1
Reducing mechanical activation-induced amorphisation
of salbutamol sulphate by co-processing with selected
carboxylic acids
2
3
4
5
6
7
8
9
10
11
12
13
14
Vincent Curtin, Youness Amharar, Kieran H. Gallagher, Sarah Corcoran, Lidia Tajber,
Owen I. Corrigan and Anne Marie Healy*
15
(Grant No. 07/SRC/B1158 and Grant No. 12/IP/1408).
16
17
Abstract
18
The unintentional generation of amorphous character in crystalline active
19
pharmaceutical ingredients (APIs) is an adverse consequence of mechanical
20
activation during dosage form manufacture. In this study, we assess and compare
21
the ability of low glass transition temperature (Tg) dicarboxylic acids to mitigate
22
amorphisation of a model API, salbutamol sulphate (SS), on both co-milling and co-
23
mixing.
24
SS processed alone, as well as co-milled and co-mixed composites of the API with
25
glutaric acid (GA), adipic acid (AA) and pimelic acid (PA) were characterised by
26
powder X-ray diffraction (pXRD), differential scanning calorimetry (DSC) and dynamic
27
vapour sorption (DVS).
28
Milling and dry mixing of SS both resulted in pXRD amorphous materials. No
29
amorphous content of SS was detected by DVS on co-milling with 50 % w/w GA,
School of Pharmacy and Pharmaceutical Sciences, University of Dublin, Trinity
College, Dublin 2, Ireland.
Author Information: *University of Dublin, Trinity College, School of Pharmacy
and Pharmaceutical Sciences, Panoz Institute, Dublin 2, Ireland. E-mail:
healyam@tcd.ie. Tel: +353 (0) 1 896 1444.
Acknowledgement: This work was supported by the Science Foundation Ireland
1
30
while amorphisation was more than halved, relative to the API milled alone, on co-
31
milling with 50 % w/w AA and PA, respectively. Co-mixing with each excipient also
32
resulted in a decrease in API amorphicity, although the extent of reduction was
33
considerably less compared to the co-milling experiments.
34
The solubility (Solexcipient) of each excipient in amorphous SS was determined by
35
thermal methods. No further reduction in API amorphisation was achieved on co-
36
mixing with 50% w/w excipient, compared to concentrations corresponding to the
37
solubility of each excipient in the amorphous API (SolGA = 36%, SolAA=21%,
38
SolPA=22%). PXRD confirmed gradual dissolution over time of GA in amorphous SS
39
on co-mixing. In contrast to co-mixing, co-milling SS at excipient weight fractions
40
above their respective solubilities in the amorphous drug resulted in further
41
reductions in API amorphisation. This is thought to be due to the generation of a
42
molecular dispersion of amorphous API, supersaturated with excipient, thereby
43
leading to a more pronounced composite Tg lowering effect.
44
The results indicate that co-processing with low Tg excipients is an effective strategy
45
at minimising amorphisation of an API on mechanical activation.
46
Keywords Mechanical activation, amorphous, glass transition temperature,
47
salbutamol sulphate, crystalline.
48
49
50
51
52
53
54
55
2
56
57
1. Introduction
58
It is well recognised that subjecting drugs to routine pharmaceutical unit operations
59
can result in changes to the solid state properties of the material (Morris et al., 2001;
60
Lin et al., 2010). This can result in the transformation of a highly ordered crystalline
61
material to a disordered, high energy amorphous state (Feng et al., 2008; Patterson
62
et al., 2007). An amorphous system, despite its advantages of enhanced solubility
63
and dissolution rate, will be thermodynamically metastable relative to its crystalline
64
form and can revert back to the lower energy state with time. This can have serious
65
implications in terms of manufacturability, processability and stability (Hancock and
66
Zografi, 1997; Yu, 2001).
67
68
Mechanical activation is a term used to describe the increase in free energy of a solid
69
on mechanical processing, and is generally assumed to arise from operations such as
70
milling, grinding, compaction and compression, which results in breakage or fracture
71
of a solid (Hüttenrauch et al., 1985). However it was shown by Mosharraf and
72
Nystrom in 1999, that ‘gentler’ operations such as dry mixing can activate solids. The
73
authors noted a considerable change in the solid state structure of griseofulvin on
74
mixing, which also resulted in an increase in solubility. More recently, Hockerfelt et
75
al in 2009 reported that griseofulvin could be completely amorphised by dry
76
blending without particle breakage.
77
78
Milling, when compared to dry mixing, is a considerably more energetic process and
79
is known to amorphise several APIs (Gusseme et al., 2008; Willart et al., 2012; Willart
3
80
et al., 2001). Balani et al (2010) noted that amorphisation of milled salbutamol
81
sulphate could be reduced by co-milling the API with crystalline excipients, such as α
82
lactose monohydrate, adipic acid and magnesium stearate. They suggested that the
83
effects observed were associated with the presence of excipient acting as seed
84
crystals, coupled with a temperature increase from the milling operation.
85
demonstrated that co-milling sulfadimidine with a series of low glass transition
86
temperature (Tg) excipients was an effective strategy to minimise amorphisation of
87
the drug (Curtin et al., 2013). It was highlighted that excipients which showed good
88
solubility in the API exerted a Tg lowering effect, resulting in composite Tg values
89
lower than that of the API alone, which could mitigate amorphisation of the API. In
90
contrast, a crystalline excipient (malic acid) which displayed very poor solubility in
91
the API was found to promote amorphisation of the drug.
We
92
93
In this work we evaluate and compare, for the first time, the influence of two
94
mechanical operations on a model API co-processed with low Tg excipients. The
95
objective of the study was to compare the effect of two different modes of
96
mechanical activation (milling and dry mixing) on salbutamol sulphate (SS) and to see
97
if co-processing with low Tg excipients could prevent or minimise amorphisation of
98
the API. SS was chosen because of its high glass transition temperature (120 °C) and
99
its known propensity to amorphise when milled (Balani et al., 2010; Griesdale et al.,
100
2011; Brodka-Pfeiffer et al., 2003). The excipients chosen were glutaric acid, adipic
101
acid and pimelic acid based on their structural similarity and their low glass transition
102
temperatures according to the Tg=Tmx0.7 rule (Fukuoka et al., 1989; Kerc and Srcic,
103
1995). Their respective solubilites in amorphous SS were determined by thermal
4
104
methods. Freeze drying was used to produce amorphous API: excipient systems
105
which enabled composite Tg values to be determined. The milled and dry mixed API,
106
as well as the API:excipient co-processed systems were characterised with respect to
107
their X-ray diffraction, thermal and vapour sorption properties.
108
109
2. Materials and Methods
110
2.1. Materials. Salbutamol sulphate (SS) raw material (molar weight (Mw) = 576.70
111
g.mol-1) was purchased from Camida Ltd. (Clonmel, Ireland). Adipic acid (AA) (Mw =
112
146.14 g.mol-1), glutaric acid (GA) (Mw = 132.11 g.mol-1), and pimelic acid (PA) (Mw =
113
160.17 g.mol-1) were purchased from Sigma-Aldrich, Ireland.
114
diameter 5mm were purchased from Sigma-Aldrich, Ireland. Ethanol (99.5 %, v/v)
115
was purchased from Corcoran Chemicals, Ireland. Water, ultra-pure, was prepared
116
from a Millipore Elix advantage water purification system. Acetonitrile, HPLC grade,
117
was purchased from Fisher Scientific (Ireland). Triethylamine was purchased from
118
Sigma Aldrich, Ireland.
Glass beads of
119
120
2.2. Methods.
121
2.2.1. Milling. Ball milling of SS and excipients was performed with a PM 100 high
122
energy planetary mill (Retsch, Germany) at room temperature. 2.5 g of material was
123
placed in stainless steel milling jars of 50 cm3 volume with three stainless steel balls
124
of diameter 20 mm, corresponding to a ball to powder mass ratio of 40:1. The speed
125
of the solar disk was set at 400 rpm. Every 20 minutes of milling was followed by a
126
pause period of 10 minutes to avoid overheating (Curtin et al., 2013). Total milling
5
127
time was kept constant at 3 hours (h) corresponding to an effective milling time of 2
128
h. SS was co-milled with GA at different weight percentages of excipient (XGA = 5, 20,
129
35, 50 %). SS was also co-milled with 21 % and 50 % w/w AA, and 22 % and 50 %
130
w/w PA. API:excipient co-processed systems are named in accordance to the %
131
weight fraction of each component. For instance SS95:GA5 refers to a system
132
composition of 95% w/w SS and 5% w/w GA.
133
performed at least in duplicate.
All milling experiments were
134
135
2.2.2. Dry mixing
136
Dry mixing experiments were performed using a Turbula mixer operating at 64 rpm
137
(360 ml tubes, W.A. Bachofen, Switzerland) based on the method used previously by
138
Hockerfelt et al., 2009. The API and API:excipient fractions were mixed with glass
139
beads in the weight proportion 1:99 of powder to beads. A total powder content of
140
0.78 g was used for the dry mixing experiments. Experiments were performed at
141
different times, ranging from 2 h to 24 h. SS was dry mixed with each excipient at a
142
concentration corresponding to its solubility in the amorphous drug. The 24 h mixes
143
were also performed at 50 % w/w excipient. All mixing experiments were performed
144
at least in duplicate.
145
146
2.2.3. Freeze drying
147
API:excipient mixtures of various weight fractions ranging from 95:5 to 50:50 were
148
dissolved in deionised water and stirred for 24 hours at room temperature to yield
149
aqueous solutions with a concentration of 5% w/w total solid. The solutions were
150
then sonicated for 10 minutes and filtered through a 0.45 micron syringe filter to
6
151
ensure the absence of any residual crystals. The resulting aliquots (5 ml) were then
152
poured into 50 ml plastic tubes, which were immersed in liquid nitrogen for 15 min
153
and then loaded into a VirTis wide mouth filter seal glass flasks and attached to one
154
of the manifold ports of a benchtop VirTis 6K freeze-dryer model EL (SP Scientific,
155
USA). A Vacuum of 29-31 mtorr was obtained by the use of an Edwards 5 RV5 rotary
156
vane dual stage mechanical vacuum pump (Edwards, England). After 48 hours of
157
freeze drying the tubes were removed and capped. This was performed in duplicate
158
for all mixtures.
159
160
2.2.4. Spray drying
161
PA was spray dried as a solution in ethanol/water 7:3 v/v using a Buchi-290 Mini
162
Spray Dryer, as previously described (Curtin et al., 2013). The inlet temperature was
163
maintained at 78 °C, outlet temperature was between 42 – 49 °C and a feed
164
concentration of 0.4 % w/v was used. Spray drying was performed in the open mode
165
configuration using compressed air as the drying medium with a pump rate of 30 %
166
(8 ml/min), aspirator setting of 100 % and air flow of 40 mm (473 L/h).
167
168
2.2.5. Powder X-ray diffraction.
169
Powder X-ray diffraction (pXRD) measurements were performed on samples placed
170
on a low background silicon sample holder using a Rigaku Miniflex II desktop X-ray
171
diffractometer (Rigaku, Tokyo, Japan), as previously described (Curtin et al., 2013).
172
The pXRD patterns were recorded from 5 ° to 40 ° on the 2 theta scale at a step of
173
0.05 ° s-1. The X-ray tube composed of Cu anode (λCUKα = 1.54 Å), was operated
7
174
under a voltage of 30 kV and current of 15 mA. PXRD patterns were recorded at
175
least in duplicate.
176
For quantification of crystalline content by pXRD, the amorphous standard was
177
prepared by milling the API for 2 hours and the crystalline reference standard was
178
obtained by placing SS as received in the DVS at 90% RH until no mass loss events
179
were observed in the kinetic profile. The method employed was based on the
180
method reported by Clas et al. (1995). Zinc oxide was used as an internal standard to
181
determine if there were variations in peak position and intensity.
182
Different weight fractions of crystalline SS (Xc) (Xc = 0.1, 0.25, 0.5, 0.75, 0.9) were
183
prepared by mixing the relevant quantities of crystalline and amorphous reference
184
samples (total mixed sample weight 100 mg) with 10 mg Zinc oxide using an agate
185
mortar and pestle.
186
independent set of reference standards for each weight fraction. Each set of
187
reference standards was analysed by pXRD three times, giving a total of six results
188
for each weight fraction.
189
The pXRD peak used for the quantification of crystalline SS in each sample was at
190
approximately 15.3° 2 theta and was selected due to the lack of interference across
191
all diffraction patterns. Rigaku Peak Integral software was used in the determination
192
of peak intensity for each sample using the Sonnefelt-Visser background edit
193
procedure.
194
195
196
8
This procedure was repeated once to produce a second
197
2.2.6. Thermal analysis.
198
Differential scanning calorimetry (DSC) experiments were conducted using a Mettler
199
Toledo 821e with a refrigerated cooling system (LabPlant RP-100), as previously
200
described (Curtin et al., 2013). Nitrogen was used as the purge gas. Sealed 40 μl
201
aluminium pans with three vent holes were used throughout the study, and sample
202
weights varied between 5 and 13 mg. The system was calibrated for temperature
203
and cell constant using indium and zinc.
204
implemented in all DSC measurements. Analysis was carried out and monitored by
205
Mettler Toledo STARe software (version 6.10) with a Windows NT operating system.
A heating rate of 10 °C min-1 was
206
207
2.2.7. Solubility determination by thermal analysis.
208
The solubilities of GA, AA and PA in amorphous milled SS were estimated by thermal
209
analysis, as previously described by Curtin et al., 2013, and will be referred to as
210
Solexcipient hereafter (e.g. SolGA refers to solubility of GA).
211
API:crystalline excipient mixtures containing between 50% and 90% w/w of
212
crystalline excipient were manually mixed using a pestle and a mortar. The resulting
213
powders were then heated on a heating plate at 160 °C under nitrogen atmosphere
214
in order to prevent chemical degradation, subsequently quenched to 25 °C and kept
215
at this temperature for 2 days. At this stage a mixture of crystalline excipient and
216
amorphous API saturated with excipient was obtained, which was confirmed by
217
pXRD. The melting enthalpy of the resulting crystalline phase was then determined
218
by DSC at 20 °C min-1 from 25 °C to 180 °C for SS:AA and 130 °C for SS:GA and SS:PA,
219
respectively, and the average of two values plotted as a function of excipient mass
220
fraction.
9
Firstly, amorphous
221
2.2.8. Thermogravimetric Analysis (TGA)
222
Thermogravimetric analysis (TGA) experiments were conducted on a Mettler Toledo
223
TG 50 apparatus using a method previously described (Bianco et al., 2012). Weighed
224
samples (5–10 mg) were analysed in open aluminium pans placed on a Mettler MT5
225
balance. Samples were heated from 25 to 220 °C at a scanning rate of 10 °C/min
226
under nitrogen purge. Mass loss of the samples recorded were analysed by the
227
Mettler Toledo STARe software. Experiments were performed in duplicate.
228
229
2.2.9. Dynamic vapour sorption (DVS).
230
Sorption isotherms and kinetic profiles of the unprocessed, milled, mixed and co-
231
processed systems were obtained using DVS (Advantage, Surface Measurement
232
Systems, Alperton, UK). The temperature was 25.0 ± 0.1 °C and water was used as
233
the probe vapour. Co-milled samples were dried at 0 % RH and then subjected to
234
step changes of 10 % RH up to either 70 % or 90 % RH, and the reverse for
235
desorption. Second/third sorption cycle isotherms were also determined. Co-mixed
236
samples for each API:excipient system were initially exposed to step changes of 10 %
237
RH up to 90 % RH. Further samples were then simply exposed to humidities below
238
and above the threshold crystallisation RH, the humidity at which crystallisation and
239
mass loss first occurred on the initial DVS run. Final RH value was system specific
240
and was determined on the basis of when crystallisation was complete for milled,
241
mixed and co-processed SS. Absence of a mass loss in the second sorption cycles
242
indicated full crystallisation. The crystalline nature of the systems was confirmed by
243
pXRD. As previously described, the sample mass was allowed to reach equilibrium
244
defined as dm/dt ≤ 0.002 mg/min over 10 min, before the RH was changed (Curtin et
10
245
al., 2013).
246
calculations were performed using eq 1., as described previously by Balani et al.
247
2010.
248
Sample weights were between 7 – 15 mg.
% amorphous content = 100 x
Amorphous content
(1)
249
where Δm is the difference in the mass uptake (%) of the API between the first and
250
second sorption cycles at a system specific RH, ms is the sample mass in the DVS, md
251
is the mass of the API in the overall sample mass, and Δm100 is the difference in
252
mass uptake (%) between the first and second sorption cycles of the amorphous
253
milled standard at a system specific RH.
254
least in duplicate.
All DVS experiments were performed at
255
256
2.2.10. Melt quench
257
PA was melt quenched by melting the material in an aluminium pan in the DSC by
258
heating from 25 to 120 °C at 10 °C min-1 and rapidly immersing the molten material
259
in liquid nitrogen. The sample was subsequently reheated on a DSC from 25 to 120
260
°C at 10 °C min-1.
261
262
2.2.11. HPLC analysis
263
The HPLC method was adopted from that of Malkki et al., 1990. The analytical
264
column used was a LiChrosorb RP-18 (250 mm length, diameter 4 mm, particle size
265
10 µm S/N 009897). The column was attached to a Waters 1525 Binary Pump with
266
an in built degasser, a Waters 717 plus autosampler and a Waters 2487 Dual λ
267
Absorbance Detector. Samples were analysed using Breeze software Version 3.30.
11
268
The mobile phase consisted of 0.02M phosphate buffer containing 750 µl
269
triethylamine per 1 litre and acetonitrile (95/5 v/v). Phosphate buffer was adjusted
270
to pH=3.0 after adding triethylamine. Separation was carried out isocratically at
271
ambient temperature (about 22°C) and a flow rate of 1.5ml/min, with UV detection
272
at 265 nm. The measured retention time for SS was approximately 4.9 min (n=2).
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
12
292
3. Results
293
3.1
294
The solubilites of GA, AA and PA in the amorphous API at 25°C were determined by
295
thermal analysis, as described in section 2.2.7. A linear relationship was observed
296
between the composition and the enthalpy of fusion of the excipient for all three
297
systems and the Solexcipient was estimated as the zero enthalpy interception point by
298
extrapolation (Figure 1). The melting enthalpies of the pure excipients were omitted
299
from the plot.
300
‘amorphous API and crystalline excipient’ where the endothermic peak represents a
301
dissolution, to a system of ‘pure excipient‘ where the endothermic peak is an actual
302
melting event.
303
SolGA, SolAA and SolPA were 36 %, 21 % and 22 % respectively. It should be noted that
304
the method of preparation of the mixtures of crystalline excipient and amorphous
305
API saturated with excipient did not lead to drug degradation, as indicated by HPLC.
Solubility determination of GA, AA and PA in amorphous SS.
This is because it is not appropriate to compare a system of
306
307
3.2
308
of unprocessed, milled and dry mixed SS.
309
SS raw material displayed Bragg peaks characteristic of a crystalline material. The
310
API was identified as pure form 1 with characteristic diffraction peaks of this solid
311
state form apparent at 10.7o, 11.9o and 12.6o 2 theta (Palacio et al., 2006). SS, in
312
contrast, was amorphous when processed by either milling for 2 hours or dry mixing
313
for 8 hours (Figure 2).
314
The corresponding DSC thermograms are displayed in Figure 3. SS revealed a single
315
endothermic event with a temperature onset of 185 °C. Milled SS revealed a broad
13
Powder X- Ray diffraction, thermal behaviour and vapour sorption analysis
316
endothermic event between 30 – 100 °C believed to be due to release of loosely
317
bound moisture (mass loss of 2 % was noted by TGA analysis up to 100 °C). The glass
318
transition temperature of milled and dry mixed SS occurred at 120 °C. An exotherm
319
just prior to the melting event for the milled material could be due to crystallisation
320
of the amorphous drug. However the event corresponds to the temperature onset
321
of significant mass loss by TGA analysis (data not shown) and may also be due to
322
thermal degradation of the API.
323
The sorption kinetic profiles of the respective systems are shown in Figure 4. SS took
324
up around 0.3% mass at 50% RH followed by a series of sharp mass losses at higher
325
RH. The vapour uptakes for the second and third sorption cycles were comparable
326
with no mass loss events evident. The average % mass uptake at a specific RH from
327
the second and third cycles was then taken as the reference value for crystalline SS.
328
The initial mass losses are attributed to crystallisation events, associated with the API
329
raw material containing a small, but quantifiable amount of amorphous content.
330
Milled and dry mixed SS displayed very similar sorption properties. Both processed
331
samples lost nearly 4% mass during the initial drying step. This was preceded by a
332
mass uptake of nearly 9% at 60% RH in both cases, followed by a very sharp, steep
333
mass loss due to an amorphous to crystalline transformation of the drug. Further
334
mass losses at 70%, 80% and 90% RH indicated that crystallisation was not fully
335
complete after the initial mass loss at 60% RH. Second sorption cycles for both
336
milled and mixed API were almost superimposable and no mass loss here indicated
337
that samples had fully crystallised. The quantification results are independent of the
338
crystallisation kinetics, provided the method allows the sample to reach equilibrium
339
at each partial pressure and the sample has crystallised (Mackin et al., 2002). For all
14
340
mono and binary systems analysed, crystallisation was complete before the start of
341
the second sorption cycle.
342
determined by comparing the difference in % mass uptake between the first and
343
second sorption cycles. Comparing the % mass uptake at 50 % RH for crystalline and
344
amorphous milled SS, with that of the drug as received revealed an amorphous
345
content of 4% ± 1% for SS raw material (Table 1).
Amorphous content quantification by DVS was
346
347
3.3
348
The pXRD patterns of the SS:GA co-milled composites are displayed in Figure 5. Co-
349
milling with just 5% w/w GA, a concentration well below SolGA of 36 %, resulted in an
350
amorphous halo similar to that for the API milled alone. Processing at excipient
351
concentrations close to, at, and above the SolGA revealed Bragg peaks specific to the
352
API and indicated that the degree of amorphisation of SS was reduced.
353
The DSC thermograms for the co-milled systems are displayed in Figure 6. An
354
exotherm at around 125 °C was observed for the SS95:GA5 co-milled system which
355
was attributed to crystallisation of the amorphous API.
356
exothermic peak was present in the thermogram of the API milled alone, suggesting
357
that co-milling the drug with just 5 % w/w GA generated a less thermally stable
358
system that that of API milled on its own. No crystallisation exotherms were noted
359
for the co-milled systems at higher weight fractions of excipient, as expected from
360
the pXRD data.
361
polymorphic conversion in the excipient followed by its melting (McNamara et al.,
362
2006). The binary system displayed a eutectic-like behaviour very similar to that
363
reported for sulfadimidine co-milled with GA (Curtin et al, 2013). The melting
15
SS co-milled with GA
No corresponding
The thermal events between 75 – 100 °C correspond to a
364
endotherm is at approx 90° C for all co-milled systems and is at a lower temperature
365
when compared to the melting of either the excipient (99 °C) or drug (185 °C). The
366
melting of excess API at higher temperature is distorted by the simultaneous thermal
367
degradation of the drug.
368
Isotherm plots of SS milled as well as SS:GA co-milled composites are displayed in
369
Figure 7. The SS95:GA5 system crystallised at a lower RH compared to the API milled
370
alone, indicative of a less physically stable system. The mass uptake at 40 % RH was
371
used in the quantification calculations and from Figure 7 it can be seen that uptakes
372
were lower for the co-milled systems compared to the API milled alone.
373
PXRD was used as complementary quantification technique. The calibration curve of
374
peak intensity versus crystalline SS weight fraction (Xc) is displayed in Figure 8 and
375
quantification results are listed in Table 1. The limit of detection was determined by
376
the IUPAC method, as reported by Long and Winefordner (1983), and corresponded
377
to a crystalline weight fraction of 0.1. The limit of quantification was calculated, also
378
by the IUPAC method, to be 0.18.
379
380
Table 1. Summary of amorphous content of SS raw material, SS milled and SS:GA co-
381
milled systems as quantified by DVS and pXRD.
System
% Amorphous content (±S.D.)
DVS
pXRD
SS
4 (1)
0
SS milled
100
100
SS95:GA5
88 (3)
100
SS80:GA20
13 (6)
0
16
SS65:GA35
8 (4)
0
SS50:GA50
0
0
382
383
3.4
384
SS was co-milled at compositions corresponding to SolAA and SolPA (21 % and 22 %
385
w/w), and above (50 % w/w). The pXRD patterns of the co-milled systems are
386
displayed in Figure 9 and the API amorphous content values, as quantified by DVS
387
using eq. 1, and pXRD, are displayed in Table 2. The presence of Bragg peaks in the
388
co-milled systems, corresponding to the crystalline API, indicated that co-milling with
389
these excipients was effective at reducing the degree of amorphisation of the milled
390
drug (Fig. 9). Interestingly, co-milling at weight fractions of excipient above Solexcipient
391
produced a further reduction in API amorphicity, as indicated by both pXRD and DVS
392
(Table 2).
SS co-milled with AA and PA
393
394
Table 2. Summary of amorphous content of SS:AA and SS:PA co-milled systems as
395
quantified by DVS and pXRD.
% Amorphous content (± S.D.)
System
DVS
pXRD
SS79 : AA21
63 (2)
75 (7)
SS50 : AA50
46 (5)
62 (12)
SS78 : PA22
64 (2)
60 (15)
SS50 : PA50
42 (4)
48 (10)
396
397
398
3.5
17
SS co-mixed with the dicarboxylic acids.
399
SS was co-mixed separately with the three dicarboxylic acids, at compositions
400
corresponding to Solexcipient and for varying times. The amorphous content of the co-
401
mixed composites were quantified by DVS. The kinetic profiles for all the three
402
systems, after dry mixing for 24 h, are displayed in Figure 10 below.
403
404
The mass loss at 70% RH for the SS:GA dry mixed composite indicated that co-mixing
405
did not completely eliminate amorphisation of the API.
406
subsequent mass losses on crystallisation were higher for the SS:AA and SS:PA
407
systems indicating that co-mixing with these excipients was less effective than co-
408
milling at mitigating API amorphisation.
Mass uptakes and
409
410
A plot of amorphous content versus dry mixing time is shown in Figure 11. SS dry
411
mixed for 4 h was 93% amorphised, and was completely amorphous after 8 h, as
412
mentioned earlier. Co-mixing experiments were performed for as long as 24 h until a
413
plateau was observed whereby no further changes in API amorphisation with mixing
414
time were noted. After 24 h, co-mixing SS with PA had a negligible effect on
415
amorphisation of the API, co-mixing with AA lowered the amorphous content by 8 %
416
± 2 %, whereas the co-mixed API composite with GA had an amorphous content
417
reduction of more than 75 %, relative to the API dry mixed alone.
418
419
4. Discussion
420
Milling and dry mixing are two mechanical operations which, although differing
421
considerably in terms of energy input, were capable of completely amorphising SS.
18
422
The dry mixing process took considerably longer than milling (8h v 2h), most
423
probably due to the milder conditions to which the API particles are exposed.
424
The excipients chosen for this study were three dicarboxylic acids which are
425
crystalline in nature. GA and AA were resistant to amorphisation when exposed to
426
milling, spray drying or melt quench (Curtin et al., 2013). In previous work, the Tg
427
values of GA and AA were calculated by the Tg = Tm x 0.7 rule to be -14 °C and 25 °C
428
respectively (Curtin et al., 2013). Moreover these two excipients were proven to be
429
very effective at mitigating amorphisation of sulfadimidine on co-milling (Curtin et
430
al., 2013).
431
When comparing the two sets of quantification results by DVS and pXRD in Table 1, it
432
should be acknowledged that, with pXRD it is difficult to detect amorphous content
433
below 10 %, whereas in the case of DVS, the detection limit is often lower than 1 %
434
(Lehto et al., 2006). For instance, quantification results based on pXRD indicated
435
that API amorphisation was eliminated on co-milling with just 20% w/w GA, in
436
contrast to DVS which detected an amorphous content of 13%.
437
We noted a relationship between the solubility of excipient in the API, the calculated
438
solubility parameter and the degree of amorphisation reduction on co-milling. PA is
439
another dicarboxylic acid, differing only in alkyl chain length from that of GA and AA
440
and has a calculated Tg of -8 °C. PA could not be amorphised when milled, spray
441
dried or melt quenched (data not shown). All three excipients were found to have
442
good solubility in amorphous SS, with AA and PA having similar values of 21 and 22
443
%, respectively, with GA being more soluble, at 36 %.
444
19
445
SS cannot be melted without degradation, therefore freeze drying was used to
446
produce amorphous composites for each API:excipient system. Figure 12 displays a
447
plot of experimental Tg of freeze dried samples and % crystallinity of co-milled
448
systems as a function of GA weight fraction. With XGA < SolGA the composite systems
449
were amorphous. With XGA > SolGA crystallinity associated with the excipient was
450
observed. It should be noted that the melting enthalpy measured for the SS:GA
451
composite, at XGA=50, was in good agreement with the value predicted from the plot
452
presented in Figure 1. In accordance to what was reported in the study with
453
sulfadimidine (Curtin et al., 2013), the crystallinity of the SS:GA co-milled systems
454
approached near maximum and then levelled off at excipient weight fractions very
455
close to SolGA. The freeze dried amorphous composites, at SolGA, had a Tg minimum
456
value of 5 °C. Amorphous composites at a composition corresponding to SolGA would
457
be saturated with excipient and therefore exhibit the lowest possible composite Tg.
458
Increasing the amount of excipient should have no further effect on the Tg and this is
459
why the Tg curve flattens when XGA > SolGA.
460
461
As previously reported, the identification of a single Tg value intermediate to that of
462
the API and excipient was taken as an indicator of a mixed, single phase system (Qian
463
et al., 2010). In contrast, a phase separated amorphous system will display two
464
distinct Tg values. Saunders et al. (2004) noted that a Tg could be detected for a
465
binary system containing as low as 1 % amorphous lactose by using very high
466
scanning rates. As GA, AA and PA are not amenable to amorphisation on their own,
467
in this particular case phase separation of API and excipient in the amorphous state
468
is unlikely. The miscibility of the SS:GA, SS:AA, and SS:PA systems were confirmed by
20
469
the presence of a single Tg value in the freeze dried amorphous samples. At
470
Solexcipient, the SS:AA system had a Tg minimum value of 34 °C and the SS:PA system
471
had a Tg minimum value of 31 °C, respectively, indicating that only GA was capable
472
of producing a composite with an intermediate Tg below that of room temperature
473
(5°C) (Table 3). GA had the lowest calculated Tg value, it was the most soluble of the
474
excipients in the API, it produced the lowest composite Tg on freeze drying, and it
475
proved to be the most effective excipient at mitigating amorphisation on both co-
476
milling and co-mixing.
477
478
It was anticipated that maximal amorphisation reduction of SS on co-milling would
479
be achieved at excipient weight fractions corresponding to SolAA (21%) and SolPA
480
(22%). However this was not the case and, as indicated in Table 2 and 3, a further
481
amorphous content reduction of more than 15 % was observed by DVS, in both
482
cases, on co-milling with 50 % w/w AA and PA. The purpose of using pXRD as a
483
quantification technique in this study was to support the results of the analysis by
484
DVS and demonstrate that the reduction in API amorphicity on co-milling at excipient
485
weight fractions beyond Solexcipient could indeed be observed from an orthogonal
486
method. Results from both DVS and pXRD illustrated that such a reduction in API
487
amorphous content was observed (Table 2). Balani et al. 2010, previously co-milled
488
SS with AA at different weight fractions of excipient and our quantification results by
489
DVS are in agreement with the quantification results reported in that study.
490
Moreover the authors noted an excipient concentration dependent effect on the
491
amorphous content of the API on co-milling. A possible explanation is that the co-
492
milling operation leads to amorphous SS, supersaturated with excipient, which
21
493
conceptually could lead to a greater composite Tg reduction and promote further
494
crystallisation from the amorphous state.
495
amorphous dispersion by co-milling was also extensively reported by Mahieu et al.,
496
2013; Caron et al., 2007; Willart et al., 2008.
The production of a supersaturated
497
498
Table 3. Summary of experimental solubilites, Tg minimum values of freeze dried
499
amorphous systems, and amorphous content values, as quantified by DVS, for co-
500
milled and co-mixed API:excipient systems. Xexcipient refers to the weight fraction of
501
excipient.
% Amorphous content by DVS (± S.D)
Tg
System
Solexcipient (%)
minimum
(°C)
Co-milling (3 h)
Co-mixing (24 h)
Xexcipient =
Xexcipient =
Xexcipient =
Xexcipient =
Solexcipient
50 % w/w
Solexcipient
50 % w/w
SS:GA
36
5
8 (4)
0
22 (9)
21 (4)
SS:AA
21
34
63 (2)
46 (5)
92 (2)
87 (4)
SS:PA
22
31
64 (2)
42 (4)
100
99 (1)
502
503
The reduction of API amorphisation by co-mixing was less pronounced than by co-
504
milling. On co-mixing for 24h at compositions corresponding to Solexcipient, the API
505
was still almost 100 % amorphous with PA, while AA and GA lowered API amorphous
506
content by 8 % and 78 %, respectively (Table 3). In order to assess whether a
507
supersaturation effect occurred on co-mixing, as was observed for co-milling,
508
experiments were performed at excipient weight fractions of 50 %w/w. However no
509
difference in API amorphous content on co-mixing at the two different
22
510
concentrations of each excipient was noted (Table 3). Dry mixing is a much milder
511
mechanical operation than milling and with less energy input the likelihood of
512
obtaining a supersaturated state is less.
513
514
Amorphous SS, prepared by dry mixing, was then mixed with 35% w/w crystalline GA
515
and samples were periodically removed over time and evaluated by pXRD. It was
516
noted that Bragg peaks of excipient became less intense initially due to dissolution of
517
the excipient in the amorphous API (Figure 13).
518
peak intensity was noted when the excipient was dry mixed alone (data not shown).
519
After 3 h the system had recrystallised (Figure 13d). This supports the hypothesis
520
that, upon mechanical activation, initial amorphisation of the API is preceded by
521
dissolution of the excipient, which is solubility dependent, and acts to lower the
522
composite Tg and promote crystallisation of the drug.
523
524
525
526
527
528
529
530
531
532
533
23
No comparable reduction in Bragg
534
535
5. Conclusion
536
Both milling and dry mixing SS resulted in a crystalline to amorphous solid state
537
change in the API. We have shown in this work that amorphisation of SS could be
538
effectively eliminated on co-milling with 50% w/w GA, while it could be more than
539
halved on co-milling with 50% w/w AA and PA. Dry mixing of SS with these low Tg
540
excipients also resulted in a less disordered API, although the extent of reduction
541
was less pronounced relative to the corresponding co-milled data. GA was the most
542
effective excipient at mitigating API amorphisation on both mechanical co-processing
543
techniques. It displayed the highest solubility of the excipients in the amorphous
544
drug, and it exerted the most pronounced Tg lowering effect in the amorphous
545
composites. The reduction in API amorphisation on co-mixing achieved at excipient
546
concentrations corresponding to their solubility in the amorphous drug was the
547
same compared to systems with 50% w/w excipient. In contrast, co-milling at the
548
higher concentration of excipient produced a further reduction in API amorphisation.
549
This is thought to be due to a supersaturation effect arising from the more energetic
550
milling process, enabling more excipient to further lower the composite Tg, and
551
promote crystallisation. We have shown that initial amorphisation of the API is
552
preceded by dissolution of the excipient which induces a Tg lowering effect on the
553
resultant composite and induces crystallisation of the drug.
554
low glass transition temperature dicarboxylic acids via two different modes of
555
mechanical activation proved effective at mitigating amorphisation of the API.
556
557
24
Co-processing SS with
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
6.
References
Balani, P. N.; Ng, W. K.; Tan, R. B. H.; Chan, S. Y. Influence of excipients in
comilling on mitigating milling-induced amorphisation or structural disorder
of crystalline pharmaceutical actives. J. Pharm. Sci. 2010, 99 (5), 2462−2474.
Bianco, S.; Caron, V.; Tajber, L.; Corrigan, O. I.; Nolan, L.; Hu, Y.; Healy, A. M.
Modification of the Solid-State Nature of Sulfathiazole and Sulfathiazole
Sodium by Spray Drying. AAPS PharmSciTech. 2012, 13 (2), 647-660.
Brodka-Pfeiffer, K.; Langguth, P.; Hausler, H.; Graβ, P. Influence of mechanical
activation on the physical stability of salbutamol sulphate. Eur. J. Pharm.
Biopharm. 2003, 56 (1), 392−400.
Caron, V.; Willart, J. F.; Danede, F.; Descamps, M. The implication of the glass
transition in the formation of trehalose/ mannitol molecular alloys by .ball
milling. Solid State Commun. 2007, 144 (7−8), 288−292.
Clas, S. D.; Faizer, R.; O’Connor, R. E.; Vadas, E. B. Quantification of
crystallinity in blends of lyophilized and crystalline MK-0591 using X-ray
powder diffraction. Int. J. Pharm. 1995, 121, 73-79.
Curtin, V.; Amharar, Y.; Hu, Y.; Erxleben, A.; McArdle, P.; Caron, V.; Tajber, L.;
Corrigan, O. I.; Healy, A. M. Investigation of the Capacity of Low Glass
Transition Temperature Excipients to Minimize Amorphization of
Sulfadimidine on Comilling. Mol. Pharmaceutics 2013, 10, 386−396.
Feng, T.; Rodolfo, P.; Carvajal, M. T. Process induced disorder in crystalline
materials: Differentiating defective crystals from the amorphous form of
griseofulvin. J. Pharm. Sci. 2008, 97 (8), 3207–3221.
Fukuoka, E.; Makita, M.; Yamamura, S. Glass state of pharmaceuticals III.
Thermal properties and stability of glassy pharmaceuticals and their binary
glass systems. Chem. Pharm. Bull. 1989, 37, 1047−1050.
Griesdale, L. C.; Jamieson, M. J.; Belton, P. S.; Barker, S. A.; Craig, D. Q. M.
Characterization and quantification of amorphous material in milled and
spray-dried salbutamol sulfate: A comparison of thermal, spectroscopic, and
water vapor sorption approaches. J. Pharm. Sci. 2011, 100 (8), 3114-3129.
Gusseme, A. D.; Neves, C.; Willart, J. F.; Rameau, A.; Descamps, M. Ordering
and disordering of molecular solids upon mechanical milling: the case of
fanaserine. J. Pharm. Sci. 2008, 97 (11), 5000−5012.
25
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
Hancock, B. C.; Zografi, G. Characteristics and significance of the amorphous
state in pharmaceutical systems. J. Pharm. Sci. 1997, 86 (1), 1−12.
Hockerfelt, M. H.; Nystrom, C.; Alderborn, G. Dry mixing transformed microparticles of a drug from a highly crystalline to a highly amorphous state.
Pharm. Dev. Technol. 2009, 14 (3), 233–239.
Hüttenrauch, R.; Fricke, S.; Zielke, P. Mechanical activation of pharmaceutical
systems. Pharm Res. 1985, 2 (6), 302–306.
Kerc, J.; Srcic, S. Thermal analysis of glassy pharmaceuticals. Thermochim.
Acta 1995, 248, 81−95.
Lehto, V. P.; Tenho, M.; Heikkilä, K. V.; Harjunen, P.; Päällysaho, M.; Välisaari,
J.; Niemelä, P.; Järvinen, K. The comparison of seven different methods to
quantify the amorphous content of spray dried lactose. Powder Technol.
1006, 167, 85-93.
Lin, S. Y.; Hsu, C. H.; Ke, W. T. Solid-state transformation of different
gabapentin polymorphs upon milling and co-milling. Int. J. Pharm. 2010, 396
(1–2), 83–90.
Long, G. L.; Winefordner, J. D. Limit of detection: A closer look at the IUPAC
definition. Anal. Chem. 1983, 55 (7), 713-724.
Mackin, L.; Zanon, R.; Park, J. M.; Foster, K.; Opalenik, H.; Demonte, M.
Quantification of low levels (<10%) of amorphous content in micronized
active batches using dynamic vapour sorption and isothermal
microcalorimetry. Int. J. Pharm. 2002, 231 (2), 227−236.
Mahieu, A.; Willart, J. F.; Dudognon, E.; Danede, F.; Descamps, M. A New
Protocol To Determine the Solubility of Drugs into Polymer Matrixes. Mol.
Pharmaceutics 2013, 10, 560−566.
Malkki, L.; Tammilehto, S. Decomposition of salbutamol in aqueous
solutions. I. The effect of pH, temperature and drug concentration. Int. J.
Pharm. 1990, 63 (1), 17-22.
McNamara, D. P.; Childs, S. L.; Giordano, J.; Iarriccio, A.; Cassidy, J.; Shet, M.
S.; Mannion, R.; O’Donnell, E.; Park, A. Use of a glutaric acid cocrystal to
improve oral bioavailability of a low solubility API. Pharm. Res. 2006, 23 (8),
1888−1897.
Morris, K. R.; Griesser, U. J.; Eckhardt, C. J.; Stowell, J. G. Theoretical
approaches to physical transformations of active pharmaceutical ingredients
during manufacturing processes. Adv Drug Deliv Rev. 2001;48 (1), 91–114.
26
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
Mosharraf M, Nyström C. The effect of dry mixing on the apparent
solubility of hydrophobic, sparingly soluble drugs. Eur. J. Pharm. Sci. 1999, 9
(2), 145–156.
Palacio, M. A.; Cuffini, S.; Badini, R. ; Karlsson, A. ; Palacios, S. M. Solid-state
characterization of two polymorphic forms of R-albuterol sulfate. J. Pharm.
Biomed. Anal. 2007, 43 (4), 1531-1534.
Patterson, J. E.; James, M. B.; Forster, A.H.; Lancaster, R. W.; Butler, J. M.;
Rades, T. Preparation of glass solutions of three poorly water soluble drugs
by spray drying, melt extrusion and ball milling. Int. J. Pharm. 2007, 336 (1),
22-34.
Qian, F.; Huang, J.; Hussain, M. A. F. Drug–polymer solubility an miscibility:
stability consideration and practical challenges in amorphous solid
dispersion development. J. Pharm. Sci. 99 (2010) 2941–2947.
Saunders, M.; Podluii, K.; Shergill, S.; Buckton, G.; Royall, P. The potential of
high speed DSC (Hyper-DSC) for the detection and quantification of small
amounts of amorphous content in predominantly crystalline samples. Int. J.
Pharm. 2004, 274, 35-40.
Shah, B.; Kakumanu, K.; Bansal, A, K. Analytical techniques for quantification
of amorphous/crystalline phases in pharmaceutical solids. J. Pharm. Sci.
2006, 95 (8), 1641-1665.
Willart, J. F.; Gusseme, A. D.; Hemon, S.; Odou, G.; Danede, F.; Descamps, M.
Direct crystal to glass transformation of trehalose induced by ball milling.
Solid State Commun. 2001, 119 (8-9), 501-505.
Willart, J. F.; Descamps, M. Solid state amorphisation of pharmaceuticals.
Mol. Pharmaceutics 2008, 5 (6), 905−920.
Willart, J. F.; Carpentier, L.; Danede, F.; Descamps, M. Solid state vitrification
of crystalline griseofulvin by mechanical milling. J. Pharm. Sci. 2012, 101 (4),
1570−1577.
Yu, L. Amorphous pharmaceutical solids: preparation, characterization and
stabilisation. Adv. Drug Delivery Rev. 2001, 48 (1), 27−42.
693
694
27