Advertisement

Sharp Absorption Peaks in THz Spectra Valuable for Crystal Quality Evaluation of Middle Molecular Weight Pharmaceuticals

  • Tetsuo Sasaki
  • Tomoaki Sakamoto
  • Makoto Otsuka
Article
  • 82 Downloads

Abstract

Middle molecular weight (MMW) pharmaceuticals (MW 400~4000) are attracting attention for their possible use in new medications. Sharp absorption peaks were observed in MMW pharmaceuticals at low temperatures by measuring with a high-resolution terahertz (THz) spectrometer. As examples, high-resolution THz spectra for amoxicillin trihydrate, atorvastatin calcium trihydrate, probucol, and α,β,γ,δ-tetrakis(1-methylpyridinium-4-yl)porphyrin p-toluenesulfonate (TMPyP) were obtained at 10 K. Typically observed as peaks with full width at half-height (FWHM) values as low as 5.639 GHz at 0.96492 THz in amoxicillin trihydrate and 8.857 GHz at 1.07974 THz for probucol, many sharp peaks of MMW pharmaceuticals could be observed. Such narrow absorption peaks enable evaluation of the crystal quality of MMW pharmaceuticals and afford sensitive detection of impurities.

Keywords

Terahertz spectroscopy High resolution and high accuracy Middle molecular weight pharmaceuticals Sharp absorption peaks 

1 Introduction

Discovery of novel pharmaceuticals by conventional small molecular weight (SMW) chemical substances is increasingly difficult, while the number of diseases targeted for treatment is also increasing. SMW pharmaceuticals generally have low target selectivity, which may be attributed to their molecular size. Large molecular weight (LMW) pharmaceuticals such as antibody pharmaceuticals have attracted much attention as a new molecular modality and represent the majority of new pharmaceutical products [1]. LMW pharmaceuticals have superior bioavailability with few side effects due to their high target specificity and selectivity compared to conventional SMW pharmaceuticals [2]. However, the use of LMW pharmaceuticals is limited because they must be administered by injection. Furthermore, they are rather expensive because they are produced via cell culture.

Middle molecular weight (MMW) pharmaceuticals, with molecular weights of about 400–4000, are of interest for their possible use in new medications [3, 4]. MMW pharmaceuticals have potentially superior bioavailability and few side effects due to their high target selectivity and are less expensive to produce. Supplying MMW pharmaceuticals in crystal form, as is done for conventional SMW pharmaceuticals, is important for quality uniformity during manufacturing and long-term stability during storage [5].

Crystal quality related to lattice defect content and impurity levels must be strictly controlled during the manufacturing process. The lattice defects are thought to be directly linked to solubility in living body and stability during long-term storage; X-ray diffraction (XRD) is normally used for this evaluation. Accurate analysis of crystal quality for MMW pharmaceuticals by XRD is difficult due to the presence of many overlapping reflection lines from the numerous atoms in the unit cell. Low temperatures reduce the linewidths and render the lines distinguishable, but XRD measurement at cryogenic temperatures is cumbersome. On the other hand, transmittance spectroscopy can be easily carried out even at cryogenic temperatures.

The presence of trace amounts of impurities is inevitable and their monitoring and control is very important [6]. Liquid chromatography (LC) is one of typical techniques having high sensitivity that is used for the detection of impurities in pharmaceuticals. However, it is challenging to distinguish similar molecular species, structures, and weights using LC, and it is also not suitable for larger molecules. The sources of impurities in pharmaceutical products include the starting materials, by-products, intermediates, and degradation products that resemble each active pharmaceutical ingredient (API) molecule in the dosage form [7]. Recently, we reported the quantitative detection of impurities at the ppm level by measuring absorption peak frequency shifts [8]. This method does not directly detect impurity molecules, but rather the influence of the impurity on the crystal structure. Thus, it has no association with the molecular species, structure, or weight. We consider this method as a novel and valuable complement to LC.

The accurate detection of trace impurities requires a spectrometer with high frequency and resolution, and samples showing sharp absorption lines. Our gallium phosphide (GaP) CW terahertz (THz) spectrometer [9] meets these requirements, having a frequency accuracy and resolution of 3 and 8 MHz, respectively [10]. High crystal quality with few defects and measurement at low temperature to reduce thermal disturbance are essential to obtain sharp spectra of target materials.

Recently, we found that crystals of a kind of MMW molecular crystal, α-cyclodextrin hexhydrate [11, 12], show sharp absorption lines [8, 13]. MMW crystals seem to provide sharper spectra than smaller crystals, judging from the limited number of spectroscopic measurements that we have obtained to date. The reason for this may be an ambiguity in the crystal structure due to a larger degree of freedom in the molecular arrangement of SMW molecules; crystalline MMW molecules can crystalize only in good crystal quality. Clearly, more research is required to explain differences in THz spectroscopic behavior between SMW and MMW molecules.

Herein, we discuss examples of crystalline MMW pharmaceuticals displaying sharp absorption spectra that are suitable for crystal quality evaluation and low-level impurity detection by a high-frequency accuracy THz spectrometer. We also discuss the use of THz spectroscopy as an analytical tool for next-generation pharmaceuticals.

2 Experimental

2.1 Samples

Amoxicillin trihydrate (C16H19N3O5S·3H2O, MW = 419.45), atorvastatin calcium trihydrate (C66H68CaF2N4O10·3H2O, MW = 1209.41), probucol (C31H48O2S2, MW = 516.84), and α,β,γ,δ-tetrakis(1-methylpyridinium-4-yl)porphyrin p-toluenesulfonate (TMPyP; C72H66N8O12S4, MW = 1363.6) were purchased from Tokyo Chemical Industry. These MMW pharmaceuticals were used without further purification. Amoxicillin is an antibiotic, atorvastatin is a lipid-lowering agent, probucol is an anti-hyperlipidemic drug, and TMPyP is used as a photosensitizer for photodynamic therapy.

PXRD measurements confirmed that the crystal structure of amoxicillin trihydrate was orthorhombic with the space group P212121 [14], and probucol was monoclinic with P21/c [15], by comparison with the Cambridge Crystallographic Data Centre (CCDC) database [16]. Unfortunately, we could not confirm the crystal structures of atorvastatin calcium trihydrate and TMPyP because we could not find their PXRD data references. All samples were diluted with polyethylene powder and pressed into wedge-shaped pellets about 1 mm thick to prevent etalon artifacts in the THz spectra.

2.2 THz Spectrometer

We obtained high-resolution and high-accuracy THz spectra for each material using a GaP continuous-wave THz spectrometer developed by our group [9, 10]. The light source in the spectrometer was a GaP THz-wave signal generator, in which difference frequency generation occurs in the GaP crystal via excitation of the phonon-polariton mode under small-angle non-collinear phase-matching conditions [17, 18]. Two pump beams, both amplified by ytterbium (Yb)-doped fiber amplifiers, were supplied from semiconductor lasers. Two frequencies were monitored, and feedback was controlled by a two-channel frequency meter. The detector was a superconducting transition-edge sensor bolometer cooled by a pulse tube refrigerator to achieve a nonstop spectrometer system. The frequency range was 0.6–6.0 THz, and the optimal resolution and accuracy were 8.0 and 3.0 MHz, respectively [10]. However, we varied the measurement steps from 100 MHz to 1 GHz to reduce measurement time. The system provides high stability, small size, easy operation, and low maintenance. The sample pellet was set in a cryostat cooled by liquid helium or nitrogen, and the temperature was stabilized within ± 0.05 K between 10 and 300 K by precise control of the heater power.

2.3 Fluorine Resin Coating for Hydrated Molecules

To date, we have applied high-resolution THz spectroscopy to more than 50 MMW pharmaceuticals, and results have shown that hydrated samples often exhibit sharp peaks. However, previous measurements did not show sharp absorption peaks in THz spectra, even when hydrated samples were measured at low temperatures. Crystal transformation into anhydrous or amorphous forms was observed due to water molecule desorption resulting from the partial drop in pressure that occurred when the sample was cooled. Thus, sharp absorption peaks could be observed via a low-temperature measurement method [19]. Fluorine resin spin coating is an effective technique for preventing water molecule desorption and maintaining the crystal structure without parasitic THz absorption. Fluorine resin (Fluorosurf®, FG-3050C-8.0L, Fluorotechnology, Japan) was spin coated at 1700 rpm for 20 s on both sides for a thickness of 8 μm. This process had a negligible effect on the absorption measurements [19]. Figure 1 shows THz absorption spectra at 70 K for amoxicillin trihydrate 10 wt% pellets with and without fluorine resin coating after measurements for 4 h at room temperature under a vacuum. A comparison of spectra with and without the coating showed that water molecules in the uncoated sample were desorbed, and the crystal structure was transformed by annealing at 300 K under a vacuum. Sharp absorption peaks were observed only in the coated sample, and they were present even after we increased the temperature to room temperature and cooled the sample again. The high electronegativity of the fluorine atom and high density of the fluorine resin prevented absorption and desorption of water molecules to and from the crystal.
Fig. 1

THz absorption spectra at 70 K of amoxicillin trihydrate 10 wt% pellets with and without fluorine resin coating after measurements for 4 h at room temperature under a vacuum

3 Results and Discussion

Four samples were selected as examples to demonstrate the sharp absorption lines observed in THz spectra at low temperatures. It was difficult to determine the optimal density for each polyethylene pellet for two reasons. First, there is a strong temperature dependence even between spectra acquired at 10 K and those at 70 K. We could not estimate the optimal density from the data at 70 K easily obtained with inexpensive liquid nitrogen cooling. Second, there is a strong frequency dependence for spectral peak intensities. In general, peaks that appear at a higher frequency have higher absorption. When density is tailored for higher frequencies, absorption peaks are too small at lower frequencies. In this case, high transmittance of the sample and long wavelengths of THz waves result in interference fringes in the spectra.

3.1 Amoxicillin Trihydrate

There are 53 atoms per molecule and four molecules per unit cell of amoxicillin trihydrate. There are 633 vibrational modes, including the infrared inactive modes. Figure 2 shows the temperature dependence of THz absorption spectra for the 10 wt% pellet of amoxicillin trihydrate coated with fluorine resin. The measurement temperatures were 10 K for cooling with liquid helium and 70, 140, 210, and 300 K for cooling with liquid nitrogen. Only 10-, 70-, and 300-K spectra are shown here. The frequency range was limited to 0.6–3.65 THz, because the full peak of strong and sharp absorptions was difficult to detect at higher frequencies. The frequency step was 1 GHz with an accuracy better than 100 MHz. At 300 K, there was a distinct absorption peak near 2.2 THz and minor peaks at 1.7, 2.7, 2.9, and 3.6 THz. At 70 K, there were 18 strong absorption peaks at 1.196, 1.362, 1.540, 1.693, 1.754, 1.855, 2.172, 2.288, 2.371, 2.423, 2.475, 2.675, 2.758, 2.854, 2.946, 3.051, 3.254, and 3.398 THz. The peak at 3.398 THz could not be evaluated because of measurement saturation. At 10 K, there were 10 new peaks and 28 total absorption peaks observed at 0.968, 0.987, 1.060, 1.191, 1.355, 1.535, 1.688, 1.751, 1.848, 1.974, 2.035, 2.067, 2.174, 2.265, 2.293, 2.379, 2.430, 2.476, 2.681, 2.763, 2.857, 2.949, 3.055, 3.148, 3.201, 3.256, 3.408, and 3.541 THz. Absorbance at 3.055 and 3.408 THz could not be evaluated because of saturation. All peak frequencies and absorbance observed at 10 and 70 K are listed in Table 1. For a detailed analysis of absorption peaks at the lowest frequency, we applied THz spectroscopy in the range of 0.95–1.00 THz with a 100-MHz step to the 30 wt% diluted pellet at 10 K. Two peaks were observed (Fig. 3). We performed least squares fitting to a Lorentz function to evaluate peak frequencies at 0.96492 and 0.98450 THz and full width at half-height (FWHM) of 5.639 and 6.180 GHz. Fitting by a Lorentz function instead of a Gaussian function assumes that the crystal is nearly ideal and without defects.
Fig. 2

Temperature dependence of THz absorption spectra for the 10 wt% pellet of amoxicillin trihydrate coated with fluorine resin

Table 1

Absorption peak frequencies and absorbance observed at 10 and 70 K in amoxicillin trihydrate

10 K

fp (THz)

Int.

0.968

0.20

0.987

0.30

1.060

0.09

1.191

0.31

1.355

0.73

1.535

1.48

1.688

0.28

1.751

1.12

1.848

0.41

1.974

0.24

2.035

0.17

2.067

0.18

70 K

fp (THz)

Int.

   

1.196

0.36

1.362

0.47

1.540

0.98

1.693

0.35

1.754

0.71

1.855

0.46

   

10 K

fp (THz)

Int.

2.174

0.67

2.265

0.43

2.293

1.18

2.379

1.64

2.430

2.32

2.476

1.42

2.681

2.43

2.763

2.87

2.857

2.05

2.949

1.34

3.055

sat.

3.148

0.39

70 K

fp (THz)

Int.

2.172

0.61

 

2.288

0.94

2.371

1.34

2.423

1.76

2.475

1.16

2.675

1.62

2.758

1.68

2.854

1.62

2.854

1.62

2.946

1.01

3.051

sat.

10 K

fp (THz)

Int.

3.201

0.48

3.256

0.62

3.408

sat.

3.541

0.86

        

70 K

fp (THz)

Int.

 

3.254

0.71

3.398

sat.

         
Fig. 3

THz spectra at 10 K in the range of 0.95–1.00 THz with a 100-MHz step for the 30 wt% pellet of amoxicillin trihydrate coated with fluorine resin. Least squares fitting by Lorentz functions was applied to two peaks

3.2 Atorvastatin Calcium Trihydrate

There are 160 atoms in atorvastatin calcium trihydrate. However, the number of vibrational modes is unknown, because the number of molecules in a unit cell and the crystal structure are unknown. Figure 4 shows the temperature dependence for the THz absorption spectra of the 10 wt% pellet of atorvastatin calcium trihydrate coated with fluorine resin. The measurement temperatures were 10, 70, 140, 210, and 300 K, but only 10-, 70-, and 300-K spectra are shown here. The frequency range was 0.6–5.8 THz, and the frequency step was 1 GHz with an accuracy better than 100 MHz. At 300 K, minor peaks were observed at 2.1, 3.2, 4.8, and 5.6 THz. At 70 K, there were 20 absorption peaks at 0.974, 1.222, 1.315, 1.468, 1.565, 1.821, 2.096, 2.328, 2.512, 2.656, 2.862, 3.189, 3.317, 3.458, 4.116, 4.492, 4.822, 5.084, 5.26, and 5.635 THz. At 10 K, there were 13 new peaks and 33 total absorption peaks observed at 0.985, 1.086, 1.212, 1.243, 1.288, 1.323, 1.474, 1.567, 1.806, 2.088, 2.327, 2.467, 2.519, 2.665, 2.876, 3.210, 3.275, 3.327, 3.437(s), 3.484, 3.611, 3.741, 3.836, 3.914, 4.135, 4.360, 4.510, 4.610, 4.710, 4.842, 5.082, 5.26, and 5.654 THz. For the peak at 3.437 THz, (s) denotes the shoulder of the large neighboring peak at 3.484 THz. All peak frequencies and absorbance observed at 10 and 70 K are listed in Table 2. The peak at 5.26 THz could not be evaluated because of saturation at 10 and 70 K.
Fig. 4

Temperature dependence of THz absorption spectra for the 10 wt% pellet of atorvastatin calcium trihydrate coated with fluorine resin

Table 2

Absorption peak frequencies and absorbance observed at 10 and 70 K in atorvastatin calcium trihydrate

10 K

fp (THz)

Int.

0.985

0.52

1.086

0.27

1.212

0.40

1.243

0.36

1.288

0.30

1.323

0.42

1.474

0.77

1.567

0.89

1.806

0.69

2.088

0.61

2.327

1.04

2.467

0.68

70 K

fp (THz)

Int.

0.974

0.22

 

1.222

0.22

  

1.315

0.18

1.468

0.28

1.565

0.41

1.821

0.26

2.096

0.24

2.328

0.54

 

10 K

fp (THz)

Int.

2.519

1.38

2.665

0.93

2.876

1.38

3.210

1.56

3.275

1.13

3.327

1.19

3.437(s)

1.35

3.484

2.05

3.611

0.89

3.741

0.88

3.836

0.75

3.914

0.72

70 K

fp (THz)

Int.

2.512

0.56

2.656

0.39

2.862

0.65

3.189

0.74

 

3.317

0.64

 

3.458

0.93

    

10 K

fp (THz)

Int.

4.135

1.91

4.36

0.61

4.510

1.81

4.610

0.82

4.710

4.842

1.90

5.082

2.11

5.26

sat.

5.654

1.39

   

70 K

fp (THz)

Int.

4.116

0.99

 

4.492

0.91

  

4.822

1.27

5.084

1.34

5.26

sat.

5.635

1.05

   

3.3 Probucol

There are 83 atoms per molecule and four molecules per unit cell of probucol. There are 993 vibrational modes, including the infrared inactive mode. Figure 5 shows the temperature dependence of THz absorption spectra for the 20 wt% pellet of probucol. The measurement temperatures were 10, 70, 140, 210, and 300 K, but only 10-, 70-, and 300-K spectra are shown here. The frequency range was limited to 0.6–5.1 THz, because full peaks for strong and sharp absorptions above 5.1 THz are difficult to detect at low temperature. The frequency step was 1 GHz with an accuracy better than 100 MHz. At 300 K, minor peaks were observed at 1.5 and 2.2 THz, and the peaks increased monotonically with frequency. At 70 K, there were 13 absorption peaks observed at 1.083, 1.516, 1.609, 1.765, 1.870, 2.008, 2.104, 2.383, 2.500, 3.015, 4.280, 4.662, and 4.904 THz. At 10 K, there were six new peaks and 19 total absorption peaks observed at 1.080, 1.397, 1.539, 1.564, 1.630, 1.768, 1.913, 2.043, 2.125, 2.427, 2.538, 2.919, 3.059, 4.182, 4.273, 4.615, 4.684, 4.841, and 4.909 THz. All peak frequencies and absorbance observed at 10 and 70 K are listed in Table 3. For a detailed analysis of absorption peaks at the lowest frequency, we applied THz spectroscopy in the range of 1.05–1.11 THz with a 200-MHz step to the 50 wt% diluted pellet at 10 K (Fig. 6). We applied least squares fitting to a Lorentz function to evaluate the peak frequencies at 1.07974 THz and FWHM of 8.857 GHz. As noted earlier, the Lorentz function assumes a nearly ideal crystal.
Fig. 5

Temperature dependence of THz absorption spectra for the 20 wt% pellet of probucol

Table 3

Absorption peak frequencies and absorbance observed at 10 and 70 K in probucol

10 K

fp (THz)

Int.

1.080

0.32

1.397

0.17

1.539

0.62

1.564

0.23

1.630

0.34

1.768

0.37

1.913

0.26

2.043

0.58

2.125

0.52

2.427

0.56

2.538

0.27

2.919

0.35

70 K

fp (THz)

Int.

1.083

0.15

 

1.516

0.26

 

1.609

0.22

1.765

0.14

1.870

0.17

2.008

0.32

2.104

0.28

2.383

0.36

2.500

0.24

 

10 K

fp (THz)

Int.

3.059

0.77

4.182

1.06

4.273

1.37

4.615

0.98

4.684

0.89

4.841

0.90

4.909

1.60

     

70 K

fp (THz)

Int.

3.015

0.50

 

4.280

0.93

 

4.662

0.93

 

4.904

1.01

     
Fig. 6

THz spectra at 10 K in the range of 1.05–1.11 THz with a 200-MHz step for the 50 wt% pellet of probucol. Least squares fitting by Lorentz functions was applied

3.4 α,β,γ,δ-Tetrakis(1-methylpyridinium-4-yl)porphyrin p-Toluenesulfonate

There are 162 atoms in α,β,γ,δ-tetrakis(1-methylpyridinium-4-yl)porphyrin p-toluenesulfonate (TMPyP). However, the number of vibrational modes is unknown because the number of molecules per unit cell and the crystal structure are unknown. Figure 7 shows the temperature dependence of THz absorption spectra for the 5 wt% pellet of TMPyP. The measurement temperatures were 10, 70, 140, 210, and 300 K, but only 10-, 70-, and 300-K spectra are shown here. The frequency range was 0.6–6.0 THz, and the frequency step was 1 GHz with an accuracy better than 100 MHz. At 300 K, wide peaks were observed at 1.5, 2.7, 3.7, and 5.3 THz. At 70 K, there were 24 absorption peaks at 0.974, 1.115, 1.309, 1.379, 1.481, 1.625, 1.780, 2.038, 2.072, 2.453, 2.630, 2.724, 2.891, 3.146, 3.414, 3.756, 4.041, 4.240, 4.458, 4.658, 4.937, 5.179, 5.313, and 5.513 THz. At 10 K, there were five new peaks and 29 total absorption peaks observed at 0.994, 1.127, 1.254, 1.332, 1.394, 1.510, 1.598, 1.659, 1.802, 1.870, 2.057, 2.124, 2.242, 2.477, 2.662, 2.748, 2.895, 3.058, 3.167, 3.438, 3.783, 4.077, 4.257, 4.469, 4.695, 4.970, 5.174, 5.312, and 5.541 THz. All peak frequencies and absorbance observed at 10 and 70 K are listed in Table 4. The absorbance at 3.783 THz could not be evaluated because of saturation at 10 K. The absorption peaks of TMPyP were wide, even at 10 K, which may be because of overlapping modes or splitting of modes by defects. It may be possible to resolve these modes by acquiring spectra at much lower temperatures.
Fig. 7

Temperature dependence of THz absorption spectra for the 5 wt% pellet of TMPyP

Table 4

Absorption peak frequencies and absorbance observed at 10 and 70 K in TMPyP

10 K

fp (THz)

Int.

0.994

0.62

1.127

0.33

1.254

0.18

1.332

0.31

1.394

0.26

1.510

0.48

1.598

0.31

1.659

0.36

1.802

0.43

1.870

0.29

2.057

0.71

2.124

0.98

70 K

fp (THz)

Int.

0.974

1.115

0.30

 

1.309

0.30

1.379

0.24

1.481

0.38

 

1.625

0.29

1.780

0.33

 

2.038

0.62

2.072

0.72

10 K

fp (THz)

Int.

2.242

0.50

2.477

0.78

2.662

0.81

2.748

0.46

2.895

0.53

3.058

0.48

3.167

0.55

3.438

1.03

3.783

sat.

4.077

0.66

4.257

0.56

4.469

0.68

70 K

fp (THz)

Int.

 

2.453

0.60

2.630

0.61

2.724

0.46

2.891

0.48

 

3.146

0.51

3.141

0.90

3.756

1.60

4.041

0.61

4.240

0.46

4.458

0.56

10 K

fp (THz)

Int.

4.695

0.71

4.970

0.60

5.174

0.54

5.312

0.69

5.541

0.95

       

70 K

fp (THz)

Int.

4.658

0.57

4.937

0.42

5.179

0.44

5.179

0.58

5.513

0.67

       

3.5 Application of THz Spectroscopy to Evaluate Crystal Quality of MMW

Of the 50 MMW pharmaceuticals we analyzed previously with THz spectroscopy, none exhibited sharp absorption peaks at room temperature. Because absorption peaks broaden and overlap, it is difficult to quantitatively evaluate the spectra. In addition, only some MMW pharmaceuticals show sharp absorption peaks at low temperatures, such as the four samples used in the present study; this is a prerequisite for identification and evaluation. It is thought that sharp absorption peaks of some MMW pharmaceuticals cannot be observed at low temperatures because of low crystal quality. Thus, high crystal quality with low contents of defects is important for obtaining sharp absorption peaks. Given the strong temperature dependence of spectra acquired at 10 and 70 K, measurements at the lowest possible temperature may be preferable.

Defects in organic molecules or crystals can be detected with high sensitivity because the frequency shifts of absorption peaks depend on the concentration of the defect [20, 21, 22]. Recently, we reported high-sensitivity detection at the ppm level for aspartic acid as impurity in asparagine monohydrate crystals via high-accuracy measurements of peak frequency shifts. Absorption peaks derived from asparagine monohydrate crystals could be fit with a Gaussian function; thus, it was assumed that the crystal was defective. The linewidth was estimated at 12 GHz, which contributed to the detection limit of the frequency shift [8]. If a narrower linewidth can be obtained, such as for amoxicillin trihydrate reported herein, it may be possible to detect much lower impurity concentrations.

Nowadays, absorption peaks in a THz spectrum of an SMW molecular crystal can be well assigned to vibrational modes via quantum calculations using density functional theory (DFT) under periodic boundary conditions [23, 24, 25, 26, 27, 28, 29, 30]. However, it is still difficult to assign absorption peaks in MMW spectra to vibrational modes via calculations because there are too many atoms in the crystal. Thus, experimental analysis including database construction is important for future development of new medications, and this study is the first step.

If a sample has high crystalline perfection, it may be possible to obtain a sharp spectrum for higher weight molecules at low temperature using high-resolution spectroscopy. To date, there are few examples of SMW measurements at 10 K; however, our data suggest that sharper absorption peaks can be obtained for MMW than for SMW pharmaceuticals. Thus, we believe that it may be possible to observe sharp absorption peaks in LMW crystals, such as proteins.

4 Conclusion

We applied high-resolution THz spectroscopy to MMW pharmaceuticals and successfully obtained sharp absorption peaks for amoxicillin trihydrate, atorvastatin calcium trihydrate, probucol, and TMPyP at low temperatures. The FWHM was as narrow as 5.639 and 6.180 GHz at 0.96492 and 0.98450 THz, respectively, for amoxicillin trihydrate and 8.857 GHz at 1.07974 THz for probucol. Such narrow peaks are expected to improve the detection of impurities in organic crystals by THz spectroscopy. These results demonstrate the usefulness of high-resolution THz spectroscopy for crystal quality evaluation and low levels of impurity detection of MMW pharmaceuticals.

Notes

Acknowledgements

We thank Dr. Jun-ichi Nishizawa for his important suggestion about high-resolution THz spectroscopy.

Funding Information

This work was partly supported by Grants-in-Aid for Scientific Research (B) (No. 16H03882) from the Japan Society for the Promotion of Science.

References

  1. 1.
    B. Torre, F. Albericio, Molecules, 22, 368 (2017).CrossRefGoogle Scholar
  2. 2.
    P. Declerck, Generics and Biosimilars Initiative Journal, 1, 13 (2012).CrossRefGoogle Scholar
  3. 3.
    T. Kimura, J. Bioanal. Biomed, 9, 263 (2017).Google Scholar
  4. 4.
    I. Fujii, Farumashia, 52, 116 (2016).Google Scholar
  5. 5.
    N. Blagden, M. de Matas, P.T. Gavan, P. York, Advanced Drug Delivery Reviews, 59, 617 (2007).CrossRefGoogle Scholar
  6. 6.
    E. Chelam, Proc. Indian Acad. Sci. (Math. Sci.) 18, 283 (1943)Google Scholar
  7. 7.
    International Conference on Harmonization; “Impurities in New Drug Substances”, (1995).Google Scholar
  8. 8.
    T. Sasaki, T. Sakamoto, M. Otsuka, Analytical Chemistry, 90, 1677 (2018).CrossRefGoogle Scholar
  9. 9.
    T. Sasaki, T. Tanabe, J. Nishizawa, Optics and Photonics Journal, 4, 8 (2014).CrossRefGoogle Scholar
  10. 10.
    T. Sasaki, T. Tanabe, J. Nishizawa, J. Jpn. Soc. Infrared Science & Technology, 26, 74 (2016).Google Scholar
  11. 11.
    S. Tenjarla, P. Puranajoti, R. Kasina, T. Mandal, J. Pharm. Sci., 87, 425 (1998).CrossRefGoogle Scholar
  12. 12.
    H. Akasaka H. T. Endo, H. Nagase, H. Ueda, S. Kobayashi, Chem Pharm Bull, 48, 1986 (2000).Google Scholar
  13. 13.
    T. Sasaki, K. Itatani, T. Sakamoto, J. Nishizawa, Proc. IRMMW-THz 2012, Australia (2012).Google Scholar
  14. 14.
    S. Feng, N. Shan, K. Carpenter, Organic Process Research & Development, 10, 1212 (2006).CrossRefGoogle Scholar
  15. 15.
    J. Gerber, M. Caira, A. Lötter, J. Crystallogr. Spectrosc. Res., 23, 863 (1993).CrossRefGoogle Scholar
  16. 16.
    C. Groom, I. Bruno, M. Lightfoot, S. Ward, Acta Cryst. B72, 171 (2016).Google Scholar
  17. 17.
    T. Tanabe, K. Suto, J. Nishizawa, T. Kimura, K. Saito, J. Appl. Phys. 93, 4610 (2003).CrossRefGoogle Scholar
  18. 18.
    J. Nishizawa, T. Tanabe, K. Suto, Y. Watanabe, T. Sasaki, Y. Oyama, IEEE Photo. Tech. Lett., 18, 2008 (2006).CrossRefGoogle Scholar
  19. 19.
    T. Sasaki, Proc. IRMMW-THz 2017, Mexico (2017).Google Scholar
  20. 20.
    J. Nishizawa, K. Suto, T. Sasaki, T. Tanabe, T. Tanno, Y. Oyama, F. Sato, Proc. Japan Academy B, 82, 353 (2006).CrossRefGoogle Scholar
  21. 21.
    J. Nishizawa, T. Sasaki, K. Suto, M. Ito, T. Yoshida, T. Tanabe, Int. J. Infrared Millimeter Waves 19, 291 (2008).CrossRefGoogle Scholar
  22. 22.
    J. Nishizawa, T. Tanno, T. Yoshida, K. Suto, Chemistry Letters, 36, 134 (2007).CrossRefGoogle Scholar
  23. 23.
    D.Allis, T.Korter, Chem Phys Chem, 7, 2398 (2006).CrossRefGoogle Scholar
  24. 24.
    S.Saito, T.Inerbaev, H.Mizuseki, N.Igarashi,R.Note,Y.Kawazoe, Chem Phys Lett, 423, 439 (2006).CrossRefGoogle Scholar
  25. 25.
    P.Jepsen, S.Clark, Chem Phys Lett, 442, 275 (2007).CrossRefGoogle Scholar
  26. 26.
    R.Williams, E.Heilweil, Chemical Physics, 373, 251 (2010).CrossRefGoogle Scholar
  27. 27.
    O.Kambara, K.Tominaga, J.Nishizawa, T.Sasaki, H.Wang, M.Hayashi,Chem Phys Lett, 498, 86 (2010).CrossRefGoogle Scholar
  28. 28.
    F.Zhang, M.Hayashi, H.Wang, K.Tominaga, O.Kambara, J.Nishizawa, T.Sasaki, J Chem Phys, 140, 174509 (2014).CrossRefGoogle Scholar
  29. 29.
    F. Zhang, H. Wang, K. Tominaga, M. Hayashi, J. Phys. Chem. A, 119, 3008 (2015).CrossRefGoogle Scholar
  30. 30.
    T. Sasaki, O. Kambara, T. Sakamoto, M. Otsuka, J. Nishizawa, Vib. Spectrosc. 85, 91 (2016).CrossRefGoogle Scholar

Copyright information

© Springer Science+Business Media, LLC, part of Springer Nature 2018

Authors and Affiliations

  1. 1.Graduate School of Medical PhotonicsShizuoka UniversityHamamatsuJapan
  2. 2.Department of Electronics and Materials ScienceShizuoka UniversityShizuokaJapan
  3. 3.Research Institute of ElectronicsShizuoka UniversityShizuokaJapan
  4. 4.Division of DrugsNational Institute of Health SciencesKanagawaJapan
  5. 5.Research Institute of Pharmaceutical SciencesMusashino UniversityTokyoJapan

Personalised recommendations