Pharma Applications of THz-Raman Quantitative Analysis

Polymorph Identification

Many active pharmaceutical ingredients (APIs) can be crystallized in multiple forms or polymorphs. This is particularly important in poorly soluble drugs, where various polymorphs can have quite different dissolution characteristics in the gastrointestinal tract, impacting their bioavailability and hence efficacy. In the past, the polymorphism of many drugs was detected fortuitously or through manual time-consuming methods. Today, drug crystal engineering via combinatorial chemistry and high-throughput screening, makes it possible to easily and exhaustively identify stable polymorphic and/or hydrate/anhydrous forms. This means that rapid and reliable identification of polymorphs during the development, manufacturing, and quality assurance process is critical in pharmaceutical manufacturing.

Polymorphs have the same molecular composition but only differ in the way the molecules are packed in a crystal lattice. So traditional optical analysis methods struggle to identify different polymorphs. For example, FTIR, NIR, and conventional Raman analysis are all based on measuring spectral peaks due to intramolecular vibrations that indicate molecular composition. Consequently, polymorph analysis applications have often resorted to other methods such as powder x-ray diffraction (PXRD) or solid-state NMR (ss-NMR), requiring prep of captured samples and complex instrumentation. THz-Raman has been shown to be ideal for this application because it measures low-energy vibrations in samples and these vibrational spectral peaks arise mainly from intermolecular vibrations such as phonon modes and lattice vibrations. The THz-Raman spectrum is thus highly sensitive to even subtle differences in crystal forms. Also, since THz-Raman does not require any special sample preparation, it is compatible with Green Chemistry initiatives to reduce solvent use.

たとえば、Applied Spectroscopy誌の論文[1]では、Coherent（旧Ondax）とBristol-Myers Squibbの科学者が、カルバマゼピン（CBZ）、カフェイン、テオフィリン、アピキサバンなど、多くの共通するAPIの多形と非晶質を定量的に区別する目的で、THz-Raman分析を使用することに成功したことを実証しています。研究チームは、THz-Ramanがいわゆる光学フォノンを検出し、芳香族分子構造では約130 cm^-1（4THz）までの周波数範囲が期待できると説明し、各多形体が格子モードといくつかの低周波（ねじれなど）分子内振動からなる特徴的なTHz Ramanシグネチャーを持っていることを示しました。 

関連する研究で、Coherent [2]の科学者は、インドメタシン、プロブコール、アセトアミノフェンなど、他のいくつかのAPIの多形間の明確なスペクトルの違いを示しました。図1を参照してください。同科学者たちは、THz-Ramanシステムが多形の高速で明確な区別を提供し、化学的同定に使用される高周波ラマン「化学的指紋」領域を完全に補完するとまとめました。

Co-Crystal Formation

Optimizing polymorphs alone is not always sufficient to enable formulation of an API with the requisite thermal stability, dissolution rates and so on. So the pharma industry has developed a range of other approaches to modify these properties. An example is the use of cocrystals to regulate the API’s solubility, stability, and bioavailability. Cocrystals are molecular crystals formed by weak intermolecular interactions such as hydrogen bonding between two different molecular or ionic compounds. They are also of interest in explosives, agrochemicals, and pigments. In pharma, their main use is with low-solubility APIs.

A cocrystal consists of an active pharmaceutical ingredient (APIs) and a so-called coformer. Different cocrystals are often possible with the same API/coformer pairing, based on different stoichiometries. For example, the carbamazepine:4-aminobenzoic acid cocrystal system can exist in 1:1, 2:1, and 4:1 stoichiometric combinations, each with specific dissolution rates and other important physiochemical properties. Cocrystals can be formed by a variety of solid state, slurry, and solution-based methods, and cocrystal preparation remains an active area of research.

Thus, there is a common need from research through QA, screening, and in-process monitoring for a reliable and fast method to quantitatively identify specific cocrystals. Traditional optical methods are focused on spectral peaks based on covalent bonds and can at best show only subtle data differences between cocrystals and mixtures and then between different cocrystals. Powder x-ray diffraction (PXRD) and nuclear magnetic resonance (NMR) spectroscopy can provide the data but need captured and prepared samples and costly instrumentation for off-line measurements. However, since THz-Raman is an optical method that mainly detects intermolecular vibrations and lattice modes it is very well-suited to this analysis challenge as shown by the example in figure 2.

Industrial & Industrial Chemical Research [3]に掲載された論文で、Coherent（旧Ondax）、明治薬科大学（東京都清瀬市）、国立医薬品食品衛生研究所（東京都世田谷区）の研究者は、THz-Raman分光器によって共結晶形成を定量的に監視できることを示しました。具体的には、エタノールを溶媒とする反応晶析法（RCM）によって生成するカルバマゼピンと4-アミノ安息香酸の1:1および2:1共結晶を研究しました。THz-Ramanデータは、液浸プローブにファイバー結合した分光器によりリアルタイムで取得され、多変量曲線分割（MCR）に基づくデータフィッティングが行われました。また、試料の定量分析だけでなく、化学量論が異なる共結晶間の変化を追跡し、反応速度や終点までの時間を決定できることも示して、共結晶化プロセスにおけるリアルタイム有用なプロセス解析ツールとしてのTHz-Ramanの有用性が確認されました。

Crystallinity vs. Amorphous Monitoring

The low frequency Raman spectrum is dominated by sharp peaks due to crystal lattice modes and optical phonons, which means that a THz-Raman analyzer can measure the concentration of a particular crystal in a mixture of phases, i.e., act as a “crystallinity monitor” [4] The gold standard in crystal identification is powder x-ray diffraction (PXRD) which produces data in the form of a pattern of bright spots at specific diffraction angles based on the lattice properties of the material. However, amorphous materials produce a broad halo of diffracted x-rays due to their semi-random, disordered orientation of molecules, which acts as a noise floor and sets the limit of detection (LOD). This means that it is usually not possible to reliably measure crystallinity content with PXRD, when that content falls below about 5%.

しかし、THz-Raman分光法では、低濃度でも結晶成分によるシャープなピークのいくつかが明確に区別できます。ここで、LODは、背景物質によるスペクトル特徴との重複を最小限に抑えた、ターゲットに起因するスペクトル成分を選択することに依存します。図4は、マンニトール中のアセトアミノフェンの希薄混合物の最近の研究からのTHz-Ramanスペクトルを示しています。アセトアミノフェンの信号とマンニトールによる信号の重なりが少ないことから、ハイライトされた周波数領域がフィッティングの対象として選ばれました。材料間の密度差を考慮し、バックグラウンド除去とベースライン補正を行った後、このスペクトルウィンドウで積分した残留シグナルの大きさに基づいて、予測（重量比）濃度を算出しました。

図5は、この単純な計算による5種類の濃度について、計算値と実際の濃度をプロットしたものです。このプロットに基づき、LODは約1%と決定されました。低濃度では、サンプル混合物の局所的な空間的変動により、また分析領域ではスペクトルシグネチャー間の重複が残っているにもかかわらず、高い誤差が生じました。これはPXRDの典型的なLODをはるかに下回り、固体NMR（ss-NMR）と同等です。混合物の均一性またはスペクトルピーク間の分離が優れている材料のさまざまな組み合わせでは、LODがさらに低くなります。

Coherent THz-Raman分光アプリケーションラボの他のシステムからのデータは、プロセスモニタリングのケースでは、より単純な分析でも結晶化度を定量化するのに十分であることを示しています。具体的には、非晶質背景の幅広いボソン特徴に比べ、結晶性ピークは普遍的にシャープであるため、低周波領域の平均スペクトル微分（傾き）の分析だけで結晶性の程度を近似的に知ることができる場合がよくあります。

研究室内とオンラインの両方でリアルタイムに結晶性を測定できることは、製薬/化学産業の多くの分野に広く適用できる利点であることは明らかです。

Multiple Hydrates (Pseudo-Polymorphism)

Pseudopolymorphism refers to crystal phases of the same compound that vary by the amount of solvent material incorporated in the crystal lattice. For example, in the case of aqueous mixtures these take the form of hemi-, mono-, di-, and trihydrate phases, as well as mixtures of these. In another published study [5], scientists at Coherent (formerly Ondax), Meiji Pharmaceutical University (Kiyose, Tokyo) and the Japan National Institute of Health Sciences (Setagaya, Tokyo) demonstrated the ability of THz-Raman spectroscopy to determine the amount of various hydrates in solid samples of magnesium stearate (Mg-St). They noted that previous studies had showed that powder lubrication, densification, and flowability were all different for the monohydrate and dihydrate forms of Mg-St. They also noted that the identification of the magnesium salt, the relative content of stearic acid by gas chromatography, and the water content are collectively not sufficient to uniquely determine the pseudopolymorphism of Mg-St. This uncertainty can be increasingly problematic where global corporations source raw materials from different locations with varying levels of good manufacturing practices.

In this study, 10 samples of Mg-St acquired from different suppliers were analyzed by PXRD as well as thermogravimetry and differential thermal analysis (TGDTA), allowing them to be described by hydrate type(s). Conventional (chemical fingerprint) Raman at high frequencies showed very minor differences between the hydrates. But as shown in figure 6, when each pseudopolymorph of Mg-St was measured using the low-frequency region, there were differences in peaks between 30 and 60 cm⁻¹ and between 100 and 180 cm⁻¹. In particular, a peak in the spectrum of each hydrate was clearly separated in the region between 100 and 180 cm−1, and this enabled discrimination between the monohydrate and the other hydrates even where the Mg-St was a mixture of pseudopolymorphs.

High-Throughput Screening

The above applications highlight how THz-Raman spectroscopy is a convenient and simpler alternative to PXRD and ss-NMR for rapidly determining the average phase of a sample. Neither of those somewhat cumbersome older techniques can perform spatial mapping. In contrast, the ability to make Raman and THz-Raman measurements through an objective lens supports both adjustment (zooming) and scanning of the sampling area, uniquely enabling mapping (imaging) of a sample under a microscope for both composition and structure/phase. This imaging capability also enables mapping of well plates and mapping within wells for high throughput screening (HTS) applications, and Coherent offers a fully automated well plate reader – the TR-WPS – to provide turnkey analysis for most common well plate formats from 6 to 1536 wells [6].

図7は、Coherent応用研究所のTR-WPSの簡単なデモによるフォールスカラー（偽色）データです。具体的には、Excedrin®の錠剤1錠をウェルに入れ、錠剤の直径3.5 mmの領域で27,000点以上のスペクトルデータを9分間で収集しました。主成分分析（PCA）とクラスタリングを使用して、3つの構成化合物（アセトアミノフェン、アスピリン、カフェイン）と境界領域の混合物を特定しました。偏光顕微鏡（PLM）画像は、その領域で同時に撮影され、視覚的に比較するためにラマンマップでオーバーレイされました。対応するスペクトルは、高品質で強力な光度データを明確に示し、TR-WPSによって素材または多形を迅速に特定し識別して、視覚データとスペクトルデータの両方をキャプチャできることを示しています。

^®ExcedrinはGlaxoSmithKlineの登録商標です。