THE MORPHOLOGICAL ANALYSIS OF CRYSTALLINE METHADONE: A NOVEL COMBINATION OF MICROSCOPY TECHNIQUES

MDN crystal, mean thickness at (428 nm) mean retardation value. Conclusion : The SEM was superior and successfully determined MDN crystal dimensions for the first time, whilst the CM results were affected by the Rhodamine dye staining process used for visualisation. The qualitative analysis of the crystallinity status of methadone hydrochloride optimally achieved using a combination of PLM and SEM techniques


Introduction
Methadone hydrochloride (MDN) is a white crystalline powder, which is soluble in water (12 g/100 mL) and alcohol (8 g/100 mL), and partially insoluble in glycerol and ether [1,2]. It is an effective pharmacological substitution treatment for opioid dependence. MDN has been adopted in different countries as a maintenance treatment (MMT) for heroin addiction [3][4][5]. However, MDN can be diverted from legal sources and misused [6][7][8]. Tablets may be pulverized to powder to be dissolved in water or alcohol for intravenous use or by insufflation [9,10].
Since MDN is a crystalline opioid [1,2], microcrystalline tests could have a potential forensic benefit for characterizing illicit sources of the drug. Recommendations to use such tests in characterizing crystalline abused drugs have been advocated for opioids [11,12]. However, the microcrystalline test provides insufficient data about basic crystal structure of powdered drugs available on the black market [13]. This study investigated whether advanced microscopic techniques could address these issues.
We proposed further work using 3-dimensional examination of MDN crystals with SEM and CM because they have been widely used in pharmaceutical sciences to analyse the morphology of crystalline and controlled drugs [16][17][18]. Esmaeili, and Khodaei (2019) successfully used SEM and CM in characterizing targeted magnetic nano capsules, which were loaded with MDN and rifampin [19].
To our knowledge the dimensional measurements of MDN crystals, the relationship between the birefringence colours, MDN crystals thickness and the light retardation have not been investigated previously.
The aim of this study was to evaluate the combined microscopic techniques in identifying the morphological and optical properties of MDN crystals for use in forensic investigations to identify unknown powders in forensic cases.

2. 1. Optimization of MDN re-crystallisation
To optimize MDN re-crystallisation, MDN powder was mixed with HPLC grade water. Room temperature and humidity were recorded using USB datalogger. Three different techniques were explored: 1. Weighed MDN method: glass microscope slides were prepared in a ratio of 1:4 (MDN:water). MDN (2.5 mg) was spread onto the slide with 10 µL of HPLC water.
2. Hot stage method: MDN solution (10 µL of HPLC water and MDN 2.5 mg) was covered with a coverslip and sealed to prevent evaporation. The slide was heated to 50 °C (rate of 1 °C/min) with a holding time of 10 mins at this temperature. Undissolved particles of the super saturated solution of MDN was used for the crystallisation process. The heating and cooling cycles were observed using the PLM (Dialux 22EB, Leitz) and the Qi Imaging camera with a magnification of X25.
3. Closed container method: MDN solution (50 mg MDN mixed 200 µL of HPLC water at 50 °C) in a closed vial (n=2), vortexed for 10s and cooled down for 2 hrs to 3.5 °C. MDN crystal formation from each vial (10 µL on a microscope slide) was observed by PLM.

2. 2. Polarized light microscope (PLM) observation method
The PLM was used to study the optical properties of MDN particles and crystals and to determine retardation colours. The prepared slides of MDN-water solutions were observed using the PLM (Dialux 22EB, Leitz) and the Qi Imaging camera with a magnification of X25. For polarized observations a cross-polarizing filter and λ (Lambda) plate red 1 compensator were used to enhance the contrast images quality [20]. The cross polarizer was rotated until the extinction position was achieved (black background obtained) to make a contrast between the sample crystals and the slide background [20,21]. To investigate the effect of the cross-polarizer angle setting on the retardation colour of MDN particles, 0.25 mg of MDN powder was spread on a microscope slide and observed by the PLM. For standardization, delineation of the central area (25×20 mm) of the bottom side of the slide was carried out using a permanent marker. Then, this area was dynamically observed (from the top side where the methadone hydrochloride powder is distributed) with a magnification of ×40.

2. Characterizing MDN particles
The optical microscope (Nikon Labophoto) with a Panasonic camera (WVCL310) and particle size analysis software (V1999) were used to observe 3 slides of MDN (magnification of X40) to measure the particle size of 500 MDN particles by measuring the volume diameter. The volume diameter was defined as the volume of a sphere having the same volume as the particle, and it was calculated according to equation (1): where dv -volume diameter, and v -volume of the particle. An anisotropic material such as MDN has more than one refractive index (RI or n) depending upon the lattice direction; the difference between the RIs (∆n) of the material is known as birefringence [21-23].

2. 4. Scanning electron microscope observations
The SEM was previously used to characterize crystal morphology at room temperature [17,24]. MDN Crystals were prepared using the closed container method for SEM observations. PLM observations were carried out using a magnification of X25 before gold coating to correlate MDN crystals colours with the thickness measurements. Each slide was metallised with gold (15 nm thickness) using a method described by [25]. Totally, 42 crystals were observed (Six crystals for 7 slides) using SEM at an accelerating voltage of 20 kV [26]. The crystal dimension measurements were carried out in duplicate using Quartz PC software using an aerial image to determine length and width measurements, and a ver-tical view (90° tilting) to measure the thickness of each crystal.

2. 5. Confocal microscope observations
The confocal microscope generated a 3D image by scanning the optical sections of the sample using Z-stack scan, which is a digital image processing technique providing three-dimensional views [18, 27,28]. The Z-stack scanning technique was used to determine if the CM could measure crystal dimensions including its thickness [28,29].
CM samples were prepared as follows: MDN solution (closed container method) required 0.0001 % g/mL warm rhodamine (Rh) solution (50 °C) as a crystallising solvent for the confocal microscopy. The Rh solution (0.0001 % gm/mL) was prepared by mixing 0.1 g of rhodamine dye with 100 mL of HPLC water. Then 1 mL of Rh solution (0.1 % g/mL) was diluted to a 0.0001 % g/mL solution. Seven slides were prepared, and 6 crystals were observed from each slide (n=42). The confocal images used a magnification of ×40 at excitation level 543 nm.

Statistical analysis
MDN crystals (n=42) were analysed to determine the crystal length, width, and thickness using SPSS 24 software. Independent samples t-test and Mann-Whitney U test were used to carry out the inferential statistics between the SEM and CM measurements. The mean values of MDN crystals thickness were measured and used in the following equation: Retardation (r)=thickness (t)×birefringence. (2) As the thickness was measured in µm, the retardation was usually measured in nm, and the birefringence represented the difference between the biggest and smallest refractive indices (n2 and n1 respectively). The equation that was used to calculate the retardation was [21]: After obtaining the mean values of retardation, the prediction of the birefringence colours range of MDN crystals was obtained using Michel-Levy Birefringence Chart.

Results
Satisfactory MDN crystallisation was recorded for the closed container method: single diamond shaped MDN crystals were formed with the lowest MDN crystalline aggregation as compared with the other methods. During the MDN re-crystallisation process, the relative humidity (% RH) was found to impact on MDN re-crystallisation. Although between 24 % and 70 % RH, satisfactory MDN crystal formation was achieved using the closed container-method. However, when the RH>70 % no MDN crystal were produced. Room temperature changes did not alter the MDN re-crystallising or characterization process. The polarized and non-polarized photomicrographs of re-crystallised MDN using the closed container method showed better consistency of shape and number of crystals formed. The crystals were diamond in shape and exhibited yellow retardation colours (Table 1). Therefore, the closed container method was shown to outperform other methods and, hence, was used in all further experimentations. MDN crystal preparations yielded a diamond shape as demonstrated in the orthographic projection ( Fig. 1), which shows an intrinsic angle of 62°.

Characterization of MDN crystals using SEM and CM.
The dimensions of MDN crystals determined by using SEM and CM. SEM images of MDN crystals at 90º tilting of the slide enabled the determination of the thickness of the crystals in µm (Fig. 2). Six single crystals separated per slide were eligible for 3D imaging using the SEM.
The retardation range for MDN crystal thickness (measured by SEM) was 195 -839 nm. The mean SEM thickness of MDN crystals was 6.62 µm, yielding a retardation value at 428 nm (Fig. 4). According to the Michel-Levy birefringence colour chart, therefore most MDN crystals should demonstrate a yellow colour (at the extinction position) as seen in the PLM photomicrographs in Table 1. The cross polarizer setting effects on the retardation colours of MDN particles were studied and the photomicrographs of MDN particles at different cross polarizer positions in relation to the extinction position could be seen in Table 2. In addition to the colour change observations, some of MDN particles were no longer visible at the extinction position, while they can be seen at other positions of the cross polarizer (Table 2).

Discussion
This study established the optimal conditions and methods needed to prepare and characterize MDN crystals using different microscopic techniques. MDN crystal thickness, birefringence colour and retardation relationship were established for the first-time using Michel-Levy birefringence colour chart.
In accordance with Hubach and Jones original work, water was used as an inexpensive, safe, easy to use solvent. MDN recrystallisation was optimized by slow heating and cooling the supersaturated MDN solution in a closed container. An important finding was the impact of humidity >70 % on MDN recrystallisation. Mullin (2001) has noted that nucleation and crystal growth rate can be affected in this way [30].
The PLM continues to be the best tool for characterizing anisotropic materials and crystalline drugs [14, 20,22] such as methadone. We confirmed the formation of the typical diamond shaped MDN crystal with 62° angle value. Different retardation colours were shown with the PLM such as purple, blue and yellow. This was attributed to the effect of the anisotropic material (MDN) on the polarized light as when white light passed through the polarizer, the MDN travelling at different velocities (acting as a beam splitter), splits the plane polarized light into two parts, each one vibrating in two separate planes perpendicular to each other (ordinary and extraordinary rays). One ray is retarded in respect to the other (this optical condition is known as retardation). When the ordinary and extraordinary rays pass through the analyzer, they interfere or recombine causing the specimen to appear coloured [31]. Thus, optical interference at the analyzer plane, which destroys some wavelengths and reinforces others [15,20,21], is behind the appearance of different colours of MDN crystals and particles. represents the volume diameter of MDN particles, and the orange series represents the cumulative percentages. More than fifty percent of the particles had volume diameter equal or less than 28.7 µm Continuation of Table 2   1  2  3 After 270° clockwise rotation At 270° interference colours changed as compared with the interference colours of MDN particles at the extinction position; however, the same interference colours of MDN particles noticed at 90°of the clockwise rotation were recorded After 360° clockwise rotation After 360° clockwise rotation, the same interference colours were obtained at the extinction position as previously observed In addition, the appearance of different MDN crystal and particle colours may be attributed to the difference in MDN sample thickness, as sample thickness and retardation values affect the appearance of interference colours [21,32]. Differentiation of the interference colours of MDN crystals was achieved by identifying the range of MDN crystals thickness, and correlating the retardation, thickness, and birefringence of anisotropic crystals (equation (2)). This outcome, coupled with the aid of Michel-Levy birefringence colour chart, can help to enrich the pharmaceutical and forensic database about MDN crystals characterization.
It was also observed that the analyzer-polarizer angle affected the retardation colours of MDN particles. This was due to the creation of different wavelength interferences of the polarized light that passed through the anisotropic material (MDN) at the analyzer level, producing different interference colours at 90°, 180°, and 270° as compared to the interference colours at the extinction position. Therefore, MDN particle colours change if the angle between the polarizer and the analyzer was altered. Hence, to standardize the diagnosis of MDN for forensic purposes, MDN particles and crystals should be observed at the extinction position when the cross polarizers are perpendicular to each other.
Some MDN particles that did not exhibit interference colours (or appeared black) at the extinction site can be attributed to the fact that the direction of one of the principal refractive indices was parallel to the vibration direction of the polarizer. This led to cancellation of the second vector component; therefore, the second vector equalled zero. In this case, the emerging light from the material had the same vibration direction as the polarizer, i.e., the emerging light direction was perpendicular to the analyzer direction. Thus, the analyzer absorbed the emerging light and prohibited its transformation on to the microscope objectives leading to the appearance of black spots [21,33].
The rotary stage of the SEM enabled the determination of the actual dimensions of MDN crystals in a standard manner using the tilting procedure at 90°. A MDN crystal was confirmed as diamond shaped with an intrinsic angle of 62°, mean length 46.4±15.2 µm, width 28.03±8.2 µm and thickness of 6.62±2.9 µm. However, the mandatory gold coating step meant that this diagnostic test is expensive and requires trained personnel in addition to the difficulties of the tilting procedure to 90°, encouraged the trial of an alternative technique.
The CM using Rhodamine B dye to characterize MDN could determine dimensions using horizontal (x, y) imaging over a wide area of a crystal, and Z-stacks to determine sample thickness [34]. MDN re-crystallisation with warm super saturated rhodamine solution helped to overcome the disadvantages mentioned by Singh et al. in 2012 [18] in providing a more uniform staining of MDN crystals, but there was a difficulty in measuring the thickness accurately and significant differences in the dimensions measured using Z-stack scanning. This was attributed to the imaging process requiring emitted light from the stained MDN crystals. The poorer resolution of dimen-sions achieved by the effect of rhodamine B dye accounted for the significant difference between SEM and CM dimensions for MDN crystals. CM was shown to be less reliable in measuring MDN crystals than the SEM.
The demonstration of a correlation between the thickness of a MDN crystal and the retardation value of MDN using the Michel-Levy birefringence chart helped to explain the range of retardation colours exhibited by MDN crystals under the PLM. Moreover, the predominant yellow colour of MDN crystals observed by using the PLM reflected the retardation value of MDN crystals with a mean thickness of 6.62±2.9 µm. This was confirmed by the mathematical calculation thus confirming the colour yellow as a defining characteristic of the MDN crystal.
Besides confirming the colours range of MDN crystals, the Michal-Levy birefringence chart was embodied in the determination of retardation value for MDN crystals. In a previous study, Beaufort et al. (2014) calculated the thickness of calcite crystals using PLM (with a cross polarizer) in relation to the Michal-Levy birefringence colour chart. However, they could not verify the crystal thickness when greater than 4.5 µm [35] and reported this as a limitation for the PLM; a shortcoming confirmed in our study. In contrast, using SEM thickness measurements of MDN crystals with the aid of Michel-Levy birefringence colour chart, MDN crystals retardation colour range was successfully determined to be between 195-839 nm.
The success of determining a procedure to characterize MDN crystals may be encouraging in terms of application to other types of controlled crystalline drugs with the potential usefulness in forensic cases to identify powders of unknown sources, particularly new psychoactive substances (NPS).
Study limitations. The number of measured crystals was limited to 42 crystals due to the technical difficulties and high cost required for measuring MDN crystals thickness using the SEM. In addition, more crystalline addictive drugs should have been studied to establish a library of crystalline addictive drugs that could be used in identifying the unknown powders of these materials.
Prospects for further research. Studying the cooling and time effects on the MDN crystallisation process as they are important parameters that can influence MDN crystals formation.
1. Studying the efficacy of these techniques in characterizing methadone hydrochloride crystals that can be prepared from crushed tablets.
2. Establishing the relationship between MDN birefringence colours, crystal thickness and the Michael-Levey birefringence colour chart using the SEM to establish a diagnostic method for detecting the source or the supplier of MDN.

Conclusion
The closed container method was best to re-crystallise MDN at room temperature with relative humidity <70 % RH. This procedure produced recognised diamond shaped crystal, with a diagnostic angle of 62°. Thickness of MDN crystal and the refractive indices were contributing factors to the resultant retardation colour, expected to be yellow.
Due to the ambiguous correlation between the birefringence theory and the properties of MDN crystals, a combined use of microscopic techniques was developed and correlated for the first time using the SEM that the mean values for a single MDN crystal length (46.4±15.2 µm), width (28.03±8.2 µm) and thickness (6.62±2.9 µm).

Conflict of interest
The authors declare that they have no conflict of interest in relation to this research, whether financial, personal, authorship or otherwise, that could affect the research and its results presented in this paper.

Financing
This work was funded by the Ministry of Higher Education and Scientific Research of Iraq through an educational scholarship to Noor R. T. Al-Hasani.