Sie sind auf Seite 1von 8

APT 1250 No.

of Pages 8, Model 5G
2 May 2016

Advanced Powder Technology xxx (2016) xxx–xxx


Contents lists available at ScienceDirect

Advanced Powder Technology

journal homepage:

2 Original Research Paper

4 Fast and scalable preparation of starch nanoparticles by stirred media
5 milling
8 Chetankumar M. Patel, Mousumi Chakraborty, Z.V.P. Murthy ⇑
9 Department of Chemical Engineering, S. V. National Institute of Technology, Surat 395007, Gujarat, India

a r t i c l e i n f o a b s t r a c t
2 4
15 Article history: A study on the preparation of starch nanoparticles (SNPs) by stirred media milling was conducted in the 27
16 Received 15 January 2016 present work. The effects of milling time and solids concentration on the particle size were investigated. 28
17 Received in revised form 19 April 2016 It was found that the particle size of native starch was reduced from 3–20 lm to around 245 nm in 29
18 Accepted 25 April 2016
90 min by stirred media milling. The stability analysis of SNPs carried out by dynamic light scattering 30
19 Available online xxxx
(DLS) technique revealed good stability over a period of one week. The morphology of prepared starch 31
nanoparticles was characterized by field emission gun-scanning electron microscopy (FEG-SEM), 32
20 Keywords:
Transmission electron microscopy (TEM). The structural changes on SNPs were investigated by wide- 33
21 Stirred media milling
22 Starch nanoparticles
angle X-ray diffraction (WAXD) and Fourier transform infrared spectroscopy (FTIR). The stirred media 34
23 Nanosuspension milling was found to produce SNPs effectively without the use of any surfactants or additives. The effec- 35
24 Stability tive, low cost, and easily scalable physical method for the production of starch nanoparticles was 36
25 presented. 37
Ó 2016 Published by Elsevier B.V. on behalf of The Society of Powder Technology Japan. All rights 38
reserved. 39

43 1. Introduction nanoparticles of different types [10–14]. Recently bio-based nano- 63
materials such as nanocellulose, nanochitin and nanostarch have 64
44 Starch is a major source of carbohydrates and the great reserves received greater interest due to their environmentally friendly nat- 65
45 of polysaccharide occur in the different plants. It is considered a ure [15]. Starch nanoparticles are expected to play a major role 66
46 polymeric carbohydrate consisting of anhydrous glucose units because of their significantly different and unique properties than 67
47 linked primarily through a-D-(1-4) glycosidic bonds. A starch basi- their bulk counterpart. A comprehensive review of starch nanopar- 68
48 cally consists of two molecules – the linear and helical amylose ticles preparation, characterization, and applications was reported 69
49 (20–30%) with molecular weight of 106, and the branched amy- by Le Corre et al. [16]. The authors reported three widely used 70
50 lopectin (70–80%) with a molecular weight of about 1010 [1,2]. methods for preparation of starch nanoparticles, viz., acid hydroly- 71
51 The starch occurs as a semi-crystalline microparticles and has a sis, regeneration, and mechanical method. The preparation of 72
52 semi-crystalline structure with crystallinity in the range of starch nanoparticles was reported by different mechanical meth- 73
53 15–45% [3]. It is abundantly available low cost natural biopolymer, ods such as an environmentally friendly, high pressure homoge- 74
54 which has been widely employed as the most important ingredi- nization [17], high-pressure homogenization and miniemulsion 75
55 ents in many industries such as food, medicine, textile, and chem- cross-linking technique [18], and by high-intensity ultrasonication 76
56 ical industries [4]. Starch and its products are non-toxic, method [19]. 77
57 biocompatible, biodegradable, renewable, inexpensive, and widely Nanomilling in the stirred media mills is an efficient process for 78
58 available; makes them promising alternative in many applications the preparation of ultrafine materials owing to its advantageous 79
59 [5,6]. features, viz., ease of operation, simple construction, high size 80
60 Newer and advanced applications of starch have been continu- reduction rate, and low wear contamination [20]. Stirred media 81
61 ally explored in the last decade [6–9]. The starch has been widely mills have been successfully employed for the preparation of dif- 82
62 used polysaccharides for synthesis and stabilizing agent for ferent types of nanoparticles [21–25]. This technique has huge 83
potential, which is evident from the commercialization of several 84
⇑ Corresponding author. Tel.: +91 2612201641, +91 2612201648; fax: +91 poor water soluble nanodrugs by the nanomilling method [26]. 85
2612227334. The grinding technique is important in the processing of biomate- 86
E-mail addresses:, rials, for example, particle size reduction is a first step in the 87
(Z.V.P. Murthy).
0921-8831/Ó 2016 Published by Elsevier B.V. on behalf of The Society of Powder Technology Japan. All rights reserved.

Please cite this article in press as: C.M. Patel et al., Fast and scalable preparation of starch nanoparticles by stirred media milling, Advanced Powder Tech-
nology (2016),
APT 1250 No. of Pages 8, Model 5G
2 May 2016

2 C.M. Patel et al. / Advanced Powder Technology xxx (2016) xxx–xxx

88 biomass conversion [27]. Fewer research reports were available in experiment was also conducted in a vertical laboratory stirred 148
89 the area of preparation of bio-based nanomaterials by stirred media mill operating in batch mode. The mill consists of stainless 149
90 media milling. Several relevant reports on the preparation of sub- steel vessel of 0.5 L capacity with cooling water jacket and a pin 150
91 micron and nanosize biomaterials by stirred media mills were type stirrer with four cylindrical rods placed at right angles. Milling 151
92 found in the literature such as preparation of nanoparticles of experiment was carried out at a stirrer speed of 1200 rpm, a grind- 152
93 starch [28], cellulose [29], chitosan [30]. The objective of the pre- ing media size of 0.4 mm, solid mass fraction of 0.1 and total 1 kg 153
94 sent work is to explore the feasibility of the production of starch of grinding media were used in batch mode operation. More details 154
95 nanoparticles on the large-scale by wet stirred media milling of this mill system can be found in our previous work [32]. 155
96 method. The effects of milling time and solids concentration on
97 the quality of prepared starch nanoparticles were investigated. Sta-
2.3. Characterization 156
98 bility of prepared SNPs was also evaluated. In addition, changes in
99 the structural properties of starch nanoparticles were investigated
Particle size measurements of the ground sample were carried 157
100 by SEM, TEM, XRD, and FTIR analysis.
out with a dynamic light scattering (DLS) technique (Malvern Zeta- 158
sizer Nano ZS90, UK). All the measurements were performed at 159
101 2. Materials and methods 25 °C temperature with a measurement angle of 90°. A disposable 160
polystyrene cuvettes with 4 optical sides were utilized for the size 161
102 2.1. Materials measurements. The sample volume of around 0.5 mL was collected 162
after regular time intervals and diluted with around 40 mL of Mil- 163
103 The maize starch was purchased from M/s. National chemical lipore water. The samples were then degassed by sonication for 164
104 laboratory, Vadodara, India. The particle size of the native starch 1 min using an ultrasonic probe (Ultrasonic, Germany) prior to par- 165
105 sample range was between 3 lm and 20 lm (see Section 3.5). All ticle size analysis. An average value of particle size and the polydis- 166
106 the suspensions were prepared in the Millipore water (Millipore, persity index (PdI) was reported from three measurement runs. 167
107 Elix, Bangalore, India). High density (6000 kg/m3) yttria stabilized The starch nanosuspension was taken with mannitol (5%, w/v), 168
108 zirconia (ZrO2) grinding media (chemical composition: 93% ZrO2, which was used as a cryoprotectant in lyophilizer Alpha 1-4 LD 169
109 5% Y2O3 and 2% others); purchased from M/s. Saint Gobain ZirPro, plus (Martin Christ, Germany). The suspension is placed in a round 170
110 France, was used for the nanogrinding experiments. bottom flask and freeze dried at 45 °C. The lyophilized powder 171
was dispersed into water using an ultrasonic probe for particle size 172
analysis by DLS technique. The Scanning Electron Microscopy 173
111 2.2. Experimental procedure
(SEM) images were obtained with a S-3400N (HITACHI, JAPAN). A 174
small amount of milled starch suspension was drop casted on a 175
112 The wet nanomilling experiments were conducted in a Dyno-
clean silicon wafer and dried under vacuum. The samples were 176
113 mill (Research laboratory mill, WAB, Switzerland) operating in a
then coated with a gold layer by Ion sputtering using E-1010 177
114 recirculation mode. The mill consists of a silicon carbide chamber
(HITACHI, JAPAN). TEM images were obtained with a Tecnai-20 178
115 (80 mL capacity) and a hardened chrome alloy DYNOÒ-
(Philips, Netherlands), which at accelerating voltage of 200 kV pro- 179
116 Accelerator. The grinding media and suspension of initial starch
vides 0.27 nm point resolution. The drop of nanosuspension was 180
117 powder prepared in the MilliporeÒ water were charged into the
placed on the carbon coated copper grid and dried at room temper- 181
118 mill. More details of the mill used in the present work are given
ature before analysis. The XRD patterns of the freeze dried powder 182
119 in our previous work [31]. The pre-suspension of starch was pre-
were recorded using D8 Advanced Diffractometer (Bruker AXS, 183
120 pared by mixing required amount of maize starch in water. The
Germany) with Cu Ka (k = 1.5406 Å) at scan speed of 0.25°/s with 184
121 suspension was sonicated by the ultrasonic probe (Ultrasonic, Ger-
0.05° step size for 2h ranging from 5° to 35°. The FTIR spectra 185
122 many) for 3 min at 35% amplitude to uniformly disperse the starch
(450–4000 cm 1) of the native starch and SNPs powder presented 186
123 particles into water before milling. This pre-suspension was used
on KBr films were acquired on a Spectrum BX FT spectrophotome- 187
124 as feed for the media milling experiment. The milling experiment
ter (Perkin Elmer, Germany). 188
125 in Run-5 was started with lower solids loading, and solids concen-
126 tration was gradually increased thereafter. This was accomplished
127 by dispersing fifteen grams of starch into 100 mL of water. The mill 3. Results and discussion 189
128 charge was prepared by mixing 60 mL of this starch slurry with
129 50 mL of water, and the experiment was started. The remaining 3.1. Preparation of starch nanosuspension by stirred media milling 190
130 40 mL of the original starch slurry was fed to the mill at every
131 10 min in equal volume of 10 mL. The sample volume of around The size reduction results from the compression and shear 191
132 0.5 mL was collected after regular time intervals for particle size stresses induced by high speed moving grinding media with an agi- 192
133 analysis. Milling experiments were carried out at a stirrer speed tator. The quality and product particle sizes obtained from a 193
134 of 3500 rpm, a grinding media size of 0.4 mm, and with a grinding milling process are determined by the number of stress events 194
135 media filling ratio of 70% (v/v). The selection of stirrer speed was undergone by the particles (stress number, SN) and the stress 195
136 based on the requirement of maintenance of the minimum flow intensity (SI) or stress energy (SE) at each stress event. Thus, opti- 196
137 rate of suspension during milling in a Dynomill. Effects of grinding mization of operating parameters to obtain desired particle pro- 197
138 media size were not investigated in the present work, highest duct quality is carried out by evaluating the specific energy or 198
139 media filling ratio was selected as recommended by the supplier the stress intensity. Further, the particles to be ground in a grind- 199
140 of the mill. The cooling water temperature of 5 °C was maintained ing process and the resulting smaller particles are not subjected to 200
141 to control temperature the during milling process for all the Runs, the same number of stresses and to the same stress intensities. 201
142 except in the Run-3. The same product temperature of 20 °C was There exist distributions of stress events and stress intensity, gov- 202
143 maintained in the beginning of all the experimental Runs. Other erned by the operating parameters. Further, the overall specific 203
144 conditions employed in the different experimental Runs are shown energy consumption of the mill is given by the product of stress 204
145 in Table 1. Higher starch loading was used in milling experiment to number and stress intensity (Em = SE.SN). Therefore, the product 205
146 explore the effectiveness of stirred media milling technique for fineness can be correlated either with the stress number or the 206
147 large-scale production of starch nanoparticles. The milling specific energy input at constant stress intensity [33]. More details 207

Please cite this article in press as: C.M. Patel et al., Fast and scalable preparation of starch nanoparticles by stirred media milling, Advanced Powder Tech-
nology (2016),
APT 1250 No. of Pages 8, Model 5G
2 May 2016

C.M. Patel et al. / Advanced Powder Technology xxx (2016) xxx–xxx 3

Table 1
Experimental milling conditions employed in the study.

Run Milling conditions

Solid mass Stirrer speed Total milling time Cooling water Temperature of product before Temperature of product after
fraction (–) (rpm) (min) temperature (°C) milling (°C) milling (°C)
Run-1 0.05 3500 90 5 20 31
Run-2 0.08 3500 90 5 20 32
Run-3 0.10 4000–5000 30 20 20 40
Run-4 0.10 4000–5000 30 5 20 34
Run-5 0.05–0.10 3500–5000 60 5 20 42

208 on the application of the Kwade’s stress model [33] can be found in break [36,37]. Starch nanoparticles with Z-Average size of 241
209 our previous publication [31], where we have clearly shown the 245 nm and PdI of 0.210 were obtained at Cm = 0.05 after 90 min 242
210 application of this model to evaluate the milling results based on of milling. When processed at higher solids concentration 243
211 stress number and stress intensity. (Cm = 0.08); the Z-Average size, and PDI values were 253 nm, and 244
212 The milling results were evaluated by particle size distribution 0.179 respectively after 90 min of the milling. The differences in 245
213 (PSD) parameters as a function of milling time, namely, Z-Average Z-average size and PdI values obtained by two experimental runs 246
214 size, and PdI. The Z-Average size is used to evaluate the mean size were not too high. This suggests that it is possible to process starch 247
215 of particles, whereas polydispersity index (PdI) is used to indicate a slurry with higher concentration with comparable end results in 248
216 measure of the width of PSD. The Run-1 and Run-2 were conducted terms of the mean particle size and polydispersity index. 249
217 at different solid mass fractions. The smaller values of Z-Average
218 size and the PdI are desirable for the better product quality. The 3.2. Preparation of starch nanosuspension using high solids 250
219 Z-Average size and PdI both decreased with an increase in the concentration 251
220 milling time. The decrease in both parameters was much faster
221 during the initial stage of milling; however, decrease in their val- We tried to increase solid mass fraction to process large amount 252
222 ues becomes smaller with the growing milling time and eventually of starch, but we could not able to run the mill satisfactorily at a 253
223 achieved a plateau in both the cases (Fig. 1a and b). The change in higher solids loading (at and above Cm = 0.1). The laboratory mill 254
224 particle size decreases with milling time. This could be attributed used in the present work requires constant suspension flow for 255
225 to two factors affecting particle breakage: since strength is inver- proper functioning, which could not be maintained at higher solids 256
226 sely proportional to the square of particle size, at a constant stress- loading. We could able to operate the mill at a higher solid mass 257
227 ing energy input, the probability of achieving successful particle fraction of 0.1 for 30 min, but the stirrer speed had to be increased 258
228 breakage decreases with decreasing size and decreased defects in to 4500 rpm to maintain proper product flow. We observed inter- 259
229 smaller particles, and secondly, with more contacts which are esting changes in physiochemical properties of starch suspension 260
230 unsuccessful in initiating breakage, strain hardening occurs and during the milling. The results of three Runs carried out at higher 261
231 causes an increase in the particle strength at that unique contact solids loading (at Cm = 0.1) are represented in Table 2. The rapid 262
232 [34,35]. The Z-Average size decrease faster in the beginning of increase in the temperature of product suspension was observed 263
233 the milling process indicates that breakage was the dominant when Run-3 was carried out at the higher temperature and the 264
234 and aggregation of the particles occurred at a much smaller rate. higher stirrer speed (see Table 1). Starch microparticles are sub- 265
235 However, with an increase in milling time, Z-Average size jected to mechanical forces as well as thermal energy evolved dur- 266
236 remained almost same or increase in Z-Average size was small ing the milling process. An increase in the milling time causes 267
237 due to dynamic equilibrium attained between breakage and aggre- reduction and disappearance of birefringence, peaks of X-ray 268
238 gation. This slowing down in breakage rate is due to the fact that diffraction patterns, and the endothermic event characteristic of 269
239 finer particles were produced with progress in milling, which are gelatinization. This corroborates that crystallinity of the milled 270
240 more difficult to capture between the beads and are harder to samples produced from the starch nanoparticles becomes rather 271

Fig. 1. Effects of solid mass fraction and milling time on: (a) Z-average size; (b) PdI.

Please cite this article in press as: C.M. Patel et al., Fast and scalable preparation of starch nanoparticles by stirred media milling, Advanced Powder Tech-
nology (2016),
APT 1250 No. of Pages 8, Model 5G
2 May 2016

4 C.M. Patel et al. / Advanced Powder Technology xxx (2016) xxx–xxx

Table 2 are significantly different for Run-5 at a probability level of 5% 317

Z-Average size and PdI of starch nanoparticles obtained in different runs as a function (p < 0.05). 318
of milling time.
We allowed grinding for at least 10 min after each addition of 319
Milling time Z-Average size (nm) and mean PdI (in parentheses) fresh slurry, that time was sufficient to break particles to much 320
(min) smaller sizes, and we did not observe binary distribution of parti- 321
Run-3 Run-4 Run-5
15 307.9 ± 3.27 aA
309.8 ± 3.15aA
292.0 ± 6.65 bB cles in the Run-5. This could be attributed to the effectiveness of 322
(0.332) (0.468) (0.373) DYNOÒ-Acceleratoris in the size reduction. The observation, we 323
30 285.1 ± 0.42cC 288.2 ± 2.534cC 281.0 ± 2.84dD corroborated even for other material such as fenofibrate nanopar- 324
(0.189) (0.243) (0.222) ticles in our previous publication [25], where we could able to pro- 325
45 – – 271.8 ± 1.53e
duce much smaller particles within 10 min. The initial particle size 326
60 – – 266.4 ± 3.22f of 10 lm was reduced to around 500 nm within 10 min of media 327
(0.180) milling in Dynomill. 328
The faster preparation of SNPs could be ascribed to the effec- 329
Values represent the means ± standard deviation; n = 3.

Values followed by same lower case superscript in a column are not significantly tiveness of the patented DYNOÒ-Accelerator employed in the mill. 330
different (p < 0.05). The DYNOÒ-Accelerator creates an intense hydraulic flow of grind- 331
Values followed by same capital superscript in a row are not significantly different ing media and product, resulting into high energy input with opti- 332
(p < 0.05). mum distribution throughout the entire grinding chamber. The 333
stability of bead flow and regulation of energy input with 334
DYNOÒ-Accelerator allows a use of smallest beads. The use of 335
272 less pronounced [38]. The formation of gel like structure began
smaller beads at low energy input with a wide range of product 336
273 merely after the 15 min of the milling. The suspension became
viscosities at high flow rates is feasible. We certainly believe that 337
274 highly viscous after 30 min of the milling, and mill had to be
this process can be further improved by the use of smaller grinding 338
275 stopped because of the difficulty of the flow of viscous suspension
media of size 0.1 mm and 0.2 mm, and would be investigated fur- 339
276 through the mill. It is possible to transform this gel like structure
ther in the future work. 340
277 into a dispersion of SNPs in the water by gentle sonication.
278 To avoid the effects of thermal energy during milling and to
279 delay gelatinization of starch and gel formation; Run-4 was con- 3.4. Preparation of starch nanosuspension by laboratory batch mill 341
280 ducted at a lower temperature by reducing cooling water temper-
281 ature down to 5 °C. The gelatinization of starch could be delayed by The problem to process higher solids concentration is due to 342
282 lowering the product temperature. However, it was difficult to run requirement of strong and constant slurry flow during milling in 343
283 the mill beyond 30 min even with such low cooling water temper- Dynomill to ensure better performance. Thus, it is this limitation 344
284 ature. We obtained highly viscous starch suspension after 30 min of the mill that caused difficulty in processing larger solids loading, 345
285 of milling by Run-4, which when kept at a room temperature spon- but we firmly believe that production scale mill can be effectively 346
286 taneously changed to gel like structure. This gel can also be chan- employed to produce large-scale SNPs by media milling. To sup- 347
287 ged to stable dispersion by the sonication. Further experimentation port our claim that higher solids concentration can be handled, 348
288 needs to be done to study the effects of lower processing temper- we used another available media mill, which operates in batch 349
289 ature on the size reduction and the gelatinization of starch. mode. Fig. 2 shows the change in the Z-Average size and PdI during 350
batch milling. It is clearly observed that both Z-Average size and 351
PdI decreased with an increase in the milling time. The trend is 352
290 3.3. Preparation of starch nanosuspension by gradual increase of solids
similar to one obtained using Dynomill, however Z-Average size 353
291 concentration
and PdI after 15 min of milling was not shown in Fig. 2. This is 354
because the result obtained after 15 min of milling in batch milling 355
292 The final particle size obtained from Run-3 and Run-4 is higher
process was not of good quality and reproducible as measured by 356
293 as seen by the larger Z-Average size in Table 2 in comparison to
DLS technique, hence we have not shown results after 15 min of 357
294 Run-1 and Run-2. To reduce particle size further, we adopted a
milling. This observation indicates that batch milling was not able 358
295 completely different approach. The milling experiment in Run-5
to reduce particle size effectively within 15 min of milling. This 359
296 was started with lower solids loading, and solids concentration
297 was gradually increased thereafter. Since an initial solid loading
298 was low, it was possible to start experiment at lower stirrer speed
299 of 3500 rpm. However, stirrer speed had to be gradually increased
300 to 5000 rpm to maintain the product flow rate. With this strategy,
301 we were able to control the temperature of product slurry, which
302 enabled us to run the mill for 60 min. Consequently, smaller parti-
303 cles of size 266 nm were obtained as shown in Table 2. The other
304 strategies may be explored, for instance, the milling can be started
305 with higher solids concentration followed by the gradual reduction
306 in the concentration of starch slurry. Different structural transi-
307 tions were observed during milling of starch under different condi-
308 tions. The similar structural transitions were reported when starch
309 particle and water mixture were processed by high pressure
310 homogenization [17]. The effects of different processing conditions
311 on the milling results are shown in Table 2. Statistical analysis was
312 performed by Student’s t test to estimate significant differences
313 among means of Z-Average size and PdI obtained by different runs
314 at 0.95 confidence level. It is clearly seen from the Table 2 that the
315 Z-Average size and PdI values obtained in Run-3 and Run-4 at dif- Fig. 2. Effect of milling on Z-average size and PdI in laboratory mill working in
316 ferent milling times were not significantly different; however, they batch mode.

Please cite this article in press as: C.M. Patel et al., Fast and scalable preparation of starch nanoparticles by stirred media milling, Advanced Powder Tech-
nology (2016),
APT 1250 No. of Pages 8, Model 5G
2 May 2016

C.M. Patel et al. / Advanced Powder Technology xxx (2016) xxx–xxx 5

360 shows the effectiveness of DYNOÒ-Accelerator on particle size study is required to be done to understand redispersion behavior 402
361 reduction. Thus, we firmly believe the difficulty experienced in of starch nanoparticles in the absence of dispersant. In addition, 403
362 Dynomill with higher solids concentration is only due to constant the issue of stability of SNPs for a long time should be addressed 404
363 slurry flow problem and it is possible to produce SNPs with higher by stabilizing suspension with appropriate stabilizer. 405
364 solid concentration in other types of stirred media mills. Further, stability of starch nanoparticles obtained by Run-2 was 406
365 Most of the reported methods for preparation of SNPs could not monitored for 7 days. The results, presented in Fig. 3b, clearly 407
366 process large amount of starch and involve many steps. Large demonstrate that Z-Average size and PdI values remain almost 408
367 amount of organic solvent is needed for solvent evaporation, constant for the observed period. Some fraction of starch particles 409
368 whereas the precipitation method suffers from many complicated settled for this sample as well, but it is pretty much evident that 410
369 processing-steps [18]. Acid hydrolysis is practically difficult due to stability of starch nanoparticles obtained over longer milling time 411
370 low yield, and long time of treatment [39]. The recently reported of 90 min was much better than obtained by shorter milling time 412
371 method of synthesis of starch nanoparticles in ionic liquid-in-oil of 15 min. The redispersibility of prepared SNPs was also investi- 413
372 microemulsion system also requires many process steps and use gated by freeze drying of prepared SNPs dispersion. The homoge- 414
373 of chemicals [40]. The processing of starch slurry with higher solids nous dispersion was prepared by sonication (35% amplitude) of 415
374 concentration by high pressure homogenization and ultrasonica- freeze dried SNPs in water for 5 min. Fig. 4 shows the particle size 416
375 tion is not reported in the literature. The media milling process is distribution of redispersed freeze dried sample obtained by Run-2 417
376 capable to fabricate nanomaterials in quantities of kilograms after 90 min of the milling. As seen from Fig. 4, freeze dried SNPs 418
377 [41]. In these contexts, we demonstrated that SNPs could be can be effectively dispersed by mild ultrasonication and obtained 419
378 rapidly prepared with higher solids concentration by the stirred particle size distribution compares well with freshly prepared SNPs 420
379 media mill on a large-scale. dispersion. The particle size distributions measured immediately 421
after the preparation of starch nanoparticles and after seven days 422
of storage are also shown in Fig. 4. Hardly any difference can be 423
380 3.5. Stability analysis of starch nanosuspension marked in the size distributions obtained at different conditions, 424
which implied a good stability of freshly prepared as well as freeze 425
381 The stability of starch nanosuspension was investigated using dried SNPs. 426
382 DLS technique. The stability analysis was carried out for samples
383 obtained by Run-4 and Run-2 after 15 min and 90 min of milling
384 respectively. The as prepared nanosuspension was diluted to con- 3.6. Morphological analysis of starch nanoparticles 427
385 tain 0.1 wt% SNPs. We have successfully prepared stable SNPs
386 without the use of dispersant by stirred media milling. Interest- The morphological features of starch particles were investigated 428
387 ingly, particle size measured after 1 day was much smaller than by FEG-SEM and TEM analysis to understand structural changes 429
388 obtained at the time of the experiment. This could be probably caused by wet milling process. As seen from Fig. 5a, native starch 430
389 due to the settling of some of the starch nanoparticles during the microparticles had a smooth surface with particle sizes ranging 431
390 storage. The samples were manually shaken to confirm whether from 3 to 20 lm. The milling of native starch particles resulted 432
391 these particles were agglomerated or just settled. DLS measure- in the particle size reduction. SEM micrograph in Fig. 5b clearly 433
392 ments were performed on shaken and unshaken samples pre- demonstrates that gel like structure with agglomerated particles 434
393 served separately. The results, shown in Fig. 3a, for sample were obtained at a shorter processing time of 15 min, whereas 435
394 obtained by Run-4 indicate that the Z-Average size and PdI of smaller starch nanoparticles were obtained after 90 min of milling 436
395 unshaken sample were much smaller than that of shaken sample. by Run-2 (Fig. 5c). TEM micrographs were also obtained to find pri- 437
396 This could have been due to the settling of some of the large starch mary particle sizes. TEM micrograph of SNPs obtained after 15 min 438
397 nanoparticles. In addition, in both the cases Z-Average size and PdI of milling in Run-4, shown in Fig. 6a, indicates that particle sizes 439
398 did not increase with storage time for five days. This means that were comparatively larger and gel like structure is clearly visible. 440
399 particles were not agglomerated during five days; indicating that The particle sizes reduce greatly after 90 min of the milling as 441
400 sample remained stable during this period. However, both shown in Fig. 6b. Additionally, gel like structure visible in Fig. 6a 442
401 Z-Average size and PdI were increased after five days. Further confirms the results of SEM micrograph (Fig. 5b). The smaller 443

0.7 300 0.25

0.6 290
320 280
Z-Average size (nm)
Z-Average size (nm)


300 270
PdI ( -)

PdI ( -)

260 0.2
Z-Average size - unshaken 240
Z-Average size - shaken Z-Average size
240 0.1
(a) PdI - unshaken 230 (b) PdI
PdI - shaken
220 0 220 0
-1 0 1 2 3 4 5 6 7 8 -1 0 1 2 3 4 5 6 7 8
Time of storage (d) Time of storage (d)

Fig. 3. The change in Z-average size and PdI as a function of storage time of starch nanosuspension obtained by (a) Run-4. (b) Run-2.

Please cite this article in press as: C.M. Patel et al., Fast and scalable preparation of starch nanoparticles by stirred media milling, Advanced Powder Tech-
nology (2016),
APT 1250 No. of Pages 8, Model 5G
2 May 2016

6 C.M. Patel et al. / Advanced Powder Technology xxx (2016) xxx–xxx

Fig. 4. Particle size distribution of sample obtained after 90 min of milling by Run-2
as a function of storage time and redispersed freeze dried SNPs sample.

444 and highly dispersed SNPs were obtained after 90 min of milling by
445 Run-2 with a primary particle size less than 25 nm (Fig. 6b). The
446 discrepancy in size, measured by TEM and DLS is attributed to
447 the fact that the DLS technique is more sensitive to the presence
448 of a small amount of large particles, and generally gives larger size
449 compared to TEM [21].

450 3.7. XRD analysis

451 X-ray diffraction patterns of native starch and freeze dried SNPs
452 were obtained to study the effects of the milling time on the struc-
453 tural properties of starch nanoparticles (Fig. 7). Starches are classi-
454 fied in three crystalline structures called A, B, and C type based on
455 their X-ray diffraction pattern [42]. The native starch posses a dis-
456 tinct A-type crystalline structure (Fig. 7a) with diffraction peaks at
457 15°, 17°, 18° and 23° of 2h [19]. It is well established in the litera-
458 ture that milling process generally reduces the crystallinity of
459 materials [15,29,32,43]. When particle size decreases, intensity of
460 the diffraction peak decreases; in contrast, width at half height
461 increases with decrease in particle size [17]. XRD pattern of the
462 sample obtained after 15 min of milling shows that intensity of
463 all diffraction peaks decreased with the broadening of peaks, also
464 the peaks at 17° and 18° were merged after 15 min of milling
465 (Fig. 7b). After 90 min of milling all the diffraction peaks observed
466 in native starch were disappeared, and only broad hallow from 5°
467 to 30° of 2h was left (Fig. 7c). This is in agreement with reported
468 results for starch nanoparticles prepared by high pressure homog-
469 enization [17], and high power ultrasonication [19]. From these
470 results, it could be concluded that starch nanoparticles with
471 reduced crystallinity could be obtained by the milling process.

472 3.8. FTIR analysis

473 FTIR analysis was carried out to study the structural changes Fig. 5. SEM micrographs of: (a) Native starch; (b) SNPs obtained after 15 min of
474 caused by the wet milling process. The structural changes caused milling by Run-4; (c) SNPs obtained after 90 min of milling by Run-2.
475 by the wet milling process could be evaluated by observing
476 decrease in peak intensity, broadening of absorption peak, shifting
477 of bands, and appearance of new bands in the FTIR spectra [44]. starch. The several discernible absorbencies located between 1250 486
478 The FTIR spectrums of native starch and starch nanoparticles and 900 cm 1 in native starch were assigned to the C–O bond 487
479 obtained after different milling times are presented in Fig. 8 with stretching [40,44]. In comparison with native starch, FTIR spectra 488
480 assignment of main absorption bands [45]. In the FTIR spectra of of starch nanoparticles prepared by the milling process had chan- 489
481 native starch, the broad absorption band 3900–3300 cm 1 is due ged. The intensity of 3900–3300 cm 1 band had decreased and got 490
482 to O–H stretching; represents the complex vibrational stretching broader for starch nanoparticles after different milling times 491
483 associated with hydroxyl groups. The characteristic absorption (Fig. 8b and c). The band at 2920 cm 1 became much weaker and 492
484 band at 2920 cm 1 is assigned to C–H stretching, whereas charac- broadened, in the spectra of SNPs obtained, at different milling 493
485 teristic peak at 1645 cm 1 is due to the presence of bound water in times than that of native starch. The peaks absorption bands 494

Please cite this article in press as: C.M. Patel et al., Fast and scalable preparation of starch nanoparticles by stirred media milling, Advanced Powder Tech-
nology (2016),
APT 1250 No. of Pages 8, Model 5G
2 May 2016

C.M. Patel et al. / Advanced Powder Technology xxx (2016) xxx–xxx 7



T (%)
1651.57 1247.42
1158.51 994.12


4000 3500 3000 2500 2000 1500 1000 500

Wave number (cm )

Fig. 8. FTIR spectra of (a) Native starch (b) SNPs obtained after 15 min of milling by
Run-4; (c) SNPs obtained after 90 min of milling by Run-2.

located between 1250 and 900 cm 1 also changed with the broad- 495
ening and the reduction in the peak intensity. The results of the 496
FTIR analysis of SNPs are in agreement with the reported results 497
[40]. These observations substantiate the fact that structural 498
changes have been caused by milling process in prepared SNPs 499
and are consistent with the findings of XRD analysis. 500

4. Conclusions 501

Starch nanoparticles were successfully prepared by high speed 502

wet stirred media milling. The effects of solid mass fraction and 503
milling time on particle sizes and polydispersity index were evalu- 504
ated. Starch nanoparticles exhibited a great degree of stability 505
without the use of chemical stabilizers. The morphological features 506
of prepared SNPs were highly dependent on the processing time as 507
revealed by FEG-SEM and TEM investigation. The crystallinity of 508
SNPs was decreased upon prolonged milling as suggested by XRD 509
and FTIR analysis. The proposed method is suitable for large- 510
Fig. 6. TEM micrographs of: (a) SNPs obtained after 15 min of milling by Run-4; (b)
SNPs obtained after 90 min of milling by Run-2. scale preparation of SNPs with high yield of SNPs in shorter time. 511

Acknowledgments 512

Authors wish to acknowledge the Sophisticated Instrumenta- 513

tion Centre for Applied Research and Testing (SICART), Vallabh Vid- 514
hyanagar, India; Sophisticated Instrumentation Centre, S.V.N.I.T, 515
Surat, India; and Shah-Schulman Center for Surface Science & Nan- 516
otechnology, D. D. University, Nadiad, India; for providing TEM, 517
SEM, and XRD analysis services respectively. 518
Intensity (A.U.)

(a) References 519

[1] M. Li, P. Liu, W. Zou, L. Yu, F. Xie, H. Pu, H. Liu, L. Chen, Extrusion processing and 520
characterization of edible starch films with different amylose contents, J. Food 521
Eng. 106 (2011) 95–101. 522
(b) [2] O.A. Odeku, Potentials of tropical starches as pharmaceutical excipients: a 523
review, Starch/Stärke 65 (2013) 89–106. 524
[3] H.F. Zobel, Starch crystal transformations and their industrial importance, 525
(c) Starch/Stärke 40 (1988) 1–7.
[4] A.-M. Shi, D. Li, L.-J. Wang, Y.-G. Zhou, B. Adhikari, Spray drying of starch 527
submicron particles prepared by high pressure homogenization and mini- 528
emulsion cross-linking, J. Food Eng. 113 (2012) 399–407. 529
[5] C. Seidel, W.M. Kulicke, C. Hess, B. Hartmann, W. Lazik, Synthesis and 530
0 10 20 30 40 characterization of cross-linked carboxymethyl potato starch ether gels, 531
Starch/Stärke 56 (2004) 157–166. 532
2θ(Degree) [6] C. Xiao, Current advances of chemical and physical starch-based hydrogels, 533
Starch/Stärke 65 (2013) 82–88. 534
Fig. 7. WXRD patterns of: (a) Native starch; (b) SNPs obtained after 15 min of [7] P.B. Malafaya, F. Stappers, R.L. Reis, Starch-based microspheres produced by 535
milling by Run-4; (c) SNPs obtained after 90 min of milling by Run-2. emulsion crosslinking with a potential media dependent responsive behavior 536

Please cite this article in press as: C.M. Patel et al., Fast and scalable preparation of starch nanoparticles by stirred media milling, Advanced Powder Tech-
nology (2016),
APT 1250 No. of Pages 8, Model 5G
2 May 2016

8 C.M. Patel et al. / Advanced Powder Technology xxx (2016) xxx–xxx

537 to be used as drug delivery carriers, J. Mater. Sci.: Mater. Med. 17 (2006) 371– [26] F. Kesisoglou, S. Panmai, Y. Wu, Nanosizing – oral formulation development 592
538 377. and biopharmaceutical evaluation, Adv. Drug Deliver. Rev. 59 (2007) 631–644. 593
539 [8] R.C. Mundargi, N.B. Shelke, A.P. Rokhade, S.A. Patil, T.M. Aminabhavi, [27] E. Khullar, B.S. Dien, K.D. Rausch, M.E. Tumbleson, V. Singh, Effect of particle 594
540 Formulation and in-vitro evaluation of novel starch-based tableted size on enzymatic hydrolysis of pretreated Miscanthus, Ind. Crop. Prod. 44 595
541 microspheres, Carbohydr. Polym. 71 (2008) 42–53. (2013) 11–17. 596
542 [9] L. Zhang, W. Cai, J. Shan, S. Zhang, F. Dong, Physical properties and loading [28] C.-J. Chen, Y.-C. Shen, A.-I. Yeh, Physico-chemical characteristics of media- 597
543 capacity of gelatinized granular starches, Ind. Crop. Prod. 53 (2014) 323–329. milled corn starch, J. Agric. Food Chem. 58 (2010) 9083–9091. 598
544 [10] D. Visinescu, F. Papa, A.C. Ianculesc, I. Balint, O. Carp, Nickel-doped zinc [29] H. Liimatainen, J. Sirvio, A. Haapala, O. Hormi, J. Niinimaki, Characterization of 599
545 aluminate oxides: starch-assisted synthesis, structural, optical properties, and highly accessible cellulose microfibers generated by wet stirred media milling, 600
546 their catalytic activity in oxidative coupling of methane, J. Nanopart. Res. 15 Carbohydr. Polym. 83 (2011) 2005–2010. 601
547 (2013) 1456. [30] W. Zhang, J. Zhang, W. Xia, The preparation of chitosan nanoparticles by wet 602
548 [11] Y.M. Soshnikova, S.G. Roman, N.A. Chebotareva, O.I. Baum, M.V. Obrezkova, R. media milling, Int. J. Food Sci. Technol. 47 (2012) 2266–2272. 603
549 B. Gillis, S.E. Harding, E.N. Sobol, V.V. Lunin, Starch-modified magnetite [31] C.M. Patel, M. Chakraborty, Z.V.P. Murthy, Enhancement of stirred media mill 604
550 nanoparticles for impregnation into cartilage, J. Nanopart. Res. 15 (2013) 2092. performance by a new mixed media grinding strategy, J. Ind. Eng. Chem. 20 605
551 [12] A.K. Zak, W.H.A. Majid, M.R. Mahmoudian, M. Darroudi, R. Yousefi, Starch- (2014) 2111–2118. 606
552 stabilized synthesis of ZnO nanopowders at low temperature and optical [32] C.M. Patel, M. Chakraborty, Z.V.P. Murthy, Study on the stability and 607
553 properties study, Adv. Powder Technol. 24 (2013) 618–624. microstructural properties of barium sulfate nanoparticles produced by 608
554 [13] M. Chichova, M. Shkodrova, P. Vasileva, K. Kirilova, D. Doncheva-Stoimenova, nanomilling, Adv. Powder Technol. 20 (2014) 2111–2118. 609
555 Influence of silver nanoparticles on the activity of rat liver mitochondrial [33] A. Kwade, A stressing model for the description and optimization of grinding 610
556 ATPase, J. Nanopart. Res. 16 (2014) 2243. processes, Chem. Eng. Technol. 26 (2003) 199–205. 611
557 [14] Gh.H. Khorrami, A. Kompany, A.K. Zak, Structural and optical properties of (K, [34] A. Russell, P. Muller, J. Tomas, Quasi-static diametrical compression of 612
558 Na)NbO3 nanoparticles synthesized by a modified sol–gel method using starch characteristic elastic–plastic granules: energetic aspects at contact, Chem. 613
559 media, Adv. Powder Technol. 26 (2015) 113–118. Eng. Sci. 114 (2014) 70–84. 614
560 [15] H. Yousefi, M. Faezipour, S. Hedjazi, M.M. Mousavi, Y. Azusa, A.H. Heidari, [35] Salman, A. Russell, S. Aman, J. Tomas, Breakage probability of granules during 615
561 Comparative study of paper and nanopaper properties prepared from bacterial repeated loading, Powder Technol. 269 (2015) 541–547. 616
562 cellulose nanofibers and fibers/ground cellulose nanofibers of canola straw, [36] C. Knieke, M.A. Azad, R.N. Dave, E. Bilgili, A study of the physical stability of 617
563 Ind. Crop. Prod. 43 (2013) 732–737. wet media-milled fenofibrate suspensions using dynamic equilibrium curves, 618
564 [16] D. Le Corre, J. Bras, A. Dufresne, Starch nanoparticles: a review, Chem. Eng. Res. Des. 91 (2013) 1245–1258. 619
565 Biomacromolecules 11 (2010) 1139–1153. [37] E. Bilgili, R. Hamey, B. Scarlett, Nano-milling of pigment agglomerates using a 620
566 [17] D. Liu, Q. Wu, H. Chen, P.R. Chang, Transitional properties of starch colloid with wet stirred media mill: elucidation of the kinetics and breakage mechanisms, 621
567 particle size reduction from micro-to nanometre, J. Colloid Interface Sci. 339 Chem. Eng. Sci. 61 (2006) 149–157. 622
568 (2009) 117–124. [38] S. Dhital, A.K. Shrestha, M.J. Gidley, Effect of cryo-milling on starches: 623
569 [18] A.-M. Shi, D. Li, L.-J. Wang, B.-Z. Li, B. Adhikari, Preparation of starch based functionality and digestibility, Food Hydrocolloid. 24 (2010) 152–163. 624
570 nanoparticles through high-pressure homogenization and miniemulsion [39] D. Song, Y.S. Thioc, Y. Deng, Starch nanoparticle formation via reactive 625
571 cross-linking: Influence of various process parameters on particle size and extrusion and related mechanism study, Carbohydr. Polym. 85 (2011) 208– 626
572 stability, Carbohydr. Polym. 83 (2011) 1604–1610. 214. 627
573 [19] S.B. Haaj, A. Magnin, C. Petrier, S. Boufi, Starch nanoparticles formation via [40] G. Zhou, Z. Luo, X. Fu, Preparation and characterization of starch nanoparticles 628
574 high power ultrasonication, Carbohydr. Polym. 92 (2013) 1625–1632. in ionic liquid-in-oil micro emulsion system, Ind. Crop. Prod. 52 (2014) 105– 629
575 [20] O.Y. Toraman, D. Katircioglu, A study on the effect of process parameters in 110. 630
576 stirred ball mill, Adv. Powder Technol. 22 (2011) 26–30. [41] P. Wagener, M. Lau, S. Breitung-Faes, A. Kwade, S. Barcikowski, Physical 631
577 [21] S. Mende, F. Stenger, W. Peukert, J. Schwedes, Mechanical production and fabrication of colloidal ZnO nanoparticles combining wet-grinding and laser 632
578 stabilization of submicron particles in stirred media mills, Powder Technol. fragmentation, Appl. Phys. A 108 (2012) 793–799. 633
579 132 (2003) 64–73. [42] D. LeCorre, J. Bras, A. Dufresne, Influence of botanic origin and amylose content 634
580 [22] S. Sakthivel, V.V. Krishnan, B. Pitchumani, Influence of suspension stability on on the morphology of starch nanocrystals, J. Nanopart. Res. 13 (2011) 7193– 635
581 wet grinding for production of mineral nanoparticles, Particuology 6 (2008) 7208. 636
582 120–124. [43] M. Anzai, T. Hagiwara, M. Watanabe, J. Komiyama, T. Suzuki, Relationship 637
583 [23] K. Ohenoja, M. Illikainen, J. Niinimaki, Effect of operational parameters and between enthalpy relaxation and water sorption of ball-milled potato starch, J. 638
584 stress energies on the particle size distribution of TiO2 pigment in stirred Food Eng. 104 (2011) 43–48. 639
585 media milling, Powder Technol. 234 (2013) 91–96. [44] S. Palaniandy, K.A.M. Azizli, Mechanochemical effects on talc during fine 640
586 [24] S. Ouattara, C. Frances, Grinding of calcite suspensions in a stirred media mill: grinding process in a jet mill, Int. J. Miner. Process. 92 (2009) 22–33. 641
587 effect of operational parameters on the product quality and the specific [45] J.F. Mano, D. Koniarova, R.L. Reis, Thermal properties of thermoplastic starch/ 642
588 energy, Powder Technol. 255 (2014) 89–97. synthetic polymers blend with potential biomedical applicability, J. Mater. Sci. 643
589 [25] C.M. Patel, M. Chakraborty, Z.V.P. Murthy, Preparation of fenofibrate – Mater. Med. 14 (2003) 127–135. 644
590 nanoparticles by combined stirred media milling and ultrasonication 645
591 method, Ultrason. Sonochem. 21 (2014) 1100–1107.

Please cite this article in press as: C.M. Patel et al., Fast and scalable preparation of starch nanoparticles by stirred media milling, Advanced Powder Tech-
nology (2016),