Subscriber access provided by the University of Exeter
Article
Detection of impurities in organic crystals by highaccuracy terahertz absorption spectroscopy Tetsuo Sasaki, Tomoaki Sakamoto, and Makoto Otsuka Anal. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.analchem.7b03220 • Publication Date (Web): 28 Dec 2017 Downloaded from http://pubs.acs.org on December 29, 2017
Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
Analytical Chemistry is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.
Page 1 of 7 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Analytical Chemistry
Detection of impurities in organic crystals by high-accuracy terahertz absorption spectroscopy Tetsuo Sasaki,*,† Tomoaki Sakamoto,‡ and Makoto Otsuka.§ †Research Institute of Electronics, Shizuoka University, Hamamatsu, Shizuoka 432-8011, Japan ‡Division of Drugs, National Institute of Health Sciences, Setagaya, Tokyo 158-8501, Japan §Research Institute of Pharmaceutical Sciences, Faculty of Pharmacy, Musashino University, Nishi-Tokyo, Tokyo 2028585, Japan ABSTRACT: Quantitative detection of impurities in organic crystals was demonstrated by accurately measuring absorption frequencies using a continuous wave gallium phosphide terahertz spectrometer. THz spectra of L-asparagine monohydrate doped with L-aspartic acid at 0.05 to 12.5 wt% were obtained at 10 K. The three lowest frequency absorption peaks were baseline-resolved, allowing them to be examined independently. Using a least squares curve fitting technique, impurities were detected at levels as low as 500 ppm. The sensitivity and detection limits of the technique depended strongly on the nature of both the host and the impurities. The projected limit of detection using the current system, given optimal materials, was estimated to be 51.7 ppm. In addition to quantitative assessments, impurities may also be identified by comparing frequency shifts of multiple absorptions.
Terahertz (THz) spectroscopy is an important tool for the evaluation of various materials, since unique molecular vibrations that allow one to identify materials and/or determine crystal structures are found in the THz range.1,2 As with infrared (IR) spectroscopy, quantitative analyses are possible in the THz region by comparing the relative intensity of absorption peaks. Generally, such analyses make use of a standard curve that correlates material content with absorption peak intensities or integrated peak area. Therefore, the limit of detection (LOD) depends strongly on the signal-to-noise (S/N) ratio of the spectroscopic measurement and the degree of peak separation. Pharmaceuticals, for example, contain several ingredients, resulting in spectra with multiple peaks. For accurate quantitation, these peaks must not overlap. Previous reports have shown that defects in organic crystals can be detected as slight frequency shifts in the absorption lines of THz spectra, and that the magnitudes of these shifts are dependent on defect concentration.3-5 Thus, both sensitivity and detection limit depend on accurate frequency measurements. Continuous wave (CW) THz spectrometer has merits on high frequency accuracy and resolution compared with pulse type spectrometers. Photomixers pumped by CW near-infrared laser beams were utilized as frequency tunable THz light sources for such spectrometers6-10. Although Gallium Arsenide (GaAs)6-9 or Indium Phosphide (InP)10 based semiconductor devices were utilized as the photomixers in most of the case, Gallium Phos-
phide (GaP) crystal can cover wide THz frequency range mainly because it has low absorption coefficient for THz wave11. The recent development of our highly accurate, highresolution CW GaP THz spectrometer has made stable, longterm measurements possible. 12, 13 This report describes the use of a high-accuracy CW GaP THz spectrometer to analyze L-asparagine (L-Asn; C4H8N2O3) monohydrate crystal powder that had been doped with Laspartic acid (L-Asp; C4H7NO4) at low temperature (10 K). This system was capable of detecting small impurities in organic crystals, such as those used in pharmaceutical applications.
Experiment A. Sample preparation Crystals of L-Asn monohydrate maintain their crystal form (orthogonal, space group P212121) at L-Asp doping levels below 15 mol%, with L-Asp being substituted for L-Asn. 14, 15 In this study, weight percentage was used instead of mole percentage, since the molecular weights of L-Asn and L-Asp are similar. L-Asn (Sigma-Aldrich, A0884; purity: > 98.0 %) containing L-Asp (TCI, A0546; purity: > 99.0 %) at doping levels from 0 to 12.5 wt% were dissolved in deionized water at 40oC in beakers. The beakers were kept in a clean draft chamber until crystallization and the crystals were dried completely.
ACS Paragon Plus Environment
Analytical Chemistry Whole, needle-shaped crystals were collected and ground into a powder with an agate mortar and pestle. Crystals containing 0, 5.0 and 12.5 wt% L-Asp were analyzed on an X-ray diffractometer (RINT Ultra III, Rigaku, Japan) at room temperature. Only reflection peaks corresponding to L-Asn monohydrate16 (orthorhombic, space group P212121) were observed up to 12.5 wt% L-Asp. The samples were mixed with polyethylene (PE) powder and pressed into 3.0 wt% pellets weighing 300 mg, with a diameter of 20 mm and an approximate thickness of 1 mm. B. THz absorption measurements Figure 1 shows a schematic of the CW GaP THz spectrometer system. The light source is based on difference frequency generation (DFG) in the GaP crystal. Two IR beams were delivered from a distributed feedback (DFB; Toptica DLDFB) laser and a mode hopping-free, external-cavity diode laser (ECDL; Spectra Quest Lab. λ-master 1040). Each beam was amplified to a constant power of 5 W at any wavelength using polarization-maintained ytterbium (Yb)-doped optical fiber amplifiers (YbFAs; Keopsys CYFA-PB series). The wavelength of the DFB laser was from 1,071.7 to 1,074.4 nm (279.0–279.7 THz), while that of the ECDL was continuously tunable from 1,000 to 1,080 nm (277.6–299.8 THz), but limited to between 1,050 and 1,075 nm (278.9–285.5 THz) in our system by the available range of the YbFAs. Wavelengths were monitored by a two-channel frequency meter (HighFinesse WS-7) and laser power was maintained using thermal power meters and feedback control at each frequency to establish simultaneous high frequency accuracy and power stability. The two beams were combined in a GaP crystal with a small phase-matching angle at each frequency using a lens and a linear stage. The GaP crystal was tilted on a rotary stage to prevent total reflection of generated THz waves at the back surface. THz output beams were collected by parabolic mirrors and focused on a niobium transition edge superconductor (TES) bolometer cooled by a low vibration pulse tube cooler (QMC instruments QNbB/PTC) after passing through the sample pellet set in a cryostat (Oxford instruments Microstat He2) cooled by liquid helium/nitrogen. The sample temperature could be controlled between 5 and 400 K with better than ±0.05 K stability. The entire THz-wave path was purged with dry air to a dew point less than 213 K, to avoid the effects of water vapor absorption. Lock-in detection of the THz wave with a chopping frequency of approximately 750 Hz afforded high sensitivity and long-term signal stability. The resulting spectra were highly reproducible, obviating the need to use a double beam method to eliminate the effects of beam power fluctuations. Although the highest absolute frequency accuracy and resolution were obtained at 3.0 and 8.0 MHz7, respectively, such conditions would require too much time per spectrum. In this study, a measurement frequency of 100 MHz was chosen as a compromise between spectral acquisition time and accuracy and resolution. Measurement sequences were controlled using software developed in our laboratory with a personal computer. The resulting spectrometer boasts a wide frequency range (0.6–6.0 THz) in a liquid helium-free, non-stop system.
TES Bolometer
DFB : Distributed-Feedback laser ECDL: mode-hop free External Cavity Diode Laser YbFA: Yb doped Fiber Amplifier Chopper DFB
YbFA
ECDL
Dry air purged
Linear Stage
Vacuum Chamber Lenz
Collimater Beam Splitter
GaP Rotary Stage
Power Meter
YbFA Collimater Power feedback
Frequency feedback
Power Meter
Controller 2 channel Frequency Meter
Stage Controller
Controller
Lock-in Amp.
Personal Computer
Figure 1. Schematic of the CW GaP THz spectrometer. Results and Discussions Figure 2 shows a THz absorption spectrum of neat, recrystallized L-Asn monohydrate at 10 K. Although several sharp absorption peaks can be observed, most of those beyond 2.7 THz were overlapping and not independently distinguishable. However, the absorption peaks around 1.71 (labeled as Peak 1), 2.44 (Peak 2) and 2.61 (Peak 3) THz were baselineresolved, allowing accurate measurement of frequency shifts. Therefore, this paper focuses on these three absorption peaks. The temperature dependence of each peak was -121 MHz/K for Peak 1, -244 MHz/K for Peak 2, and -168 MHz/K for Peak 3. Since the sample temperature stability was better than ±0.05 K, the maximum peak frequency deviations due to temperature fluctuations were 12.1, 24.4 and 16.8 MHz, respectively. Wavenumber (cm -1) 50 100 150
0 1.5
Peak1
Absorbance
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Page 2 of 7
Peak3 Peak2
1
PE
0.5
0
1
2 3 4 Frequency (THz)
5
Figure 2. THz spectrum of L-Asn monohydrate in a 3.0 wt% polyethylene (PE) pellet at 10 K. Figure 3 shows enlarged views of Peak 1 (a), Peak 2 (b) and Peak 3 (c) in the THz absorption spectrum of pure L-Asn monohydrate, measured at 10K in 200-MHz steps. Acquisition times for each spectrum were about 40minutes, depending on measurement points of 1,000. All three peaks exhibited an asymmetric tail on the low-frequency side. This may be due to
ACS Paragon Plus Environment
Page 3 of 7 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Analytical Chemistry the presence of multiple low-frequency peaks, arising from initial defects in the host crystal, overlapping with the primary peak. Although further investigations are needed to reveal the mechanisms underlying the impurities in the host crystal, the above hypothesis is consistent with the results of this paper. Commercially available L-Asn initially contains a maximum of 2% impurities. This is one possible reason for the observed peak tailing and broadening of absorption peaks. We applied a curve-fitting algorithm to each absorption peak by employing a least squares method with two or three Gaussian curves. There may, in fact, be several kinds of defects, such that one or two Gaussian profiles are not sufficient for a good fit. Peak 2, for example, contains an absorption due to the PE matrix at around 2.35 THz, necessitating an additional peak under the fitted curve. Peak 3 incorporates an additional high-frequency absorption (possibly a component of Peak 4), requiring three Gaussian curves for an acceptable fit.
Table 1. estimated
IP11
1.645 +0.011 - 0.014
0.102 +0.015 - 0.012
fP21
IP21
2.3752 0.1573 +0.0016 +0.0050 - 0.0028 - 0.0057
fP31
./0 (1) = ∑6 4506 789 :−
(
