A new method for classification of salts and cocrystals using solid-state UV spectrophotometry

Article abstract of DOI:10.1248/cpb.c18-00743

Salt and cocrystal formulations are widely used as techniques to improve physicochemical properties of compounds. Some spectrometric techniques to distinguish cocrystals from salts have been reported; however, it has not been possible to adapt these formu

Full text of DOI:10.1248/cpb.c18-00743

Vol. 67, No. 9
Chem. Pharm. Bull. 67, 945–952 (2019)
945
Current Topics
Front-Line Preformulation Studies
Regular Article
A New Method for Classification of Salts and Cocrystals
Using Solid-State UV Spectrophotometry
,
a
b
c
d
Akihito Kiguchiya,* Reiko Teraoka, Toshiyasu Sakane, and Etsuo Yonemochi
a
Technology Research & Development Division, Sumitomo Dainippon Pharma Co., Ltd.;
b
3
–1–98 Kasugade-naka, Konohana-ku, Osaka 554–0022, Japan: Department of Pharmaceutical
Technology, College of Pharmaceutical Sciences, Himeji Dokkyo University; 7–2–1 Kamiono,
c
Himeji, Hyogo 670–8524, Japan: Department of Pharmaceutical Technology, Kobe Pharmaceutical
d
University; 4–19–1 Motoyamakita-machi, Higashinada-ku, Kobe 658–8558, Japan: and Department
of Physical Chemistry School of Pharmacy and Pharmaceutical Science Hoshi University;
2
–4–41 Ebara, Shinagawa-ku, Tokyo 142–8501, Japan.
Received September 25, 2018; accepted April 1, 2019
Salt and cocrystal formulations are widely used as techniques to improve physicochemical properties of
compounds. Some spectrometric techniques to distinguish cocrystals from salts have been reported; however,
it has not been possible to adapt these formulations with many compounds, because of limitations, high dif-
ficulty, and exceptions. Therefore, we focused on the possibility of UV spectrometry, which had not been
reported as a classification technique for salts and cocrystals. The integration values of solid-state UV/visible
(
Vis) spectra of indomethacin salts were larger than those of physical mixtures of indomethacin and counter
molecules, while that of indomethacin cocrystal was not large compared with that of the physical mixture.
From these results, differences between a salt and a cocrystal were observed in their solid-state UV/Vis ab-
sorption spectra for indomethacin complexes. Therefore, it is suggested that solid-state UV/Vis absorption
spectra can be used as a new technique to classify salts and cocrystals.
Key wordsꢀ classification;ꢀsalt;ꢀcocrystal;ꢀsolid-stateꢀUVꢀspectrum;ꢀindomethacin
Introduction
paper was issued by the European Medicines Agency (EMA)
2
3)
It is often necessary to improve physicochemical properties in 2014. Thus, an environment for development of cocrystals
of a compound, such as the solubility, since the number of as an active pharmaceutical ingredient (API) is being pre-
drugꢀcandidatesꢀcategorizedꢀinꢀtheꢀBiopharmaceuticsꢀClassifi- pared.ꢀInꢀaccordanceꢀwithꢀtheꢀFDAꢀguidanceꢀandꢀEMAꢀreflec-
cation System (BCS) as class 2 and class 4 has been increas- tion paper, it is necessary to ensure that an API is a cocrystal.
1)
ing recently. Development of a salt formulation is one of the Inꢀ particular,ꢀ theꢀ FDAꢀ guidanceꢀ requiresꢀ thatꢀ cocrystalsꢀ beꢀ
most popular techniques to improve physicochemical proper- discernedꢀ fromꢀ salts.ꢀ Cocrystalsꢀ areꢀ definedꢀ asꢀ crystallineꢀ
ties of a compound, and many drugs having a salt formula- materials composed of two or more molecules within the
2
–7)
24–26)
tion as drug substance are approved.
However, salt forms same crystal lattice, as is the case with salts.
However,
are not adopted to neutral compounds, since they require that the interaction of components is different between salts and
8
)
27,28)
the compound has some dissociative group. In addition, the cocrystals.
In salts, the components are arranged in the
8)
number of counter molecules that can be used to form salts is crystal lattice predominantly based on ion pairing, while in
small, because their safety information and approval records cocrystals, the components are assembled via weak interac-
9)
as salts are considered in use as pharmaceutical salts. Thus, tions such as hydrogen bond, π–π stacking, or van der Waals
18–21)
the adaptation of salts to compounds has been limited.
interaction.
Therefore, it is possible to distinguish salts
Recently, cocrystal formulation was reported as a technique from cocrystals using techniques that obtain the differences.
to improve physicochemical properties such as solubility, sta-
Some kinds of techniques to classify salts and cocrystals
10–17)
bility, and bioavailability, as with salt form.
Cocrystals have been reported, for example the pK ꢀdifference,ꢀΔpK (pK
a
a
a
are advantageous in that they do not require that compounds (base)-pK (acid)). Generally speaking, a compound for which
a
have dissociative group, and there are many candidates for ΔpK >3ꢀ isꢀ aꢀ saltꢀ andꢀ oneꢀ forꢀ whichꢀ ΔpK <1 is a cocrys-
a
a
18–21)
29–31)
use as cocrystal formers considered as safe.
Therefore, tal.
ꢀFurthermore,ꢀtheꢀoccurrenceꢀofꢀprotonꢀtransfer,ꢀwhichꢀ
3
2,33)
cocrystals can be used widely, and have recently become one meansꢀ salt,ꢀ isꢀ avoidableꢀ underꢀ lowꢀ ΔpKa.
of the most noteworthy techniques to improve physicochemi- very useful, and need not use any sample. However, with this
cal properties. techniqueꢀ itꢀ isꢀ difficultꢀ toꢀ ensureꢀ theꢀ salt/cocrystalꢀ natureꢀ ofꢀ
Theꢀ U.S.ꢀ Foodꢀ andꢀ Drugꢀ Administrationꢀ (FDA)ꢀ issuedꢀ someꢀ compoundsꢀ suchꢀ asꢀ neutralꢀ compounds.ꢀ Single-crystalꢀ
This method is
2
2)
a guidance regarding cocrystals in 2013, ꢀ andꢀ aꢀ reflectionꢀ X-rayꢀ diffractionꢀ isꢀ theꢀ mostꢀ effectiveꢀ methodꢀ toꢀ distinguishꢀ
*
ꢀToꢀwhomꢀcorrespondenceꢀshouldꢀbeꢀaddressed.ꢀ e-mail:ꢀakihito-kiguchiya@ds-pharma.co.jp
©
2019 The Pharmaceutical Society of Japan

9
46
Chem. Pharm. Bull.
Vol. 67, No. 9 (2019)
3
4)
salts from cocrystals.
However, this method requires a of 0.015° and time per step of 0.25s.
1
1
singleꢀ crystal,ꢀ whichꢀ isꢀ oftenꢀ difficultꢀ toꢀ obtain.ꢀ Therefore,ꢀ
H-NMR Spectrometry H-NMRꢀ spectraꢀ wereꢀ obtainedꢀ
techniques that do not require a single crystal were studied. with an AV400M NMR system (Bruker) equipped with a
Solid NMR spectrometry was reported as a technique for the 5-mmꢀspinnerꢀatꢀroomꢀtemperature.ꢀApproximatelyꢀ1ꢀtoꢀ5ꢀmgꢀ
classificationꢀofꢀsaltsꢀandꢀcocrystals,ꢀandꢀgivesꢀinformationꢀofꢀ of each indomethacin complex was dissolved to 0.75mL of
35)
theꢀ locationꢀ ofꢀ non-ionicꢀ interaction. However, this method dimethylsulfoxide-d ꢀwithꢀ0.03%ꢀTMSꢀorꢀmethanol-d , and the
6
4
is based on experience rather than rationale. Raman spec- solution was inserted to an NMR tube separately. The number
trometry and IR spectrometry are also used, but require that of accumulations was 16 to 256. Phase correction was per-
36)
the compound has a carbonyl group. Thus, spectroscopic formed manually for each spectrum, and baseline correction
techniques are used widely to classify salts and cocrystals, but was applied over the entire spectral range. All spectra were
there are restrictions in their adaptation.
Theꢀ useꢀ ofꢀ UV/visibleꢀ (Vis)ꢀ spectrometryꢀ hasꢀ notꢀ beenꢀ re- of methanol at 3.3ppm.
referenced to the TMS resonance at 0ppm and methyl proton
15
15
portedꢀ asꢀ aꢀ classificationꢀ methodꢀ ofꢀ saltsꢀ andꢀ cocrystals.ꢀ UVꢀ
Solid-State
N-NMR Spectrometryꢀ ꢀS o l i d - s t a t e ꢀ N-
spectroscopy provides information on the excitation wave- NMR spectra were obtained with a 400 MR DD2 system
length of compounds and on difference of chemical structure (Agilent,ꢀCA,ꢀU.S.A.)ꢀequippedꢀwithꢀaꢀ4-mmꢀspinnerꢀatꢀroomꢀ
of compounds. In addition, changes of energy levels of elec- temperature.ꢀTheꢀNMRꢀtubesꢀwereꢀtightlyꢀfilledꢀwithꢀpowderꢀ
trons, which are characteristically related to ionizing proper- samples. This experiment was run with MAS rotation speed
ties,ꢀ canꢀ beꢀ detectedꢀ byꢀ UVꢀ spectroscopy.ꢀ Inꢀ thisꢀ study,ꢀ weꢀ ofꢀ 12ꢀkHz,ꢀ andꢀ cross-polarizationꢀ contactꢀ timeꢀ ofꢀ 10ꢀms.ꢀ Toꢀ
focusedꢀ onꢀ UV/Visꢀ spectrometryꢀ forꢀ useꢀ inꢀ theꢀ classificationꢀ optimizeꢀ theꢀ signalꢀ sensitivity,ꢀ delayꢀ timeꢀ wasꢀ adjustedꢀ toꢀ 5ꢀ
of salts and cocrystals.
or 10s. Data were collected for 46 or 92h. Phase correction
was performed manually for each spectrum, and baseline cor-
rection was applied over the entire spectral range. All spectra
Experimental
Materials Indomethacin, saccharin, L-arginine,ꢀ meglu- wereꢀreferencedꢀtoꢀtheꢀ15N-glycineꢀresonanceꢀatꢀ−347.5ppm.
mine citric acid, fumaric acid, and barium sulfate were
UV Spectrometryꢀ ꢀS olid-stateꢀ UV/Visꢀ absorptionꢀ spectraꢀ
purchasedꢀ fromꢀ FUJIFILMꢀ Wakoꢀ Pureꢀ Chemicalꢀ Corpora- wereꢀ obtainedꢀ withꢀ aꢀ UV-2450ꢀ systemꢀ (Shimadzu,ꢀ Kyoto,ꢀ
tion (Osaka, Japan). L-Lysineꢀ andꢀ tromethamineꢀ (TRIS)ꢀ wereꢀ Japan)ꢀ equippedꢀ withꢀ anꢀ integratingꢀ sphereꢀ unitꢀ (ISR-240,ꢀ
purchased from Tokyo Chemical Industry Co., Ltd. (Tokyo, Shimadzu) at room temperature. The samples were placed in
Japan).ꢀCompoundꢀAꢀ(Fig.ꢀ4)ꢀandꢀitsꢀcomplexesꢀwereꢀsynthe- sample cells. Barium sulfate was used as the reference stan-
sized at Sumitomo Dainippon Pharma Co., Ltd. Methanol and dard.ꢀTheꢀspectraꢀwereꢀacquiredꢀwithꢀaꢀ0.5-nmꢀsamplingꢀpitchꢀ
toluene were purchased from Nacalai Tesque (Osaka, Japan). in the wavelength range from 200 to 800nm. The obtained
Dimethylsulfoxide-d with 0.03% tetramethylsilane (TMS) for reflectionꢀ dataꢀ wereꢀ convertedꢀ toꢀ absorptionꢀ dataꢀ usingꢀ theꢀ
6
NMR was purchased from Cambridge Isotope Laboratories, Kubelka–Munk function.
Inc.ꢀ(MA,ꢀU.S.A.).ꢀTheꢀotherꢀsolventsꢀandꢀreagentsꢀusedꢀwereꢀ
purchasedꢀfromꢀFUJIFILMꢀWakoꢀPureꢀChemicalꢀCorporation.ꢀ Results
Waterꢀ wasꢀ obtainedꢀ fromꢀ aꢀ Milli-Qꢀ gradientꢀ systemꢀ (Merck,ꢀ
Characterization of Indomethacin Complexes The
NJ,ꢀU.S.A.).
PXRD patterns obtained for indomethacin complexes are
Preparation of Complexes Indomethacin and stoichio- shownꢀinꢀFig.ꢀ1.ꢀTheꢀpatternꢀforꢀeachꢀcomplexꢀandꢀeachꢀcoun-
metricꢀ (1ꢀ:ꢀ1)ꢀ counterꢀ moleculesꢀ wereꢀ dispensedꢀ intoꢀ separateꢀ ter molecule was different from that of indomethacin gamma
vials, and solvents were added and mixed. The following form, which was the initial crystal form. These patterns were
solvents were selected used for each complex, based on the also not in correspondence with that of indomethacin alpha
37)
literature :ꢀ 2-propanol/waterꢀ (9ꢀ:ꢀ1)ꢀ wasꢀ usedꢀ forꢀ indometha- form. In addition, no diffraction peaks derived from residual
cin arginine complex; ethanol for indomethacin saccharin indomethacin free form and counter molecules were observed.
complex and indomethacin lysine complex; toluene for in- The PXRD pattern of indomethacin saccharin complex was
38)
domethacin meglumine complex; and acetonitrile for indo- the same as that of a previous report. ꢀForꢀtheꢀotherꢀcomplex-
methacin TRIS complex. These solutions were heated at 80°C, es, except for indomethacin TRIS complex, only one crystal
36,37)
stirred for 10–30min, and then shaken with a shaking incuba- form of each was reported.
Therefore, each complex was
tor for 24h at 25°C. Solvents were removed from the vials by notꢀjustꢀaꢀphysicalꢀmixtureꢀofꢀindomethacinꢀandꢀeachꢀcounterꢀ
gradually opening the cap of the vial, and collected powders molecule, but a salt or a cocrystal. These complexes were not
wereꢀfiltered.
solvates, since residual solvents that were used in preparing
1
Compoundꢀ Aꢀ andꢀ stoichiometricꢀ (1ꢀ:ꢀ1)ꢀ counterꢀ moleculesꢀ the complexes were not detected in H-NMRꢀspectrumꢀofꢀanyꢀ
were dispensed into separate vials, and ethylacetate was added complex.
and mixed. These solutions were shaken with a shaking incu-
bator for 4h at 40°C. Solvents were removed from the vials by
The stoichiometry of these complexes, based on their
H-NMRꢀspectra,ꢀisꢀshownꢀinꢀTableꢀ1.ꢀIndomethacinꢀsaccharinꢀ
1
gradually opening the cap of the vial, and collected powders complex was demonstrated as a cocrystal and indomethacin
wereꢀfiltered.
meglumine complex was categorized as a salt in previous
37,38)
Powder X-Ray Diffraction (PXRD) Measurement reports.
The pK ꢀ differenceꢀ (ΔpK ) is used as one of the
a
a
PXRDꢀ patternsꢀ wereꢀ obtainedꢀ withꢀ aꢀ D8ꢀ ADVANCEꢀ X-rayꢀ classificationꢀmethodsꢀbetweenꢀsaltsꢀandꢀcocrystals.ꢀGenerally,ꢀ
3
9)
diffractometerꢀ (Bruker,ꢀ MA,ꢀ U.S.A.)ꢀ atꢀ roomꢀ temperature.ꢀ complexesꢀwithꢀΔpK >3 are salts. Regarding indomethacin
a
X-raysꢀ wereꢀ generatedꢀ withꢀ Cu-Kα radiation at 40kV and arginineꢀ complex,ꢀ theꢀ ΔpK between indomethacin and argi-
a
4
0ꢀmA.ꢀ IMCꢀ samplesꢀ wereꢀ placedꢀ onꢀ reflection-freeꢀ sampleꢀ nine is 9.0, as shown in Table 2. This suggests that indometha-
plates. Data were collected from 5 to 40° (2θ) with a step size cin arginine complex is a salt. Indomethacin lysine complex

Vol. 67, No. 9 (2019)
Chem. Pharm. Bull.
947
Fig.ꢀ 1.ꢀ PXRDꢀPatternsꢀofꢀIndomethacinꢀComplexes
meglumine complex, the signal derived from meglumine was
shifted from −358 to −348ꢀppm,ꢀasꢀseenꢀinꢀFigs.ꢀ2Cꢀandꢀ2D.ꢀ
Based on this, the interaction site between indomethacin and
meglumineꢀwasꢀidentifiedꢀasꢀcarboxylꢀgroupꢀofꢀindomethacinꢀ
and amine of meglumine. In indomethacin arginine complex,
the signal derived from arginine was shifted from −349 to
Table 1. Stoichiometry for Complexes
Sample
Stoichiometryꢀ(Freebaseꢀ:ꢀCounterꢀmolecule)
Indomethacin–Saccharin
Indomethacin–Meglumine
Indomethacin–Arginine
Indomethacin–Lysine
1ꢀ:ꢀ1
1ꢀ:ꢀ1
1ꢀ:ꢀ1
1ꢀ:ꢀ1
1ꢀ:ꢀ1
2ꢀ:ꢀ3
2ꢀ:ꢀ1
−
341ꢀppm,ꢀasꢀseenꢀinꢀinꢀFigs.ꢀ2Eꢀandꢀ2Fꢀfromꢀwhichꢀtheꢀinter-
actionꢀsiteꢀ betweenꢀindomethacinꢀandꢀarginineꢀ wasꢀidentifiedꢀ
as carboxyl group of indomethacin and α amine of arginine.
The value of the chemical shift difference between each
complex and each counter molecular related to interaction is
shown in Table 3. In addition, the chemical shift difference
Indomethacin–TRIS
Compound A–Citric acid
CompoundꢀA–Fumaricꢀacid
15
of each N-NMRꢀspectraꢀwasꢀnotꢀlargerꢀthanꢀ60ꢀppm.ꢀBasedꢀ
Table 2. pK ꢀValueꢀandꢀΔpK
a
a
onꢀthis,ꢀallꢀindomethacinꢀcomplexesꢀwereꢀclassifiedꢀasꢀcocrys-
tals.
35)
Sample
pKa
ΔpK ((pK (base)-pK (acid))
a
a
a
a)
RegardingꢀcompoundꢀAꢀfumaricꢀacidꢀcomplexꢀ(Fig.ꢀ3),ꢀtheꢀ
signal derived from N of imine was shifted from −132.3 to
−158.8ꢀppm,ꢀasꢀseenꢀinꢀFig.ꢀ4.ꢀOnꢀtheꢀotherꢀhand,ꢀinꢀtheꢀcaseꢀ
ofꢀcompoundꢀAꢀcitricꢀacidꢀcomplexꢀ(Fig.ꢀ3),ꢀtheꢀsignalꢀderivedꢀ
from N of imine was shifted from −132.3 to −202.8ꢀppmꢀ(Fig.ꢀ
4). Based on this, compound A fumaric acid complex and
compoundꢀAꢀcitricꢀacidꢀcomplexꢀwereꢀclassifiedꢀasꢀaꢀcocrystalꢀ
Saccharin
1.3
−2.9
3.8
9.0
6.7
3.8
—
b)
Meglumine
L-Arginine
L-Lysine
8.03
b)
13.2
b)
10.93
8.02
b)
TRIS
a)
Indomethacin
Citric acid
Fumaricꢀacid
Compound A
4.2
b)
3.128
0.9
1.0
—
b)
35)
3.03
4.05
and a salt, respectively.
Solid-State UV/Vis Spectrumꢀ ꢀT heꢀ solid-stateꢀ UV/Visꢀ
c)
a)ꢀAlhalawehꢀA.,ꢀRoyꢀL.,ꢀRodriguez-HomedoꢀN.,ꢀVelagaꢀS.ꢀP.,ꢀMol. Pharmaceu- absorption spectra of each complex and each physical mixture
tics, 9, 2605–2612 (2012). b) Handbook of Pharmaceutical Salts, 334 and 342. c)
Calculated by ADMET Predictor ver. 7.2
ofꢀ freebaseꢀ andꢀ aꢀ counterꢀ moleculeꢀ areꢀ shownꢀ inꢀ Fig.ꢀ 5.ꢀ Theꢀ
solid-stateꢀUV/Visꢀspectraꢀofꢀphysicalꢀmixturesꢀwasꢀcalculatedꢀ
from the spectra of freebase and a counter molecule, based
and indomethacin TRIS complex are also salts, because their on the stoichiometry of freebase and a counter molecule for
ΔpK values are >3.
a complex. The calculated spectra for physical mixtures were
Classification by Solid-State NMR Spectrumꢀ ꢀT h e ꢀs o l i d - obtainedꢀbecauseꢀtheꢀsolid-stateꢀUV/Visꢀspectraꢀwereꢀobtainedꢀ
a
15
state N-NMRꢀspectraꢀofꢀeachꢀcompoundꢀandꢀeachꢀindometh- byꢀaꢀreflectionꢀmethodꢀandꢀitꢀwasꢀdifficultꢀtoꢀcreateꢀaꢀuniformꢀ
acin complex, including saccharin, meglumine, and arginine, sampleꢀ surfaceꢀ forꢀ physicalꢀ mixtures.ꢀ Theꢀ solid-stateꢀ UV/Visꢀ
areꢀshownꢀinꢀFig.ꢀ2.ꢀTheꢀspectraꢀinꢀFig.ꢀ2,ꢀexceptꢀforꢀ2Cꢀandꢀ absorption spectrum of a complex was different from that of a
2
F,ꢀ haveꢀ aꢀ somewhatꢀ highꢀ degreeꢀ ofꢀ noiseꢀ becauseꢀ ofꢀ differ- physicalꢀmixtureꢀforꢀallꢀcomplexes.ꢀInꢀaddition,ꢀtheꢀsolid-stateꢀ
ence of relaxation time for the compounds. The chemical shift UV/Visꢀabsorptionꢀspectrumꢀofꢀaꢀcomplexꢀwasꢀdifferentꢀfromꢀ
of a signal derived from indomethacin was observed around that of freebase and a counter molecule for all complexes. The
−
207ppm in each indomethacin complex. In indomethacin integratedꢀUVꢀabsorptionꢀofꢀindomethacinꢀsaccharinꢀcomplexꢀ
saccharin complex, the signal derived from saccharin was was smaller than that of the physical mixture in the range
shifted from −226 to −230ꢀppm,ꢀasꢀseenꢀinꢀFigs.ꢀ2Aꢀandꢀ2B.ꢀ from 200 to 306nm, but was larger than that of the physical
Based on this, the interaction site between indomethacin and mixture in the range over 306nm. Indomethacin meglumine
saccharinꢀ wasꢀ identifiedꢀ asꢀ carboxylꢀ groupꢀ ofꢀ indomethacinꢀ complex and indomethacin arginine complex had larger ab-
and nitrogen of isothiazole of saccharin. In indomethacin sorption than their physical mixture in the range over 200nm,

9
48
Chem. Pharm. Bull.
Vol. 67, No. 9 (2019)
15
Fig.ꢀ 2.ꢀ N-NMRꢀSpectraꢀofꢀIndomethacinꢀComplexes
asꢀ shownꢀ inꢀ Fig.ꢀ 5.ꢀ Inꢀ addition,ꢀ UVꢀ spectraꢀ ofꢀ compoundꢀ Aꢀ UV/Visꢀabsorptionꢀspectra.ꢀTheꢀdifferenceꢀbetweenꢀcomplexesꢀ
complexes were changed from the respective physical mix- andꢀphysicalꢀmixturesꢀinꢀsolid-stateꢀUV/Visꢀabsorptionꢀspectraꢀ
tures,ꢀ asꢀ shownꢀ inꢀ Fig.ꢀ 6.ꢀ Therefore,ꢀ itꢀ isꢀ alsoꢀ possibleꢀ toꢀ indicates the conformation change of freebase and counter
ensureꢀ theꢀ formulationꢀ ofꢀ complexꢀ basedꢀ onꢀ theꢀ solid-stateꢀ molecules and the formation of interaction between freebase

Vol. 67, No. 9 (2019)
Chem. Pharm. Bull.
949
and counter molecule such as hydrogen bond and ionic bond.
not larger than that of physical mixture of freebase and each
TheꢀintegralꢀvalueꢀofꢀUVꢀspectraꢀofꢀeachꢀcomplexꢀandꢀeachꢀ counter molecule. On the other hand, with regard to the other
physical mixture is shown in Table 4. The integral value of indomethacin complexes and compound A citric acid complex,
solid-stateꢀUV/Visꢀabsorptionꢀspectrumꢀforꢀindomethacinꢀsac- theꢀ integralꢀ valueꢀ ofꢀ solid-stateꢀ UV/Visꢀ absorptionꢀ spectrumꢀ
charin complex and compound A fumaric acid complex were for each complex was larger than that of physical mixture of
freebase and each counter molecule.
15
Table 3.
Complex
N-NMRꢀ Chemicalꢀ Shiftꢀ Differenceꢀ betweenꢀ Freebaseꢀ andꢀ
Discussion
Inꢀ theꢀ solid-stateꢀ UV/Visꢀ absorptionꢀ spectraꢀ ofꢀ indometha-
cin complexes with saccharin, meglumine, arginine, lysine,
and TRIS, and compound A complexes with fumaric acid and
citric acid, differences from the initial, which was a physical
mixture of freebase and a counter molecule, were observed
inꢀtheꢀintensityꢀandꢀtheꢀshape.ꢀSpecifically,ꢀtheꢀintegralꢀvalueꢀ
ofꢀ solid-stateꢀ UV/Visꢀ absorptionꢀ spectrumꢀ forꢀ indomethacinꢀ
Sample
Chemical shift difference (ppm)
Indomethacin–Saccharin
Indomethacin–Meglumine
Indomethacin–Arginine
Compound A–Citric acid
CompoundꢀA–Fumaricꢀacid
3.9
4.0
9.7
72.2
26.4
Fig.ꢀ 3.ꢀ PXRDꢀPatternsꢀofꢀCompoundꢀAꢀComplexes
15
Fig.ꢀ 4.ꢀ N-NMRꢀSpectraꢀofꢀCompoundꢀAꢀComplexes

9
50
Chem. Pharm. Bull.
Vol. 67, No. 9 (2019)
Fig.ꢀ 6.ꢀ Solid-StateꢀUV/VisꢀSpectraꢀofꢀCompoundꢀAꢀComplexes
saccharin complex was similar to that of physical mixture of
indomethacin and saccharin, and that of compound A fumaric
acid complex was smaller than that of physical mixture of
freebase and fumaric acid. On the other hand, the integral
valuesꢀ ofꢀ solid-stateꢀ UV/Visꢀ absorptionꢀ spectraꢀ forꢀ theꢀ otherꢀ
indomethacin complexes and compound A citric acid complex
were larger than those of physical mixture of freebase and
each counter molecule. Based on these results, indometha-
cin saccharin complex was determined as a cocrystal, and
indomethacin meglumine complex was reported as a salt.
Compound A fumaric acid complex and compound A citric
acid were demonstrated as a cocrystal and a salt, respectively,
15
basedꢀ onꢀ theꢀ solid-stateꢀ N-NMRꢀ spectra.ꢀ Thisꢀ suggestedꢀ
that the difference was related to whether a complex was a
salt or a cocrystal. The structural difference between salts
and cocrystals lies in the interaction style between an active
ingredient and a counter molecule. Salts have ionic interaction
in the compounds, while cocrystals have nonionic interaction
such as hydrogen bond, van der Waals force, and π–π stacking
in the compound. Considering the difference between ionic
interaction and nonionic interaction, it may be suggested that
this phenomenon is caused by strength of interaction or charge
transfer. Regarding indomethacin arginine complex, indo-
methacin lysine complex, and indomethacin TRIS complex,
as shown in Table 2, the pK difference between indomethacin
a
and counter molecules are 9.0, 6.7, and 3.8, implying they are
salts,ꢀ generally.ꢀ Inꢀ theꢀ solid-stateꢀ UV/Visꢀ absorptionꢀ spectra,ꢀ
the integral value for the complex was larger than the value
for the physical mixture of indomethacin and each counter
molecule. This result indicated that indomethacin arginine
complex, indomethacin lysine complex, and indomethacin
TRIS complex were salts, and corresponded to the result
Fig.ꢀ 5.ꢀ Solid-StateꢀUV/VisꢀSpectraꢀofꢀIndomethacinꢀComplexes
based on pK and the results of Raman spectra.
a
15
Fromꢀ theꢀ resultsꢀ ofꢀ solid-stateꢀ N-NMRꢀ spectra,ꢀ threeꢀ
indomethacinꢀ complexesꢀ wereꢀ classifiedꢀ asꢀ cocrystals.ꢀ Indo-
methacin saccharin complex was reported as a cocrystal, and

Vol. 67, No. 9 (2019)
Chem. Pharm. Bull.
951
Tableꢀ 4.ꢀ IntegralꢀValueꢀofꢀSolid-StateꢀUV/VisꢀSpectra
theꢀ preparationꢀ ofꢀ thisꢀ manuscript,ꢀ andꢀ Mr.ꢀ Uedaꢀ ofꢀ Sumi-
tomoꢀDainipponꢀPharmaꢀCo.,ꢀLtd.ꢀforꢀcollectingꢀtheꢀsolid-stateꢀ
N-NMRꢀspectraꢀreportedꢀinꢀthisꢀpaper.
IntegralꢀvalueꢀofꢀUVꢀspectrumꢀ(200–800ꢀnm)
15
Sample
Complex
Physical mixture
Conflict of Interest Theꢀ authorsꢀ declareꢀ noꢀ conflictꢀ ofꢀ
interest.
Indomethacin–Saccharin
Indomethacin–Meglumine
Indomethacin–Arginine
Indomethacin–Lysine
1328.2
2098.6
3384.5
2352.2
2904.0
175.6
1557.6
899.2
796.0
1066.6
849.1
149.0
430.9
References and Notes
1)ꢀ BenetꢀL.ꢀZ.,ꢀBroccatelliꢀF.,ꢀOpreaꢀT.ꢀI.,ꢀAAPS J., 13, 519–547 (2011).
ꢀ
Indomethacin–TRIS
2
)
Kumer L., Amin A., Bansal A.K., Pharmaceutical Technology
March 2, 2008
Compound A–Citric acid
CompoundꢀA–Fumaricꢀacid
163.4
3
)
Rocheleau M. J., Current Pharmaceutical Analysis, 4, 25–32 (2008).
ꢀ
4)ꢀ “Handbookꢀofꢀpharmaceuticalꢀsalts:ꢀproperties,ꢀselection,ꢀandꢀuse,”ꢀ
ed.ꢀbyꢀStahlꢀP.ꢀH.,ꢀWermuthꢀC.ꢀG.ꢀInternationalꢀUnionꢀofꢀPureꢀandꢀ
AppliedꢀChemistry,ꢀVHCA,ꢀWiley-VCH,ꢀNewꢀYork,ꢀ2002.
thisꢀ alsoꢀ correspondedꢀ withꢀ theꢀ resultꢀ ofꢀ solid-stateꢀ UV/Visꢀ
absorption method. On the other hand, indomethacin meglu-
5
)
Bighley L. D., Berge S. M., Monkhouse D. C., “Encyclopedia of
PharmaceuticalꢀTechnology,”ꢀVol.ꢀ13,ꢀed.ꢀbyꢀSwarbickꢀJ.,ꢀBoylanꢀJ.ꢀ
C., Marcel Dekker Inc., New York, 1996, pp. 453–499.
mineꢀcomplexꢀwasꢀclassifiedꢀasꢀaꢀsaltꢀbasedꢀonꢀΔpK between
a
indomethacin and the counter molecule, and on the result of
3
6)
Raman spectrum. ꢀ Thisꢀ suggestedꢀ thatꢀ itꢀ wasꢀ difficultꢀ toꢀ
6) Gu C. H., Grant D. J. W., “Handbook of Experimental Pharmacol-
ogy:ꢀ Stereochemicalꢀ Aspectsꢀ ofꢀ Drugꢀ Actionꢀ andꢀ Disposition,”ꢀ ed.ꢀ
by Eichelbaum M., Testa B., Somogyi A., Springer, Berlin, 2003.
7)ꢀ Putraꢀ O.ꢀ D.,ꢀ Yonemochiꢀ E.,ꢀ Uekusaꢀ H.,ꢀ Cryst. Growth Des., 16,
568–6573 (2016).
8)ꢀ Aakeröyꢀ C.ꢀ B.,ꢀ Fasuloꢀ M.ꢀ E.,ꢀ Desperꢀ J.,ꢀ Mol. Pharm., 4, 317–322
2007).
9)ꢀ Moleculesꢀ fromꢀ theꢀ “generallyꢀ regardedꢀ asꢀ safe”ꢀ GRASꢀ list,ꢀ phar-
maceutical excipients, amino acids, food additives.
0)ꢀ AakeroyꢀC.ꢀB.,ꢀForbesꢀS.,ꢀDesperꢀJ.,ꢀJ. Am. Chem. Soc., 131, 17048–
15
adaptꢀtheꢀclassificationꢀmethodꢀbyꢀsolid-stateꢀ N-NMRꢀspec-
tra to all of the indomethacin complexes. On the other hand,
asꢀ weꢀ mentionedꢀ above,ꢀ theꢀ classificationꢀ resultꢀ ofꢀ solid-stateꢀ
UV/Visꢀ absorptionꢀ methodꢀ correspondedꢀ withꢀ theꢀ resultsꢀ ofꢀ
theꢀ otherꢀ methods.ꢀ Therefore,ꢀ theꢀ methodꢀ usingꢀ solid-stateꢀ
UV/Visꢀ absorptionꢀ spectraꢀ isꢀ moreꢀ usefulꢀ forꢀ classifyingꢀ
aꢀ complexꢀ asꢀ aꢀ saltꢀ orꢀ aꢀ cocrystalꢀ comparedꢀ toꢀ solid-stateꢀ
ꢀ
6
ꢀ
(
ꢀ
15
N-NMRꢀ spectra,ꢀ atꢀ leastꢀ inꢀ theꢀ caseꢀ ofꢀ indomethacinꢀ com-
plexes.
Theꢀclassificationꢀ methodꢀ byꢀ solid-stateꢀ UV/Visꢀ absorptionꢀ 11)ꢀ Remenarꢀ J.ꢀ F.,ꢀ Morissetteꢀ S.ꢀ L.,ꢀ Petersonꢀ M.ꢀ L.,ꢀ Moultonꢀ B.,ꢀ
1
17049 (2009).
spectra requires that samples are solids, because the evalua-
tionꢀisꢀcarriedꢀoutꢀwithꢀsolid-stateꢀUV/Visꢀspectraꢀofꢀanꢀactiveꢀ
agent, a counter molecule, and a complex. Thus, it is not pos-
sible to adopt complexes that have a counter molecule such
as hydrochloric acid, nitric acid, and acetic acid. However,
thisꢀclassificationꢀmethodꢀdoesꢀnotꢀrequireꢀtheꢀsampleꢀtoꢀbeꢀaꢀ
single crystal. Therefore, in the case of a complex for which a
single crystal was not obtained, this method offers a suitable
means in addition to other spectrometric techniques such as
MacPhee J. M., Guzman H. R., Almarsson O., J. Am. Chem. Soc.,
125, 8456–8457 (2003).
2)ꢀ Goodꢀ D.ꢀ J.,ꢀ Rodriguez-Hornedoꢀ N.,ꢀ Cryst. Growth Des., 9, 2252–
1
2
264 (2009).
3) Trask A. V., Motherwell W. D., Jones W., Int. J. Pharm., 320,
14–123 (2006).
4) Trask A. V., Motherwell W. D., Jones W., Cryst. Growth Des., 5,
01–121 (2005).
5)ꢀ Karkiꢀ S.,ꢀ Friscicꢀ T.,ꢀ Fabianꢀ L.,ꢀ Laityꢀ P.ꢀ R.,ꢀ Dayꢀ G.ꢀ M.,ꢀ Jonesꢀ W.,ꢀ
Adv. Mater., 21, 3905–3909 (2009).
1
1
1
1
1
solid-stateꢀ NMR,ꢀ IR,ꢀ andꢀ Ramanꢀ spectrometry.ꢀ Inꢀ addition,ꢀ 16) Sun C. C., Hou H., Cryst. Growth Des., 8, 1575–1579 (2008).
this method can be adopted to complexes regardless of their 17) McNamara D. P., Childs S. L., Giordano J., Iarriccio A., Cassidy J.,
specificꢀchemicalꢀstructure.ꢀTherefore,ꢀthisꢀsolid-stateꢀUV/Visꢀ
classificationꢀ methodꢀ canꢀ provideꢀ keyꢀ informationꢀ onꢀ com-
plexesꢀ withoutꢀ aꢀ singleꢀ crystal,ꢀ andꢀ isꢀ notꢀ limitedꢀ toꢀ specificꢀ
chemical structures.
Shet M. S., Mannion R., O’Donnell E., Park A., Pharm. Res., 23,
888–1897 (2006).
8) Vishweshwar P., McMahon J. A., Bis J. A., Zaworotko M. J., J.
Pharm. Sci., 95, 499–516 (2006).
9) Meanwell N. A., Annu. Rep. Med. Chem., 43, 373–404 (2008).
0) Schultheiss N., Newman A., Cryst. Growth Des., 9, 2950–2967 (2009).
1
1
1
2
Conclusion
21) Duggirala N. K., Perry M. L., Almarsson O., Zaworotko M. J.,
In indomethacin complexes, differences between salts and
cocrystalsꢀ wereꢀ observedꢀ inꢀ solid-stateꢀ UV/Visꢀ absorptionꢀ
spectra. The same tendency was observed in compound A
complexes.ꢀ Classificationꢀ resultsꢀ obtainedꢀ usingꢀ thisꢀ methodꢀ
corresponded to the results obtained using other techniques,
Chem. Commun., 52, 640–655 (2016).
2
2)ꢀ Regulatoryꢀ Classificationꢀ ofꢀ Pharmaceuticalꢀ Co-Crystals,ꢀ U.S.ꢀ De-
partmentꢀofꢀHealthꢀandꢀHumanꢀServices,ꢀFoodꢀandꢀDrugꢀAdminis-
tration,ꢀCenterꢀforꢀDrugꢀEvaluationꢀandꢀResearchꢀ(CDER),ꢀFebruaryꢀ
2018.
suchꢀasꢀΔpK ꢀandꢀsingle-crystalꢀX-rayꢀdiffraction.ꢀThisꢀclassi- 23)ꢀ Reflectionꢀpaperꢀonꢀtheꢀuseꢀofꢀcocrystalsꢀandꢀotherꢀsolid-stateꢀformsꢀ
a
ofꢀactiveꢀsubstancesꢀinꢀmedicinalꢀproducts,ꢀFebruaryꢀ2014.
4)ꢀ DesirajuꢀG.ꢀR.,ꢀCrystEngComm, 5, 466–467 (2003).
5) Dunitz J. D., CrystEngComm, 5, 506 (2003).
6) Aakeroy C. B., Salmon D. J., CrystEngComm, 7, 439–448 (2005).
7) Jones W., Motherwell S., Trask A. V., MRS Bull., 31, 875–879 (2006).
8) Stahly G. P., Cryst. Growth Des., 7, 1007–1026 (2007).
9) Huang K. S., Britton D., Etter M. C., Byrn S. R., J. Mater. Chem., 7,
ficationꢀmethodꢀrequiresꢀthatꢀaꢀfreeꢀform,ꢀaꢀcounterꢀmolecule,ꢀ
2
2
2
2
2
2
and a complex are solid, but it is not necessary to obtain a
single crystal of a complex. Moreover, this method can be
adapted to various compounds, irrespective of their chemical
structures. Therefore, this method may provide key informa-
tion to classify salts and cocrystals in the case of compounds
withoutꢀaꢀsingleꢀcrystalꢀorꢀspecificꢀtypeꢀofꢀchemicalꢀbond.
713–720 (1997).
3
0) Smith M., March J., March J., “March’s Advanced Organic Chem-
istry:ꢀReactionsꢀMechanismsꢀandꢀStructure,”ꢀ5thꢀed.,ꢀJ.ꢀWiley,ꢀNewꢀ
York, 2001.
Acknowledgments We thank Mr. Yoshimizu of Sumito-
mo Dainippon Pharma Co., Ltd. for helpful discussions during

9
52
Chem. Pharm. Bull.
J. Am. Chem. Soc., 128, 8199–8210 (2006).
Vol. 67, No. 9 (2019)
3
3
1)ꢀ BhogalaꢀB.ꢀR.,ꢀBasavojuꢀS.,ꢀNangiaꢀA.,ꢀCrystEngComm, 7, 551–562
(2005).
36)ꢀ KojimaꢀT.,ꢀTsutsumiꢀS.,ꢀYamamotoꢀK.,ꢀIkedaꢀY.,ꢀMoriwakiꢀT.,ꢀ Int.
J. Pharm., 399, 52–59 (2010).
2)ꢀ Putraꢀ O.ꢀ D.,ꢀ Yoshidaꢀ T.,ꢀ Umedaꢀ D.,ꢀ Higashiꢀ K.,ꢀ Uekusaꢀ H.,ꢀ Yone-
mochi E., Cryst. Growth Des., 16, 5223–5229 (2016).
3)ꢀ PutraꢀO.ꢀD.,ꢀYoshidaꢀT.,ꢀUmedaꢀD.,ꢀGunjiꢀM.,ꢀUekusaꢀH.,ꢀYonemo-
37) Yamashita H., Hirakura Y., Yuda M., Terada K., Pharm. Res., 31,
1946–1957 (2014).
3
chi E., Cryst. Growth Des., 16, 6714–6718 (2016).
4)ꢀ Umedaꢀ Y.,ꢀ Fukamiꢀ T.,ꢀ Furuishiꢀ T.,ꢀ Suzukiꢀ T.,ꢀ Tanjohꢀ K.,ꢀ Tomonoꢀ
K., Drug Dev. Ind. Pharm., 35, 843–851 (2009).
38)ꢀ Basavojuꢀ S.,ꢀ Bostromꢀ D.,ꢀ Velagaꢀ S.ꢀ P.,ꢀ Pharm. Res., 25, 530–541
(2008).
39)ꢀ Cruz-CabezaꢀA.ꢀJ.,ꢀCrystEngComm, 14, 6362–6365 (2012).
3
35) Li Z. J., Abramov Y., Bordner J., Leonard J., Medek A., Trask A. V.,