Decarboxylation of indomethacin induced by heat treatment

Article abstract of DOI:10.1016/j.ijpharm.2018.04.022

As crystalline indomethacin is heated and subsequently cooled, it transforms into glassy indomethacin. While the original crystals are off-white in color, the glass becomes blackish-brown via a yellow intermediate stage. TLC of the components of the glass

Full text of DOI:10.1016/j.ijpharm.2018.04.022

Accepted Manuscript
Decarboxylation of Indomethacin Induced by Heat Treatment
Yohsuke Shimada, Haruki Komaki, Ayako Hirai, Satoru Goto, Yasuyuki
Hashimoto, Hiromi Uchiro, Hiroshi Terada
PII:
S0378-5173(18)30239-4
https://doi.org/10.1016/j.ijpharm.2018.04.022
IJP 17429
DOI:
Reference:
International Journal of Pharmaceutics
To appear in:
Received Date:
Revised Date:
Accepted Date:
26 January 2018
23 March 2018
12 April 2018
Please cite this article as: Y. Shimada, H. Komaki, A. Hirai, S. Goto, Y. Hashimoto, H. Uchiro, H. Terada,
Decarboxylation of Indomethacin Induced by Heat Treatment, International Journal of Pharmaceutics (2018), doi:
https://doi.org/10.1016/j.ijpharm.2018.04.022
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers
we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and
review of the resulting proof before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Decarboxylation of Indomethacin Induced by Heat Treatment
Yohsuke Shimada1,2, Haruki Komaki1, Ayako Hirai1, Satoru Goto1,2*, Yasuyuki Hashimoto1,
Hiromi Uchiro1,2 and Hiroshi Terada1,2
1Faculty of Pharmaceutical Sciences, Tokyo University of Science, 2641 Yamazaki, Noda, Chiba
278-8510, Japan
2Center for Drug Delivery Research, Faculty of Pharmaceutical Sciences, Tokyo University of
Science, 2641 Yamazaki, Noda, Chiba 278-8510, Japan
*To whom correspondence should be addressed: s.510@rs.tus.ac.jp

Abbreviations used: COSY, correlated spectroscopy; DEPT, distortionless enhancement by
polarization transfer; DSC, differential scanning calorimetry; DTA, differential thermal analysis;
EI, electron ionization; HMQC, heteronuclear multiple quantum coherence; INM,
indomethacin; MS, mass spectrometry; SEM, scanning electron microscopy; TG,
thermogravimetry; XRPD, X-ray powder diffraction.
1

ABSTRACT
As crystalline indomethacin is heated and subsequently cooled, it transforms into
glassy indomethacin. While the original crystals are off-white in color, the glass becomes
blackish-brown via a yellow intermediate stage. TLC of the components of the glass revealed
three bands. The yellow component, which is generated either under hypoxic conditions or in
the dark, was elucidated by NMR spectroscopy to be a decarboxylated fragment produced by
thermal degradation. The colorless component is proposed to be formed by the opening of the
indole ring of indomethacin; the structure of this degradation product was identified by EI-MS
to be the same as the oxidative-cleavage product formed upon UV-irradiation, as previously
reported. Another band was a blackish-brown pigment whose mobility placed it close to the
TLC baseline. This oxidative-cleavage product and the blackish-brown pigment are not
generated under hypoxic conditions. However, the extent of indomethacin decarboxylation
under hypoxic conditions was found to be dependent on the heating temperature and time.
Consequently, we prepared amorphous indomethacin through control of the heating temperature
and time; heating at 160 °C for 30 min or less under hypoxic conditions is optimum for
obtaining pure amorphous indomethacin.
Keywords: indomethacin, amorphous, decarboxylation, NMR, EI-MS
2

1. INTRODUCTION
Compounds belonging to the arylalkanoic acid family derived from 2-arylacetic acids,
such as indomethacin, sulindac, diclofenac, and tolmetin, and those derived from
2-arylpropionic acids, such as naproxen, ibuprofen, fenoprofen, ketoprofen, flurbiprofen, and
loxoprofen, are anti-inflammatory, antipyretic, and analgesic agents. These non-steroidal
anti-inflammatory drugs (NSAIDs) are clinically effective in the therapy of inflammation,
chronic rheumatoid arthritis, periarthritis, osteoarthritis, spondylosis deformans, and acute gout,
by inhibiting cyclooxygenase (COX), an enzyme belonging to the prostaglandin synthetic
cascade. Indomethacin (INM) is a commonly used effective NSAID. Its abilities and
photoproducts to scavenge hydroxyl radicals and inhibit xanthine oxidase have also been
reported (Lien et al., 2013). Other NSAIDs, such as ketoprofen, have also been reported to have
radical-scavenging abilities, these NSAIDs also present photosensitivity as a side effect of their
photolysis (Takara et al., 2017). However, INM is sparingly soluble in aqueous media.
Therefore, there have been numerous attempts to improve its water solubility. One
well-documented method is to make INM amorphous, which could be accomplished by the
heating quenching method or hot melt extrusion method (ElShaer et al., 2011; Forster et al.,
2001; Inada et al., 2013; Jung et al., 2010; Kawakami et al., 2004; Shimada et al., 2013a, 2013b;
Shimada et al., 2018).
Amorphous INM is readily obtained by the heat treatment of crystalline INM at
temperatures above its melting point, followed by cooling of the fused INM. Heat-treated INM
does not crystallize following cooling, and exhibits a transition to the glass state at about 42 °C
(Shimada et al., 2013a). In addition, heat treatment of crystalline INM results in a change in
color; off-white crystalline INM turns yellow upon melting and then blackish brown. Despite
being known for some time, the mechanism associated with these color changes is unclear
(Shimada et al., 2013a).
This study investigated the mechanism of the heat-induced color change of INM by
3

means of scanning electron microscopy (SEM), thermogravimetry/differential-thermal analysis
(TG/DTA), differential scanning calorimetry (DSC), X-ray powder diffraction (XRPD),
thin-layer chromatography (TLC), electron-ionization mass spectrometry (EI-MS), and nuclear
magnetic resonance (NMR) spectroscopy.
2. MATERIALS AND METHODS
2.1 Materials
Crystalline INM was purchased from Wako Pure Chemical Industries, Ltd. (Osaka, Japan).
Glass TLC plates, coated with silica gel and F254 fluorescent indicator, were obtained from
Merck (Darmstadt, Germany). Other chemicals and solvents were of reagent grade and
commercially available.
2.2 Preparation of the glass form of INM
In a typical procedure, 20 mg of crystalline INM was heated at 200 °C for 1, 5, 15, 30, or 60
min in a sample tube immersed in an oil bath to form fused INM, after which it was cooled to
room temperature to obtain glassy INM. Heating and cooling were performed either under
aerobic or hypoxic conditions, which were achieved by purging the air inside the sample tube
with argon.
2.3 SEM
The scanning electron micrographs of crystalline INM and glass form of INM (which was
heated at 200 °C for 1 or 60 min) were taken using a miniscope (TM-1000; Hitachi
High-Technologies Co., Tokyo, Japan). Samples were fixed on a sample holder with conductive
double-sided adhesive tape. The operating conditions were: accelerating voltage 15 kV and
scanning time 40 s. The glassy samples were crushed with an agate mortar and pestle into
powder before measurements.
2.4 DSC and XRPD
Thermal analyses were performed using a DSC8230 instrument (Rigaku Co., Tokyo, Japan)
4

with 10 mg of INM placed in an aluminum pan. After the pan was sealed, the temperature was
scanned between –40 and 170 °C at 10 °C/min, using the onset temperature as the melting
temperature. XRPD analyses were performed in parallel with the DSC measurements on a
RINT2000 X-ray diffractometer (Rigaku Co.) with Cu Kα1 radiation, operating at 40 kV and 40
mA in the 10–40° 2 range, with scanning at 15 °/min.
2.5 TLC
The crystalline and glass forms of INM were dissolved in 2 mL of acetonitrile. Samples were
developed on TLC plates with ethyl acetate as the eluent, and were visualized under ultraviolet
light (254 nm). In addition, the separated products were extracted by preparative TLC method
using a sufficient amount of chloroform as eluent.
2.6 NMR spectroscopy and EI-MS
Crystalline INM and the corresponding fused samples were dissolved in 1 mL of chloroform-d1
and filtered through cotton mesh to provide a solution for 1H-NMR and 13C-NMR spectroscopy.
1
The structures of INM and decarboxylated INM were assigned by H-1H-COSY, and HMQC
techniques, which are more accurate NMR methods. Spectra were acquired using an LA-400
NMR spectrometer (JEOL Ltd, Tokyo, Japan) using tetramethylsilane as an internal reference.
EI-MS studies were performed on a Varian 910-MS TQ-FT-ICR mass spectrometer (JASCO
International Co. Ltd., Tokyo, Japan).
2.7 Preparation of intact glassy INM
In a typical experiment, 20 mg of crystalline INM in a sample tube was heated in an oil bath at
predetermined temperatures and heating times under hypoxic conditions, after which it was
cooled to room temperature to obtain glassy INM.
2.8 Solubility of crystalline or amorphous INM
The crystalline and glassy forms of INM were added to a sodium citrate-hydrochloric acid buffer
(pH 1.2) and shaken in a water bath at 310 K for 120 min. The supernatant was then filtered using a
membrane filter (Minisart RC 4, 0.20 μm; Sartorius, Göttingen, Germany). The concentration of
5

INM in the supernatant was determined using HPLC (Shimadzu Co., Kyoto, Japan), with a mobile
phase of acetonitrile:tetrahydrofuran:acetic acid:ultrapure water (50:5:0.2:44.8 v/v/v/v) and a flow
rate of 1 mL·min-1 using a reversed phase column (Capcell Pak C18; Shiseido; 5 μm, 4.6 mm Φ×
250 mm) at 313 K. The amount of INM was determined by monitoring the absorbance at 320 nm.
Curve fitting was carried out by using Microsoft Office excel solver module (Microsoft Co.,
Redmond, USA).
3. RESULTS
3.1 Changes in State and Color of Indomethacin upon Heat Treatment
Heating crystalline INM at temperatures above its melting temperature (160 °C;
Shimada et al., 2013a, 2013b) resulted in a color change, from off-white crystals to pale yellow
and then blackish-brown, depending on the heating temperature and time, with the INM melt
becoming a paste. Figure 1a shows photographic images of INM after heating at 200 °C under
aerobic conditions for various times. Upon heating, a transparent yellow paste was
obtained
Figure 1
in a few minutes, after which it turned brown and then blackish-brown. DSC of INM over the
–40 °C to 170 °C range at 10 °C/min revealed an endothermic peak at 160 °C that corresponds
to the melting of INM. However, no exothermic peak was observed upon cooling of the fused
INM to –40 °C; instead a transition signal due to the formation of the glassy state was observed
at 42 °C, as reported recently by us (Shimada et al., 2013a, 2013b). Although signals specific to
the crystalline form were observed in the XRPD pattern of crystalline INM, fused INM
exhibited halo-patterns characteristic of the amorphous state (see Supplemental Data). From the
DSC and XRPD results, cooling of the fused INM crystals results in the formation of the
amorphous state. However, as shown in Figure 1b, there was no qualitative morphology change
in the SEM image of INM after it became amorphous (following 1 and 60 min of heating).
6

Although the particle size of the crystal INM (γ type) is small, it did not have a characteristic
form in its crystal habit.
Table 1
3.2 Structural Changes of Indomethacin upon Heating
To elucidate which chemical species are generated by heating INM, the crystalline and
heat-treated INM samples were subjected to TLC with ethyl acetate as the eluent. As shown in
Fig. 2, heat-treated INM exhibits three distinct spots, with Rf values of 0.667, 0.478, and 0.289.
Of these, the spot at Rf 0.289 is apparently due to INM because the same Rf was observed for the
spot corresponding to crystalline INM. The major (Rf = 0.667) and minor (Rf = 0.478) spots
should correspond to the products generated by the heat treatment of INM; and these products
are referred to as Product-1 and -2, respectively. In addition, a thin smear was observed from the
baseline to the solvent front of the TLC plate for the amorphous INM sample, and an amount of
blackish-brown material remained at the baseline. We refer to these blackish-brown species, as a
whole, as “blackish-brown pigments.” Product-1 was yellow in color, as observed by the spot
corresponding to the INM crystals, whereas Product-2 was colorless. Hence, the blackish-brown
color of heat-treated INM is due to the blackish-brown pigments rather than Product-1 or -2.
Figure 2
We retrieved Product-1 and -2 by preparative TLC using chloroform as the eluent. The
13C-NMR spectrum of Product-1 in chloroform-d1 reveals that the signals corresponding to C-11
( 30.01) and C-12 ( 176.77) of the dissolved crystalline INM were absent, with a new signal
appearing at  8.77 (Table 1). The “lost” C-11 and C-12 signals correspond to the
C-11-methylene and C-12-carboxyl groups in crystalline INM, with the “new” signal
corresponding to a methyl group. MS reveals that Product-1 has a MW 313.1. These spectra, as
1
1
well as those acquired by H-NMR, H-1H-COSY, and HMQC spectroscopy, reveal that the
7

major component (Product-1) in the heat-treated INM is the decarboxylated INM, as was
observed when INM was photochemically degraded (Weedon and Wong, 1991). INM and
Product-1 are yellow due to conjugation between their indole rings and amide groups.
The EI-MS spectrum suggests that Product-2 has a MW of 288.0. We attempted to
identify the chemical structure of Product-2; however, the amount of Product-2 recovered was
too low for NMR spectra to be acquired. As Product-2 is colorless, it is possible that the original
indole ring in INM may have been oxidatively destroyed. Wu et al. determined the chemical
structure of their photoproduct to be a 1,2-dioxane derivative with a 6-membered ring bridging
the 2nd and 3rd positions of the indole ring, which was produced by UV irradiation and oxygen
via an acyl radical. EI-MS and NMR spectroscopy revealed that the compound with a MW of
288.0 was a fragment ion produced through the cleavage of the indole ring in the 1,2-dioxane
derivative by electron collision during EI-MS (Wu et al., 1997); therefore, we consider
Product-2 to correspond to this fragment. In addition, it was very difficult to isolate sufficient
quantities of the blackish-brown pigments for characterization by NMR spectroscopy and MS
because their extractions with various solvents, such as methanol, ethanol, ethyl acetate, acetone,
acetonitrile, or phosphate buffer (pH 7.4), proved to be very difficult. Hence, we were unable to
establish the chemical structural changes that take place in INM upon heating to form the
blackish-brown pigments. However, the blackish-brown pigments are likely to be polyphenols
and/or polymers with quinone structures that are produced by the oxidation of INM to produce
the quinone derivative of p-anisidine (p-aminoanisole), as suggested by the photochemical
degradation of INM (Weedon and Wong, 1991).
3.3 Heat Treatment of Indomethacin under Hypoxic Conditions
As oxygen was thought to trigger the polymerization of INM, we subsequently heat
treated INM under hypoxic conditions by replacing air with argon gas. INM heat-treated at
200 °C for 60 min under hypoxic conditions was yellow in color, in contrast to the
8

blackish-brown color following heat treatment under aerobic conditions. Hence, the
blackish-brown color of the product formed under aerobic conditions is proposed to be due to
the oxidative polymerization of INM. In addition, the TLC plate of INM following heating
under hypoxic conditions exhibited two spots, the Rf values of which were identical to those of
intact INM and Product-1 obtained by heat treatment under aerobic conditions. In addition, the
colors of the Product-1 bands following the heat treatment of INM under hypoxic and aerobic
conditions were the same. We confirmed by NMR spectroscopy that Product-1 obtained by
heating under hypoxic and aerobic conditions were identical. Hence, Product-2, as well as
blackish-brown pigments formed by heating under aerobic conditions, is a product from the
oxidation of INM.
1
Product-1 is proposed to be decarboxylated INM based on 13C- and H-NMR
spectroscopy and MS. Scheme 1 displays a possible mechanism for the heat-induced
decarboxylation of INM under both hypoxic and aerobic conditions. The 10-π-electron ring
system of INM is interrupted by protonation from the carboxylic acid at C-12, which attacks
C-2 carbon, adjacent to the nitrogen atom. INM then decarboxylates leading immediately to
Product-1 via an unstable intermediate structure. It is noteworthy that only Product-1 was
produced (without the formation of any Product-2) during the photochemical decarboxylation of
INM (Weedon and Wong, 1991). This indicates that the mechanism depicted in Scheme 1 and
that for the formation of Product-2 are complementary; hence Product-1 is rapidly produced
without the need for additional high energy, such as that produced during EI-MS.
Scheme 1
3.4 Temperature Dependence of the Degradation of Indomethacin
We next examined the dependence of Product-1 formation on heating temperature and
heating time under hypoxic conditions. With increasing temperature and heating time, the “new”
9

1
signal at 2.19 ppm in the H-NMR spectrum of the decarboxylated product, due to the three
methyl protons at C-(11), increased synchronously with the decrease observed for the peak at
3.69 ppm corresponding to the two methylene protons at C-11 of the intact INM. The signals at
2.19 ppm and 3.69 ppm are referred to as “S1” and “S2”, respectively. As the areas of S1 (IVS1)
and S2 (IVS2), obtained by peak integration, represent the relative amounts of Product-1 and the
original intact INM, respectively, we determined the amount of Product-1 relative to the original
amount of INM using the equation given below.
Amount of Product-1 generated (%) = IVS1/3/(IVS1/3 + IVS2/2) 100
Figure 3 shows the percentage of Product-1, generated by heating for 60 min under
hypoxic conditions, as a function of temperature. The decarboxylation of INM begins at its
melting temperature (Tm, 160 °C), and progressively increases with increasing temperature,
reaching about 30% at 200 °C. When INM was heated at 200 °C for various times under
hypoxic conditions, the percentage of Product-1 increased linearly with increasing heating time,
as shown in Fig. 4. As the generation of Product-1 is both dependent on heating temperature and
time, it is possible to determine conditions that avoid its production. After examination of the
amounts of Product-1 generated at various heating temperatures and times under hypoxic
conditions, we found INM begins to degrade to form Product-1 at the Tm of INM (160 °C) after
heating for 40 min or longer; above Tm the degradation of INM proceeds more rapidly with
increasing heating time, and Product-1 is instantaneously generated at 200 °C. Figure 5
summarizes the relationship between heating temperature and heating time for the preparation
of amorphous INM under hypoxic conditions. Clearly, amorphous INM, devoid of any
Product-1, Product-2, or blackish-brown-pigment impurities, can be obtained under hypoxic
conditions through control of the heating temperature and time, as shown by the data in Fig. 5.
Product-1 was generated by heating INM at 160 °C for more than 40 min, at 170 °C for more
10

than 10 min, and at 180 °C for more than 5 min; heating at 160 °C for 30 min or less under
hypoxic conditions is optimum for obtaining pure amorphous INM. When INM is heated under
aerobic conditions, Product-1, Product-2, and the blackish-brown component are not produced
when heated at 160 °C for less than 2 min, and at 170 °C for less than 1 min.
Figure 3, Figure 4, Figure 5
Figure 6 shows the solubility data of γ type and amorphous INM at 25 °C and pH 1.2.
Amorphous INM was prepared by heating at 170 °C for about 5 min in the shaded range shown
in Fig. 5 and cooling to room temperature. In this experiment, since INM (pKa 4.3) exists in the
molecular form in solution, the saturation value after 120 min is its solubility. From the
literature, the INM solubility is improved several times as it becomes amorphous, which was
seen in this experiment as well.
Figure 6
4. DISCUSSION
We heated crystalline INM above its melting point of 160 °C under aerobic conditions
for specific times, which resulted in the formation of two species, namely Product-1 and -2,
which were yellow and colorless, respectively, as well as blackish-brown pigments. Product-1
was determined to be decarboxylated INM, while Product-2 corresponded to the fragment ion
obtained by cleavage of the indole ring in a 1,2-dioxane derivative by EI-MS (Wu et al., 1997).
However, heating INM under hypoxic conditions generated only Product-1, with no Product-2
or blackish-brown pigments formed, indicating that the blackish-brown color of heat-treated
INM is mainly due to polyquinones produced by the oxidation of the p-aminoanisole moiety of
INM, as observed during the photochemical degradation of INM (Weedon and Wong, 1991; Wu
et al., 1997). Product-1 and Product-2 are generated through different pathways; Product-1 is
produced only by heating, while Product-2 is produced by heat and oxygen, which behaves as a
11

catalyst for the formation of the blackish-brown pigment.
The rate of formation of Product-1 through the decarboxylation of INM above its Tm
(160 °C) depends on the heating temperature and the heating time. The degradation of INM at a
fixed heating temperature increased with time in a linear manner. In addition, Figure 5 reveals
the times over which uncontaminated amorphous INM was formed at each temperature. For
example, INM was stable at 160 °C for up to 30 min, while at 200 °C it took only 1 min for
degradation to occur. It is noteworthy that the decarboxylation of INM is temperature dependent,
and that the reaction essentially does not occur after cooling to room temperature in the absence
of UV irradiation. Therefore, we were able to prepare uncontaminated stable amorphous INM
by controlling the heating temperature and time.
The reason why the melted INM is yellow in color is unknown; however the thermal
decomposition results provide clues. Since amorphous INM and Product-1 are yellow in color
but crystalline INM and Product-2 are not, the yellow color might be due to the degree of
indole-ring conjugation and the resonance structures formed, which depend on flexible
conformation not present in the crystalline form. Regardless of decarboxylation, the color
depends on whether the indole ring and the benzene ring form a crystal structure. The color is
yellow for the flexible amorphous form, but off-white for the rigid crystal form. In particular,
we believe that the dihedral angle between indole ring and benzene ring affects yellow.
The water solubility of INM is greatly improved by changing its state from crystalline
to amorphous; however amorphous INM is decomposed by light as well as heat. We examined
the degradation of INM by changing the heating temperature and time. This revealed that
amorphous INM is readily obtained without decomposition under appropriate conditions. In
addition, it has been suggested that the heat-decomposition products are reactive toward radicals.
Since the presence of free radicals may induce the decomposition of amorphous INM, it is also
important to consider protecting INM against free radicals.
12

5. REFERENCES
ElShaer, A., Khan, S., Perumal, D., Hanson, P., Mohammed, A.R., 2011. Use of amino acids as
counterions improves the solubility of the BCS II model drug, indomethacin. Curr. Drug. Deliv.
8, 363–72.
Forster, A., Hempenstall, J., Tucker, I., Rades, T., 2001. Selection of excipients for melt
extrusion with two poorly water-soluble drugs by solubility parameter calculation and thermal
analysis. Int. J. Pharm. 226, 147–161.
Hawkey, C.J., 2000. Non-steroidal anti-inflammatory drug gastropathy. Gastroenterology 119,
521–535.
Inada, A., Oshima, T., Takahashi, H., Baba, Y., 2013. Enhancement of water solubility of
indomethacin by complexation with protein hydrolysate. Int. J. Pharm. 453, 587–593.
Jung, M.S., Kim, J.S., Kim, M.S., Alhalaweh, A., Cho, W., Hwang, S.J., Velaga, S.P., 2010.
Bioavailability of indomethacin-saccharin cocrystals. J. Pharm. Pharmacol. 62, 1560–1568.
Kawakami, K., Miyoshi, K., Ida, Y., 2004. Solubilization behavior of poorly soluble drugs with
combined use of Gelucire 44/14 and cosolvent. J. Pharm. Sci. 93, 1471–1479.
Lien, G.S., Chen, C.S., Chen, W.Y., Huang, S.H., Cheng, K.T., Lin, C.M., Chao, S.H., 2013.
Photoproducts of indomethacin exhibit decreased hydroxyl radical scavenging and xanthine
oxidase inhibition activities. J. Food Drug Anal. 21, 332–336.
Shimada, Y., Goto, S., Uchiro, H., Hirabayashi, H., Yamaguchi, K., Hirota, K., Terada, H.,
2013a. Features of heat-induced amorphous complex between indomethacin and lidocaine.
Colloids Surf. B Biointerfaces 102, 590–596.
Shimada, Y., Goto, S., Uchiro, H., Hirota, K., Terada, H., 2013b. Characteristics of amorphous
complex formed between indomethacin and lidocaine hydrochloride. Colloids Surf. B
Biointerfaces 105, 98–105.
Shimada, Y., Tateuchi, R., Chatani, H., Goto, S., 2018. Mechanisms underlying changes in
indomethacin solubility with local anesthetics and related basic additives. J. Mol. Struct. 1155,
13

165–170.
Takara, A., Kobayashi, K., Watanabe, S., Okuyama, K., Shimada, Y., Goto, S., 2017. Dibucaine
inhibits ketoprofen photodegradation via a mechanism different from that of antioxidants. J.
Photochem. Photobiol. A Chem. 333, 208–212.
Weedon, A.C., Wong, D.F., 1991. The photochemistry of indomethacin. J. Photochem.
Photobiol. A Chem. 61, 27–33.
Wehling, M., 2014. Non-steroidal anti-inflammatory drug use in chronic pain conditions with
special emphasis on the elderly and patients with relevant comorbidities: management and
mitigation of risks and adverse effects. Eur. J. Clin. Pharmacol. 70, 1159–1172.
Wu, A.B., Cheng, H.W., Hu, C.M., Chen, F.A., Chou, T.C., Chen, C.Y., 1997. Photolysis of
indomethacin in methanol. Tetrahedron Lett. 38, 621–622.
14

Figure captions
Figure 1. (a) Photographic images of crystalline INM (0 min) in a sample tube, and the color
changes observed for fused INM heated at 200 °C for 1, 5, 15, 30, and 60 min under aerobic
conditions. (b) Left to right: SEM images of crystalline INM (0 min, x1000 and x2000), of
fused INM heated at 200 °C for 1 min (x500) and 60 min (x1000) under aerobic conditions,
during the color change shown in (a).
Figure 2. Thin-layer chromatograms. Left lane: INM (Rf 0.289). Right lane: three distinct spots
for fused INM; Product-1 (Rf 0.667), Product-2 (Rf 0.478), and the blackish-brown pigment
(baseline).
Figure 3. The dependence of the formation of Product-1 on heating temperature at 60 min under
hypoxic conditions. The degradation rate is calculated by (IVS1/3)/(IVS1/3 + IVS2/2)  100 and IV
values are obtained from 1H-NMR data.
Figure 4. The dependence of the formation of Product-1 on heating time. INM was heated at
200 °C for various times under hypoxic conditions. The degradation rate was calculated using
(IVS1/3)/(IVS1/3 + IVS2/2)  100, and IV values are obtained from 1H-NMR data.
Figure 5. The relationship between heating temperature and heating time for the formation of
amorphous INM devoid of Product-1, Product-2, and the blackish-brown pigments, under
hypoxic conditions.
Figure 6. Solubility of the γ-form (diamonds and gray line) and amorphous form (circles and
black line) of INM in citric acid buffer (pH 1.2). Error bars are standard deviations for n = 3,
except for those at 5, 15, and 90 min in γ INM (n = 2). The lines were constructed by curve
15

fitting using Excel solver module.
Scheme caption
Scheme 1. Mechanism for the heat-induced decarboxylation of INM.
16

1
Table 1. Molecular structures of INM and decarboxylated INM, and their H-NMR and
13C-NMR chemical shifts in chloroform-d1. The integrated values (IVs) were determined against
the 1H-NMR peaks of H-7, which correspond to one proton.
INM
Decarboxylated INM
3’
Cl
4’
3’
2’
1’
Cl
4’
2’
O
N
O
3”
1’
1
3”
8
5
2”
1
9
8
5
2”
7
6
10
9
4
N
7
6
10
13
CH3
2
13
H3C
CH3
H3C
2
4
3
O
OH
3
O
11
12
(11)
CH3
O
No.
1H-NMR
6.944(d)
IV
13C-NMR
No.
1H-NMR
6.892(d)
IV
13C-NMR
1
168.310
136.266
111.823
130.466
101.259
156.068
111.700
115.007
130.787
13.295
1
2
3
4
5
6
7
8
9
168.253
134.390
115.435
130.877
101.292
155.962
111.157
114.933
131.963
13.311
-
2
3
4
5
1.00
1.90
6
7
6.667(dd)
6.849(d)
1.00
1.01
6.661(dd)
6.892(d)
1.00
1.90
8
9
10
11
12
13
1'
2.380(s)
3.685(s)
2.91
2.02
10
2.311(s)
-
2.79
-
30.013
11
12
13
1'
176.768
55.731
-
3.820(s)
3.06
3.848(s)
3.13
55.731
133.789
131.042
129.034
138.874
8.770
133.798
131.198
129.149
139.351
2'(2'')
3'(3'')
4'
7.656(td)
7.462(td)
2.06
2.04
2'(2'')
3'(3'')
4'
7.641(t)
7.461(t)
2.22
2.08
(11)
2.192(s)
3.22
17

Figure 1
a)
b)
0 min 1 min 5 min 15 min 30 min 60 min
0 min
x 1.0k
100 μm
0 min
x 2.0k
30 μm
1 min
x 500 200 μm
60 min
x 1.0k
100 μm
18

Figure 2
Product-1
Product-2
INM
Blackish
19

Figure 3
35
30
25
20
15
10
5
y = 0.7236x - 115.85
R2 = 0.9759
0
140 150 160 170 180 190 200
Temperature ( C)
20

Figure 4
50
40
30
20
10
0
0
20
40
60
80
100
Time (min)
21

Figure 5
50
40
30
20
10
0
Tm
Degraded
Intact
Stable
140
150
160
170
180
190
200
Temperature ( C)
22

Figure 6
0.015
0.010
0.005
0
0
20 40 60 80 100 120 140
Time /min
23

Scheme 1
24

Indomethacin; NSAIDs
Cl
Cl
O
N
O
N
CH3
OH
CH3
H3C
H3C
O
O
CH3
Decarboxylation
O
Heat treatment under aerobic condition
Low
Time dependent
High
Impurity
Long
Short
Time
25