are provided in Table 1. Analogous products from four suppliers (Analtech, EMD Millipore, Machery-Nagel, and Sorbtech) were sourced, and a survey of plate performance of these products was conducted; manufacturers’ specifications for these plates are provided in Table 1. Analtech, at the time, did not have a stock of
HPTLC plates without a fluorescent brightener, and could only supply plates containing a fluorescent dye which is typically added for locating colorless UV absorbing compounds. Communications with chromatography suppliers indicated that the majority of chromatography plates are produced with a fluorescent indicator. The presence of such an indicator is unnecessary for the examination of dyestuffs, which are themselves are highly colored. Furthermore, the presence of a fluorescent indicator presents an added complication to the subsequent evaluation and analysis of developed plate. Each plate was spotted with an equivalent amount of a reference dye mixture (Analtech, test dye I), consisting of four dyes: Sudan IV, Bismarck Brown Y, Rhodamine B, and Fast Green FCF. This dye mixture was used to compare the developing characteristics of the different HPTLC plates. The plates were then dried and developed using an eluent proposed by Wiggins [13]:n-Butanol, acetone, water, and ammonia (5:5:1:2), which was selected for evaluation because it was reported by Wiggins [13] to be applicable to the widest range of dye application classes. The evaluation was completed in triplicate, developing a single plate from each manufacturer on separate days using a fresh batch of eluent. In each evaluation the plates show similar separations of the dye mixture, but also show variations in Rf and band densities. Figure 1 shows a comparison of the developed HPTLC plates from one of the evaluations. The images have been cropped to show only the developed lanes (no post processing of the images was conducted). Since the Whatman plates are no longer commercially manufactured, the data has been included here only to represent a baseline measurement, typical of results that were historically obtained by our laboratory for dye analyses. The stationary phase of the Analtech plate is gray compared to the other surveyed plates, making visualization of the yellow band of the test mixture more difficult. Yellow bands, separate from that of the dye in the test mixture, are present on the upper third of the developed plate from Sorbtech. The presence of these bands also makes visualization difficult, particularly for the yellow dyestuff, but also for minor components that may develop into this region when analyzing unknown samples. The Machery-Nagel and Sorbtech plates each show a loss of resolution of the components at the highest Rf values as well as general lane broadening. The EMD Millipore plate retains the best band density compared to the other evaluated plates and also exhibits the least lane broadening. These characteristics are likely due to the more tightly constrained particle size distribution of this plate, reported to be 5-7 µm, compared to the others which are in the range of 2-10 µm. In comparison to the discontinued Whatman plates, the EMD plate shows a comparable level of performance for the test dye that was studied. Based upon this performance comparison, the EMD Millipore plate was selected for use in the remainder of experiments reported in this paper.
2.2 Pre-Elution Historically, TLC analysis begins with “activating” a plate [4]. TLC plate activation involved predeveloping a plate with an elution solvent to remove contaminants and impurities, followed by a period of heating to drive off residual solvent and water to maximize adsorptive capacity of the stationary phase [14]. This pre-analysis step was designed to optimize development conditions and ensure consistent results from plates, which were typically made in-house.
Commercial production of TLC plates minimizes plate to plate differences with better constraints on stationary phase thickness and purity. Additionally, the micronization of the stationary phase particles for HPTLC dramatically increases the surface area for adsorption, providing a greater level of sorbent activity. Commercially produced TLC plates are typically activated prior to packaging; however, the storage conditions of the plates within a laboratory can greatly affect sorbent activity [15]. Storage of plates in areas free of moisture and solvents will minimize the need for (re-)activation in the laboratory. With the availability of uniform, pre-activated plates, the authors have explored the relevance of plate activation in everyday casework. HPTLC plates stored under ambient laboratory conditions (~21 °C, 45% RH) were scored and broken in half. One half was developed without any sample and dried in an oven (~65 °C) for 30 minutes. The second half was left in the oven to avoid exposure to additional moisture. The two halves of the plate (one “activated”, one untreated) were each spotted with the same quantity and combination of textile dyes. The plates were simultaneously developed and evaluated. In general, the pre-developed (i.e., activated) plates show broadened bands, tailing, and variation in Rf compared to neat plates (Figure 2). The loss of resolution and lane broadening observed after inhouse pre-elution suggests that the commercial production of HPTLC plates produces a reliable product that can be used without the need for additional activation. Given the potential effect that storage conditions can have upon sorbent activity [15], the performance of the test dye mixture over time was evaluated. Test dye samples were run on plates from the same lot over a period of ~3.5 years. The plates were stored in their factory packaging in a cabinet away from solvents and other chemicals. Figure 3 shows an example of plates from the same lot developed with the same test dye mixture over the course of ~3.5 years. An evaluation of the four major dyes in the test dye shows that the Rf values vary by less than 4.9% relative standard deviation (RSD) over this period, which is within the uncertainty of the method (See §2.6 Evaluating Reproducibility). Therefore, even over a period of ~3.5 years, with controlled storage conditions, no significant broadening or Rf variations were noted.
2.3 Tank Saturation Tank saturation is an important variable associated with planar chromatography. The developing chamber (interior volume ~825 cm3) is loaded with the elution solvent (~40 mL for this size tank) and allowed to equilibrate to generate a uniform environment for analysis. The time period needed to obtain saturation, and therefore equilibrium, in the tank is dependent on the composition of the mobile phase, temperature, and the volume of the developing chamber [16]. It is commonly advised that the developing tank should be lined with filter paper to encourage and retain this state, as the equilibrium can be disturbed by changing solvents, conducting multiple runs with the same solvent, or leaving the chamber open for prolonged periods. While this work has not attempted to determine the point at which equilibrium is reached, these empirical studies provide insight into whether a selected equilibrium period is appropriate for a given configuration. In this work, the mobile phase was added to a conventional rectangular glass container with a removable lid tank and fitted with filter paper along the wide, vertical wall of the chamber (interior dimensions ~12.5 x 6 x 11 cm). In this way, the plate could still be observed during development. This setup represents the simplest of the various developing setups that are available and have
been utilized for dye separations [17]; yet, it is amenable to the limitations of casework sized samples. The chamber was covered and allowed to equilibrate overnight. Plates were loaded into the tank while minimizing the amount of time that the lid was opened. Serial development of plates was performed with a two hour recovery period between runs. While this duration was initially selected because it provided sufficient time to document a developed plate, calculate Rf values, and spot and dry a new plate for analysis, the results (discussed in eluent stability) illustrate that this period is sufficient to obtain consistent data. Further work would be needed to justify shorter periods of equilibration and recovery.
2.4 Developing Distance Optimization of developing distance is the determination of the height to which the solvent front is permitted to advance on a plate. Bearing in mind that capillary flow rate of the mobile phase is not constant, and distance varies as a square root function of time, developing distance also has an impact on the speed of an analysis (i.e., greater developing distances require increasingly longer separation times). A review of the literature shows that an “optimal” range of developing distances, varying from 2 to 18 cm, have been proposed. These distances are related to either specific methods or general guidelines proposed for TLC analysis [4, 13, 18]. The selected distance directly impacts the separation and density of bands on a plate (i.e., the ability to separate closely spaced bands) and is affected by a several variables. For instance, smaller particle size of the stationary phase of HPTLC plates results in better resolution over a comparatively shorter developing distance relative to traditional TLC plates [9]. From a simplistic viewpoint, an optimal developing distance might be considered as the point when adequate separation of the components has occurred. While this is a practical approach, it does not necessarily ensure optimal resolution. From a theoretical standpoint, defining optimal resolution in planar chromatography will in turn specify the developing distance. In column chromatography, resolution is described through the van Deemter equation, which describes resolution as a number of theoretical “disks” into which the column is divided; the higher the number, the greater the resolution. Optimizing flow rate to the smallest disk height yields the greatest number of plates, within a fixed distance, and consequently the best resolution. However, in planar chromatography the flow rate of the mobile phase is not constant. Modifying the van Deemter equation to relate developing distance (instead of flow rate) to theoretical disk height allows for a calculation of optimal resolution [19]. Based on an adsorbent particle size of 6 µm (the average value of the EMD Millipore plate selected for use in this work), the optimal resolution corresponds to a developing distance of approximately 4 cm [20]. For the topic of interest presented in this work, systematic forensic dye analysis, the subject of developing distance was also explored empirically. Four HPTLC plates were developed to varying distances: 2, 2.5, 3.75, and 5 cm. The resulting plates, which were otherwise developed under the same conditions, are presented in Figure 4. The Rf of each component was measured (Table 2) and examination of the values shows a maximum variation in Rf of 0.02, which is within the standard
deviation for each band (for further discussion of this point, see the latter section on evaluating reproducibility). Short developing distances result in faster analysis times with more concentrated dye bands, but provide less separation between them, which can ultimately result in a loss of resolution if reduced beyond a certain threshold. The converse is true of longer developing distances, which come with the drawback of poorer visualization (i.e., loss of band density). Of the four distances surveyed, a developing distance of 3.75 cm(~1.5 in.) was selected as the optimal compromise between band density, resolution, and analysis time. This distance is very close to the theoretical maximum resolution for this plate, which was discovered after the empirical evaluation had been completed.
2.5 Eluent Stability The stability of a developing solution is directly related to the repeatability of an analysis. Developing solutions are typically not shelf stable for extended periods of time, and it has been suggested in the literature that developing solutions be prepared frequently, especially if highly volatile components are used [21]. With use (i.e., opening a saturated chamber, or developing multiple plates), the composition of the mobile phase can change, which, in turn, can impact the repeatability of plate development [15]. It has been suggested that any developing solution composed of more than a single component should be discarded after a single use [15]; however, this becomes costly and time consuming when the analysis of multiple plates is required. One way to evaluate the stability of an eluent is to monitor the developing characteristics of the plates by means of a standard dye mixture [17]. This practice is common in drug chemistry laboratories whereby plates are evaluated against an archived reference. To evaluate the potential impact of eluent age on plate repeatability, test dye data was examined from 39 plates for which the age of the eluent was known, both in terms of the time since eluent preparation, and the number of plates run since the eluent was changed. Figure 5a shows the average Rf for each of the six bands in the test dye, relative to the number of days since the eluent was prepared. Figure 5b shows analogous data for test dye bands relative to the number of plates run since the eluent was changed. The dashed lines in Figure 5a represent a range of one standard deviation from the mean Rf value for each band from plates developed the same day that a fresh batch of eluent was prepared. The dashed lines in Figure 5b represent a range of one standard deviation from the mean Rf value for each test dye band from the first developed plate in a fresh eluent. Examination of Figure 5a shows that on days one through three, after the preparation of a new eluent, there is a decrease in the average value of bands 3, 4, 5 and 6. However, the average Rf for bands 1 and 2 suggest an increasing trend. Therefore, no consistent trend in Rf could be identified as a function of eluent age. As expected, some variation occurs outside of the calculated one standard deviation range; however, the largest outliers are from bands 3 and 4, which are weak and diffuse, respectively, making them more prone to measurement errors. Ultimately, all variation observed over the eluent age of six days fell within 2 standard deviations, and most (77 %) fell within 1 standard deviation.
Similarly, Figure 5b provides insight into possible exhaustion of the eluent due to overuse (e.g., changes in composition). There are no consistent trends in the average band positions. While some bands show variation exceeding one standard deviation from the fresh eluent average, particularly bands 3 and 4 (which are weak and diffuse, respectively), all variation is again within 2 standard deviations of the fresh eluent average, and most (~80 %) is within 1 standard deviation, even after six plates were run in the same eluent. Together, the eluent aging and use data both demonstrate the stability of the parameters used for the analysis of multiple plates over a period of several days. Therefore, as a conservative rule of thumb, the authors proceeded with a guideline that eluent should be replaced after no more than 5 days, which is consistent with the shelf life reported by Wiggins [13], or 5 plate runs.
2.6 Evaluating Reproducibility Application of the factors that were studied up to this point provides the basis for a reproducible approach to the development of TLC plates, which in turn, provides a means by which data obtained from multiple plates can be relied upon to provide Rf values of sufficient validity for database searching. In terms of our goal of identifying the separated dyes, this offers a practical means by which to build a reference dye database. However, to ensure that the results obtained under such conditions are consistent, and to assist with the development of reasonable search tolerances, it was also necessary to develop an understanding of the expected variation in Rf values. Lederer [5] provides some guidance on the topic of Rf value tolerances and suggests that variations in the Rf of a developed band should be within 0.02, when the operating temperature is well controlled and the system is at equilibrium. When this criterion is applied to the dataset collected here, consisting of the test dye separated on 54 plates, 43% of the data would be rejected. Examination of the data reveals that this tolerance is exceeded more often for bands with a low Rf values relative to those with a high Rf values (56% for bands 1 and 2 vs. 12% for bands 5 and 6). This is consistent with the data presented in Figure 5, which indicates a general increase in the standard deviation range (dotted lines) as a function of decreasing Rf. Given the empirical evidence that band position uncertainty is not accurately represented by a fixed value, we set out to establish a more appropriate means of assigning uncertainty by evaluating test dye data. The average, range, and standard deviation for each of the six bands in the test dye mixture are presented in Table 3. A plot of the standard deviation calculated from the Rf data for each band, as a function of Rf value of each band is plotted in Figure 6. With the exception of Band 4, the data show a strong inverse linear correlation (R2=0.995). Band 4 is a comparatively broad band, relative to the other five bands of the test dye, from which it is more difficult to manually extract an Rf value. The inverse relationship between retardation factor and uncertainty provides, for the plate conditions evaluated, a means by which one can ensure the quality of a developed plate and define reproducibility of a measurement.
3. Discussions Variables of plate selection, plate pre-elution, tank saturation, developing distance, and eluent stability, were explored both through a survey of the literature and experimentation to optimize the
experimental conditions and data repeatability. The results from these evaluations provide an objective means by which to evaluate the variation of the operating conditions studied throughout this paper. While the results presented here are directly applicable to this specific set of conditions, various general lessons arose from this study, which can be applied to other systems:
Numerous planar chromatography plate configurations are available (e.g., HPTLC/TLC; size; stationary phase composition; and the addition of visualization aids). Appropriate selection of these factors for a specified analytical goal can impact the resolution of the method, the ability to recover separated analyte, and cost. An evaluation of comparable plates from several suppliers demonstrated the benefits of a specific combination of these factors to the goal of forensic dye separation.
The commercial production of HPTLC plates provides a highly uniform product that can be used without additional stationary phase activation. In fact, in this small study, a laboratory-based, pre-development process contributed undesirable artifacts.
Saturation of the developing chamber for planar chromatography is critical for achieving and maintaining repeatability. Ultimately, the time required to reach equilibrium is dependent on many factors, including chamber volume and mobile phase composition. A two hour equilibrating period was used in this research.
The multi-phase Wiggins #1 eluent [12] has a reported shelf life of one week. This work confirms that the eluent may be stable for up to five days or the development five HPTLC plates (whichever comes first). Ultimately, the stability of an eluent should be monitored through the use of a test dye mixture and defined assessment criteria, such as that discussed in this paper.
The use of a test dye mixture on each plate is a simple and beneficial way to assess the uncertainty of the method. It also provides a standard that permits the quality of inter-plate and archival data to be evaluated.
The distance to which the mobile phase is eluted is tied to developing time and band resolution. For the system evaluated here, a developing distance of 3.75 cm represents a good compromise between band density and analysis time. This distance is also very close to the theoretical optimal developing distance of 4 cm, which represents the maximum resolution for a plate with a 6 µm silica particle size.
Variation in Rf value is dependent on the complexity of the mobile phase. The multicomponent mobile phase used in this work, Wiggins [13] #1, results in an uncertainty that scales with an inverse linear relationship to Rf.
4. Conclusions Planar chromatography remains a low-cost, highly sensitive analytical tool for the isolation, separation, and purification of textile dyestuffs. When performed in the context of a forensic dye analysis, HPTLC has been generally limited to the realm of a comparative analytical technique that requires the simultaneous analysis of known and questioned samples. However, when the
variables discussed in this work are defined and controlled, the results between different plates and even between different laboratories can be reproducibly developed and objectively evaluated. While variation is inevitable in any analytical technique, an understanding of the impact of experimental parameters on the resulting data permits results to be evaluated for quality and compared over time. In forensic dye analysis, this opens the possibility of developing collections of reference data. The results of this research provide practical insight that is specific to a particular HPTLC separation configuration that is optimized for general forensic dye analysis but which may be utilized as a guideline to optimize and evaluate other experimental TLC separation configurations.
Acknowledgements This work was partially supported by the National Institute of Justice (Grant number 2012-DN-BXK042). The authors are grateful for the efforts of Stephanie Friewald, who spotted, developed, and documented the several dozen HPTLC plates for this project.
References [1] M. Grieve, The role of fibers in forensic science examinations, J. For. Sci. 28 (4) (1983) 877–887. [2] E. Groves, S. Palenik, C.S. Palenik, A Generalized Approach to Forensic Dye Identification: Development and Utility of Reference Libraries. Journal of the AOAC International. In press. [3] C. Neumann, R. Ramotowski, T. Genessay, Forensic Examination of ink by high-performance thinlayer chromatography – The United State Secret Service Digital Ink Library. Journal of Chromatography A. 1218 (2011) 2793-2811. [4] E. Stahl, Thin-Layer Chromatography: A Laboratory Handbook. Second Edition. New York: SpringerVerlag, 1969. [5] E. Lederer, M. Lederer, Chromatography. Second Edition. New York: Elsevier. 1957. [6] F. Fadil, W. McSharry, Extraction and TLC separation of food, drug, and cosmetic dyes from tabletcoating formulations, J. Pharm. Sci. 68 (1) (1979) 97–98. [7] L.C. Lee, Feasibility of High Performance Thin Layer Chromatography for the Forensic Analysis of Ballpoint Pen Inks, Problems of Forensic Science. Vol 97 (2014) 14-22. [8] ASTM E2227-13, Standard Guide for Forensic Examination of Non-Reactive Dyes in Textile Fibers by Thin-Layer Chromatography, ASTM International, West Conshohocken, PA, 2013 www.astm.org. [9] A. Zlatkis, R.E. Kaiser. HPTLC, High Performance Thin-layer Chromatography. Amsterdam: Elsevier Scientific Pub., 1977. [10] C.S. Palenik, J.C. Beckert, S.J. Palenik, Microspectrophotometry of Fibers: Advances in Analysis and Interpretation. NIJ Report 2012-DN-BX-K040 (2015). [11] E. Groves, The Discriminating Power of High Performance Thin Layer Chromatography (HPTLC) for Commercial Textile Dyestuffs. Inter/Micro; 2014 Jun 2-4; Chicago, IL. Microscope 62(3) (2014) 103. [12] C.S. Palenik, E. Groves, S. Palenik, Dye Identification in Casework: How Far Can You Go? Inter/Micro; 2015 Jun 6-8; Chicago, IL. Microscope 63(3) (2015) 115. [13] K. Wiggins, Thin Layer Chromatographic Analysis for Fibre Dyes. Forensic Examination of Fibres, Second Edition. London: Taylor & Francis, (1999) 291-310. [14] R.J. Hamilton, S. Hamilton, D. Kealey. Thin Layer Chromatography. ACOL, London, Wiley, 1987. [15] J. Touchstone, Practice of Thin Layer Chromatography. Third Edition. New York: Wiley, 1992. [16] K. Randerath, Thin-Layer Chromatography. Second Edition. New York: Academic Press. 1966.
[17] D. Laing, The standardization of TLC for comparison of fibre dyes, Journal of the Forensic Science Society. Vol. 30, No. 5 (1991) 299-307. [18] J. Dean, Chemical Separation Methods. New York: Can Nostrand Reinhold, 1969. [19] C. Poole, Multidimensionality in Planar Chromatography. Journal of Chromatography A. Vol. 703, No. 1-2 (1995) 573-612. [20] B. Spangenberg, Theoretical Basis of Thin Layer Chromatography (TLC). Quantitative Thin-Layer Chromatography: A Practical Survey. Berlin: Springer-Verlag, 2011. [21] J. Bobbit, Thin-Layer Chromatography. Third Edition. New York: Reinhold, 1963.
Tables Table 1. Specifications of the HPTLC plates examined in this study. Manufacturer Analtech EMD Millipore Machery-Nagel Sorbtech Whatman
Adsorbent Fluor. Thickness Particle size Chemistry indicator 61077 150 µm 8-10 µm Yes 13748 150-200 µm 5-7 µm No Unmodified 811032 200 µm 2-10 µm No Silica 4214056 200 µm 2-10 µm No 05-713-255 200 µm 4.5 µm No
Pre-conc. Zone size 1.5 cm 2.5 cm 2.9 cm 2.8 cm 2.0 cm
Catalog #
Plate size
Plate substrate
10 x 10 cm
Glass
Table 2. Retardation factor (Rf) measurements for the test dye bands developed varying distances. The Rf value of each band was manually measured in triplicate and the average value is provided. Band # Band 6 Band 5 Band 4 Band 3 Band 2 Band 1
Developing Distance (cm) 2.0 2.5 3.75 5.0 0.99 0.99 0.98 0.97 0.90 0.90 0.90 0.89 0.76 0.78 0.78 0.77 0.66 0.67 0.68 0.66 0.55 0.55 0.53 0.53 0.52 0.52 0.50 0.51
Min
Max
Range
0.97 0.89 0.76 0.66 0.53 0.50
0.99 0.90 0.78 0.67 0.55 0.52
0.02 0.01 0.02 0.01 0.02 0.02
Table 3. Average, standard deviation (σ), and range of Rf values for the test dye mixture (n=54). Band #
Avg. Rf
3σ
Band 6 Band 5 Band 4 Band 3 Band 2 Band 1
0.98 0.88 0.68 0.62 0.49 0.46
0.03 0.04 0.12 0.09 0.10 0.11
Range Min Max 0.96 0.99 0.85 0.91 0.60 0.78 0.56 0.68 0.38 0.55 0.34 0.53
Groves, Palenik (2018) – Reproducibility of High-Performance Thin-Layer Chromatography (HPTLC) in Textile Dye Analysis.
Figure captions: Figure 1. Comparison of the five surveyed plates. Left to right: Analtech, Sorbtech, Machery-Nagel, EMD Millipore, and Whatman. The same dye mixture has been developed on each plate under comparable conditions. The labels on the left identify the developed bands by number. Figure 2. Comparison of developed dyes on neat (left) and pre-developed (right) HPTLC plates. Figure 3. Comparison of four HPTLC plates developed over a period of several years. The plates were stored under the same conditions within our laboratory. Approximate age (months): 10, 23, 35, 40 (left to right). Figure 4. Set of four HPTLC plates developed varying distances. Left to right: 2.0, 2.5, 3.75, 5.0 cm. Figure 5. A plot of retention factors measured from 39 separate HPTLC plate runs of a test dye mixture that consists of six bands. The age and number of plates run between eluent changes was monitored. These two plots and show the variation in retention factor as a function of eluent age in terms of (A) time since preparation and (B) plates run since the eluent was changed. The dashed lines represent a range of one standard deviation from the mean for all bands from either day 1 (A) or plate 1 (B). Note that day 1 represents the day that the eluent was changed. Figure 6. Standard deviation as a function of retention using Wiggins # 1 eluent, on an EMD Millipore HPTLC plate developed a distance of 3.75 cm. The linear trendline has been calculated without Band 4 (bright pink data point), which shows a higher deviation due to the diffuse shape of this band.
Groves, Palenik (2018) – Reproducibility of High-Performance Thin-Layer Chromatography (HPTLC) in Textile Dye Analysis.
Highlights:
Forensic HPTLC analysis is not strictly a comparative tool. A test dye mixture can be used to monitor and assess HPTLC data quality. Libraries of HPTLC data of dyes are possible with controlled analysis conditions. Uncertainty across the HPTLC plate is not uniform for multi-component eluents. Libraries of HPTLC data may lead to more investigative exploration of dyed evidence.
6 5
4 3
2 1
Figure 1
Test Dye Mixture
Acid Red 92
Figure 2
Acid Orange 116
Figure 3
Figure 4
B 1.00
1.00
0.90
0.90
Retardation Factor
Retardation Factor
A
0.80 0.70 0.60 0.50
0.70 0.60 0.50
0.40
0.40 0
1
2
3
4
Days Since Eluent Change
Figure 5
0.80
5
6
7
0
1
2
3
4
5
Plates Run Between Eluent Change
6
7
0.12
Standard Deviation (3σ)
0.10
0.08 y = -0.1475x + 0.1737 R² = 0.9951 0.06
0.04
0.02
0.00 0.40
Figure 6
0.50
0.60
0.70 Retardation Factor (Rf)
0.80
0.90
1.00
Graphical abstract