Effect of Particle Shape on Inline Particle Size Measurement Techniques

Research Article

Maryam Mostafavi Sandra Petersen Joachim Ulrich Martin-Luther University Halle-Wittenberg, Center for Engineering Science, Thermal Process Engineering, Halle, Germany.

1721

Effect of Particle Shape on Inline Particle Size Measurement Techniques The influence of the particle shape on the results of different measuring techniques was investigated. Considered were single-frequency ultrasound technique, 3D optical reflectance measurement (advanced particle analyzing system with multi capture signal technology), and focused-beam reflectance measurement probes as techniques which are usable inline and in suspension density ranges usually present in industrial crystallization processes. Advantages and shortcomings of these techniques are compared and discussed. Keywords: Crystallization, Crystal size, Focused-beam reflectance measurement, Optical reflectance measurement, Single-frequency ultrasound Received: April 07, 2014; revised: May 22, 2014; accepted: July 18, 2014 DOI: 10.1002/ceat.201400212

1

Introduction

Crystallization is a unique, indispensable process step in many industrial production lines such as pharmaceuticals and agrochemicals. Features like morphology and particle size are important parameters especially in the pharmaceutical crystallization processes [1]. For instance, the efficiency, i.e., bioavailability, of an active pharmaceutical ingredient (API) is strongly modified by the dissolution rate of crystals and depends also on crystal size and crystal size distribution (CSD) [2]. It is essential to control the size and the number of crystals produced in the crystallization processes. The knowledge of timedependent particle numbers and particle size distribution of a concentrated crystal suspension (10–30 vol % of crystals) are two crucial variables in various processes [1]. To measure size distributions is frequently difficult and for control purposes often only thresholds or direction of changes are needed. Therefore, a mean particle size would be sufficient. Some measurement techniques are also not able to provide more than one parameter of a size-related value. Sometimes this is enough even though a full size distribution would give a more complete description. Several measuring principles are available to determine the CSD of the particles. Most are based on the measured signal as well as on a mathematical assumption and therefore not primary results. Offline sizing techniques such as image analysis, coulter counter, and laser diffraction or cascade sieving typically rely on suitable sampling, but with a large-scale production process, sampling is often hazardous and non-representative.

– Correspondence: Prof. Joachim Ulrich ([email protected]), Martin-Luther-Universita¨t Halle-Wittenberg, Zentrum fu¨r Ingenieurwissenschaften, Thermische Verfahrenstechnik, Hoher Weg 7, 06120 Halle, Germany.

Chem. Eng. Technol. 2014, 37, No. 10, 1721–1728

The single-frequency ultrasound technique, the ultrasound attenuation spectroscopy (UAS), the optical reflectance measurement (3D ORM; advanced particle analyzing system (APAS) with multi capture signal technology (MCST)), as well as the focused-beam reflectance measurement technique (FBRM) are established techniques for the measurement of the mean particle size, the mean chord length, or the particle size distribution and the suspension densities inline and in real time [3]. The ultrasound measuring technique (single frequency) is an ideal method for in situ determination of the crystallization parameters in the liquid phase (concentration, supersaturation) and solid phase (suspension density and mean particle size) and more preferable because most materials are ultrasonically transparent. This technique has a high reliability and belongs to the series of new inexpensive process analytical technologies (PAT) developments. Furthermore, this technique differs from UAS. An advantage of the UAS technology is the possible determination of the particle size distributions. A disadvantage of UAS is the requirement of lengthy measurement times (60 s) making it unsuitable for very fast changing dynamic systems. Also UAS is limited regarding the complexity in analyzing the obtained data. The analysis of the attenuation spectra by means of mathematical models to calculate the CSD and hence the mean crystal size is difficult [2, 4]. The reflectance measurement technique is a reliable measurement tool that can be used under fast dynamic conditions without calibration effort and is widely applied in different industries such as bioprocesses, polymerization, and crystallization [3]. The shape of crystals is an important parameter for pharmaceutical crystallization processes since the surface area of a particle influences the dissolution speed, too. The shape, e.g., needles, of crystalline materials can affect the recorded result of some sizing techniques in a way that wrong values may be calculated from the signals. Most measuring techniques assume a spherical shape for the suspended particles. However, this

ª 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

www.cet-journal.com

Research Article

1722

assumption is in most cases not true [5]. The effect of particle shape on the result of different measuring techniques in crystallization is investigated.

2

Measurement Principles

2.1 Single-Frequency Ultrasound Measurement The sound velocity in case of the single-frequency equipment is measured by recording the time required for the sound to pass the known distance between the sound generator and the receiver. In a given medium, sound travels with a velocity (n), which depends on the density (r) and adiabatic compressibility of the medium (kad). Both of these parameters depend on temperature and pressure. For liquids, the pressure dependence can be ignored [6]. Single-frequency ultrasound technique measures ultrasound velocity and attenuation. These measured parameters are applied to calculate the most important solid phase variables such as suspension density (j) and mean crystal size (d50) by applying simple mathematical models based on the equation of Urick [2, 8]. Pertig et al. [9] used different suspension densities and particle size ranges for ammonium sulfate and urea and recorded ultrasonic velocity and attenuation by single-frequency ultrasound technique. They expressed the measured ultrasound signals as a function of suspension density (j) and mean crystal size (d50) by mathematical models. Sayan et al. [10] investigated the change of ultrasound velocity as a function of particle size and suspension densities. The ultrasound velocity and attenuation of different suspension densities and different particle size ranges were determined by single-frequency ultrasound measurement technique. It was found that variation in factors such as particle size and solid concentration can have a significant effect on the ultrasound velocity measurement [10].

Figure 1. Measurement principle of the 3D ORM instrument [11]. A moving light beam is focused into a flowing solution and particles are detected by their reflection as they pass through the moving laser focus. If the particles are not in the measuring zone, the reflected light will not be collected by the sensor.

2.3 FBRM In FBRM, a laser light source produces a continuous beam of monochromatic light focused to a small spot on the surface of the probe window. A pneumatic or electrical motor rotates the optics in such a way that the rotating, focused beam of laser light is continuously scanning over particles that are passing in front of the probe as demonstrated in Fig. 2. The suspended particles scatter the laser light back to the probe where the reflected light is detected. From the duration of each back scatter of light and the rotation velocity of the optics, the chord length can be determined [11].

2.2 3D ORM 3D ORM (APAS with MCST) as well as FBRM are based on the physical effect of reflection. Again a high-quality time measurement is required. The light from a highly stabilized 10-mW laser diode is coupled in a 4-mm single-mode fiber passing the optical mechanical dynamic selective focus system of the beam wave guide. The focus is moving in a dynamic depth into the suspension. Crystals which are inside the center of the focus are measured concerning their exposed particle surface area. Only if the focus is in alignment with the system, i.e., as shown in Fig. 1 where only those particles are measured which are in the focal point, these signals are taken into account. The alignment is one property which is based on the fact that the system is using only one fiber for in- and outgoing energy/signals. This technology can only work as long as 1 ms is between two signals. Depending on the process and application, suspension densities from 2 to 40 wt % are commonly applied [3]. The outcome of the measurement is to determine the mean chord length or the mean particle size and the suspension densities inline and in real time.

www.cet-journal.com

Figure 2. Measurement principle of the laser scanner (FBRM instrument) [11].

From these measurements (FBRM and 3D ORM), the total counts, which enable to define the suspension density in case the overall sample volume does not change, and the chord lengths, which can be used to define the size of the particles, can be obtained. Tab. 1 compares the used 3D ORM and FBRM probes. Using FBRM, a focal point position of the laser is chosen at +20 mm from the probe window. In comparison, the 3D ORM (APAS with MCST) technique works with a self-selective dynamical focal point which guarantees the highest resolution of the particles. The optimal distance of the focal point from the sensor window is half the size, i.e., mean diameter, of the investigated particles. Particles which are only partially in the focus or out of the focus cannot be detected [3].



Research Article

1723

Table 1. Comparison of the 3D ORM (APAS with MCST) and FBRM technologies [3]. Features

APAS

FBRM

Rotation of optical lens

No

Yes, with 2 ms–1

Focal point

Dynamic (max. 600 mm outside window system)

Fixed (+20 mm inside sapphire window)

Fiber optic

One single-mode 4 mm fiber (fiber-optical coupler)

Two multimode 105/125 mm fibers (beam splitter)

Laser intensity

10 mW (modular laser systems up to 30 mW)

0.9 mW at the probe window

Laser wavelength

785 nm (optional 400–850 nm)

780 nm

Extraction of raw reflection signals

Yes (MCST)

Yes

Measurement range

0.5–2000 mm

0.5–2000 mm

3

Experimental

3.1 Ultrasound Probe For investigation of the particle shape, two systems with different shapes were selected: (i) sugar/water with hexagonal prism as substance shape and (ii) L-threonine/water with needle-like substance shape. First, a defined amount of L-threonine with sieved particle size ranges of 80–200 mm was considered. Under the assumption that the particle number and the volume of each particle in both mentioned systems remain the same, the amount and the size of sugar particles were calculated. During the isothermal experiments, defined amounts of these substances were added to the saturated solution and the single-frequency ultrasound technique was applied to measure the ultrasound velocity of the two systems. The setup consists of the ultrasound sensor, a controller, and a computer for data acquisition. The ultrasound sensor emits a low-frequency (2 MHz) ultrasound wave. The energy input into the system is by far too low to cause cavitation [6].

3.2 Laser Backscattering Probes Two systems for investigation of the effect of particle shape on the measured results were selected: (i) ammonium chloride/ water (distilled) as a substance with cubic shape and (ii) L-threonine/water (distilled) as a substance with needle-like shape. Saturated L-threonine/water and ammonium chloride/water solutions were prepared according to the solubility data [12, 13]. Each experiment was started with a saturated solution at 20 C. During the isothermal experiments, defined amounts of particles with different size ranges of L-threonine and ammonium chloride, separated by sieving, were added to the saturated solutions. Two different suspension densities (10–15 vol %), each with sieved particle size ranges of 80–200 mm, were measured by FBRM and 3D ORM probes. These experiments started with a suspension density of 15 vol % and were subsequently changed for each measured suspension density by adding defined amounts of saturated solution in order to avoid undesired air intake during addition


of solids [9]. The experiments were performed in a jacketed glass vessel, and an external thermostat was used to control temperature (isotherm at 20 C). The crystallizer was equipped with a propeller stirrer and two probes. All laser backscattering probes were located at an angle of approximately 45.

4

Results and Discussion

4.1 Ultrasound Device: L-Threonine and Sugar Suspensions Tests with the same number and volume of each particle were carried out at constant temperature in each case to prove the effect of the particle shape on the ultrasound measurement technique in two systems: (i) sugar/water with a non-needle-like substance shape and (ii) L-threonine/water with a needle-like substance shape. The inline monitored results of these two systems each at different temperature levels are presented in Fig. 4. It is obvious that the ultrasound technique is capable to measure accurately inline as a function of time at constant temperatures. A constant signal was achieved for both systems, meaning that the different physical properties of the liquid phase do not influence the signal for the solid phase. An effect of the particle shape on the measured signals by this measuring technique is not visible. Although the used ultrasound technique is apparently insensitive to the particle shape, the ultrasound velocity, however, is clearly sensitive to particle size. The ultrasound technique was established and validated to determine the metastable zone width (MZW) and the solubility under different conditions, and to provide information on particle size and suspension densities as also illustrated in [9].

4.2 3D ORM Device 4.2.1 Measurement of Counts in L-Threonine and Ammonium Chloride Suspensions Based on the same suspension density and particle size range, experiments were performed at constant temperature to prove the effect of particle shape on the 3D ORM measurement tech-


www.cet-journal.com

Research Article

1724

Figure 3. Experimental setup with ultrasound probe, 3D ORM (APAS with MCST), FBRM probes, and electric stirrer in the crystallizer.

nique in the two systems: (i) ammonium chloride/water (distilled) as a substance with cubic shape and (ii) L-threonine/ water (distilled) as a substance with needle-like shape. In Fig. 5, the counts of the two substances are presented which were measured by means of the 3D ORM device. There is a significant difference between the count numbers of the two systems with the same volume suspension density which is related to dissimilar reflection intensities of the two substances. These reflection intensities of the particles can be analyzed by means of the obscuration factor (OBF); see Fig. 6. This factor indicates the average reflection intensities of particles in the focus of the laser and was measured by 3D ORM measurement technique during the experiments [3].

As can be concluded from Fig. 6, the OBF for ammonium chloride has a higher value compared to L-threonine. It was found from experiments that the number of particles obtained by the 3D ORM measurement technique depends on this factor. The smaller this factor is, the lower is the absolute value of the measured counts. Therefore, more ammonium chloride particles can be counted. The OBF is also influenced by particle properties such as roughness, impurities, etc. Consequently, the difference in refractive index between particles and solution was considered. The refractive index at a solid phase and particle-solvent system of substances like ammonium chloride, ascorbic acid and a-glycine were determined by Heinrich [11] and it was found that the smaller the difference in the refractive indices between particle and solution is, the lower is the absolute value of the measured counts. The refractive indices for the substances used in this work are summarized in Tab. 2.

4.2.2 Particle Size Measurement of L-Threonine and Ammonium Chloride

Figure 4. Ultrasound velocity versus time; sugar (20–80 mm, gray line), L-threonine (20–80 mm, black line). Photos on the right side: (A) L-threonine, (B) sugar. The photos of the crystals were taken offline by a microscope.

www.cet-journal.com


The average length of the two substances measured by the 3D ORM technique obtained from the exposed particle surface area is displayed in Fig. 7. This technique does not measure the particle size directly. The outline measurement of this sizing technique needs to be deconvoluted and converted into a particle size. The d50 (average length from area measurement) recorded by this device is different from the sieved particle size range (80–200 mm). For instance, this


Research Article

1725

Figure 5. Variation of particle number (counts) versus time; ammonium chloride (15 vol %, 80–200 mm, gray signals), L-threonine (15 vol %, 80–200 mm, black signals). Photos on the right side: (A) ammonium chloride, (B) L-threonine.

extent of back-reflected light in the 3D ORM and FBRM techniques depends on the optical properties of particles and solution, e.g., refractive index, morphology, and roughness of the particles. Since the used substances have not the same mentioned properties, the measured average length of L-threonine and ammonium chloride particles are not identical, even though the used particles result from the same sieve fraction. The presented results in Fig. 7 indicate, however, that the particle size measurement is relatively constant with time and reproducible. Average length signals in the 3D ORM measurement technique are derived from the measured exposed particle surface area (EPSA). Furthermore, needle-like par-

Figure 6. Variation of obscuration factor (OBF) versus time; ammonium chloride (15 vol %, 80–200 mm, gray signals), L-threonine (15 vol %, 80–200 mm, black signals).

Figure 7. Variation of mean chord length versus time; ammonium chloride (15 vol %, 80–200 mm, gray signals), L-threonine (15 vol %, 80–200 mm, black signals).

Table 2. Refractive indices for the substances used in this work (selected refractive indices at l = 589 nm).

ticlescan be detected by their length or diameter in this measurement technique because of different orientations to the focal point. Consequently, the average length signals for needle-like particles exhibit more fluctuations compared to non-needle-like particles. Comparison of average length signals for needle-like particles and non-needle-like particles demonstrates that it is possible from the 3D ORM measurement technique to differentiate particles with varying shapes (see Fig. 7) based on the fluctuation intensity of the signals. Furthermore, this difference between average length signals of two different shapes is still present after normalization of this graph (chord length/counts versus time; not shown here). In real multiphase systems, however, often different particle shapes are present which are difficult to distinguish by only one geometrical dimension. The APAS provides information on the exposed particle surface area (EPSA) with a further geometrical dimension (two dimensions) as illustrated in Fig. 8 by the application of a pulsed laser point. As can be seen in Fig. 8 A, particles with different shape provide signals with different intensity in the diagram. Comparing

Substance

Refractive index (solid)

Refractive index (solution)

Ammonium chloride

1.64 [13]

1.38 (20 C; saturated solution) (own measurement)

L-Threonine

1.52 [14]

1.35 (20 C; saturated solution) (own measurement)

measurement technique provides values for L-threonine particles between 65 and 140 mm and for ammonium chloride particles values of ~ 30 mm; see Fig. 7. This can be ascribed to the fact that the physical principle for measuring the particle size is not the same. The measurement principle of the FBRM and 3D ORM is based on the physical effect of light reflection. The



www.cet-journal.com

Research Article

1726

and (ii) L-threonine/water (distilled) as a substance with needle-like shape. Fig. 9 presents the number of particles (counts) for two systems and two different suspension densities. As can be concluded, the number of the particles (counts) increases with higher volume suspension density in these solutions. Since the FBRM measurement principle is based on the detected backward reflected laser light, the amount of backward collected light is a function of the optical properties of particles and solution, size and shape of the considered particles. Due to the varying properties, a difference in the particle numbers (counts) can be observed in these two systems.

4.3.2 Measurement of Mean Chord Length of L-Threonine and Ammonium Chloride Comparing the two signals in Fig. 10, a constant signal was achieved for both systems. It is obvious that the FBRM can Figure 8. In situ particle size analysis by a 3D ORM. (A) The APAS provides information measure the mean chord length of different on EPSA by application of a pulsed laser beam focus. By laser pulsation, this focus laser particle shapes over time. This mean chord point moves up and down. The actual surface of a particle (area), not just the edge of length was generated from the straight line particles (one dimension as recorded by the FBRM measurement technique), can be measured. The particle size is determined by the time the particle is positioned inside measurement from one edge to the other the pulsed laser beam focus (time of flight) and the signal intensity. During the time of edge of a particle. As described in the preflight, laser and particle are interacting and the laser light is reflected back. (B) The APAS vious section for the 3D ORM technique, with MCST measures not only the time of one signal (one dimension), but an EPSA the measured mean chord length by FBRM which is related to the actual surface of the particle. APAS allows the interpretation of is different from the sieved particle size shapes due to the different height of signal derived by the measured EPSA [3, 15]. range, too. A size, i.e., mean chord length, of ~ 26 mm was found for L-threonine particles and of ~ 35 mm for ammonium chloride particles. This the signals resulting from different particle shapes, it can be difference between average length signals of two different observed that needle-like particles have higher signal intensity and these particles can be detected by their length; see Fig. 8 A. This measurement technique is also able to detect different morphologies of particles, allowing, e.g., to follow a polymorphic transformation which is extremely important in many applications, e.g., for pharmaceuticals. In real time and inline, a shape discrimination technique exhibits many advantages such as no sampling is needed, measuring in original concentration without dilution. It can give especially useful extra information on important events, e.g., crystal growth and crystal breakage [3, 16].

4.3 FBRM Device 4.3.1 Measurement of Counts in L-Threonine and Ammonium Chloride Suspensions Based on the same volume suspension density and particle size range, experiments were performed under constant temperature to prove the effect of particle shape on the results of the FBRM measurement technique in the two systems: (i) ammonium chloride/water (distilled) as a substance with cubic shape

www.cet-journal.com

Figure 9. Variation of particle number (counts) versus time (primary mode); ammonium chloride (80–200 mm, gray signals), L-threonine (80–200 mm, black signals).



Research Article

1727

Figure 10. Variation of mean chord length versus time (primary mode); ammonium chloride (15 vol %, 80–200 mm, gray signals), L-threonine (15 vol %, 80–200 mm, black signals).

shapes is still present after normalization of this graph (chord length/counts versus time; not shown here). A 1D measuring technique such as the FBRM does not allow for distinguishing different morphologies. The FBRM technique provides statistically qualified information related to the chord length measurement. However, particles with different shapes could have approximately the same spherical equivalent diameter which can be calculated if desired from the chord length values. The orientation of a particle at the probe window plays an important role in determining the characteristic length from the measured chord length. The chord length found by this technique in a suspension of needle-like particles is a strong function of the needle width. The results in Fig. 10 support the assumption that the width of the needle-like particles has a higher probability than the length to be measured by the FBRM technique. For needle-like-shaped particles, however, no similar high fluctuation can be observed as with the 3D ORM measurement technique; see Fig. 7. The FBRM measurement technique can be applied to indicate polymorphic transformation by monitoring the change in chord length distribution or variation in the mean chord length signals, respectively. Changes in these signals affirm alterations in shape or size. It is, however, never clear that the mentioned signals belong to a specified particle shape [18–24]. The FBRM is a probe-based measurement tool which is installed directly in the crystallizer without the need for sample dilution or manipulation. It has also been successfully applied as a tool for detecting a nucleation event and characterization of the metastable zone width [1].

5

Conclusions

The suspension density and mean particle size are two of the most important parameters which have to be controlled carefully in order to manage crystallization processes. Powerful sensor technologies to analyze these process parameters inline and in real time are ultrasound, 3D ORM, and FBRM techniques. The sensitivity of these techniques with respect to the shape of particles is, however, different. The ultrasound technique does not exhibit sensitivity to the particle shape according to the measured signals (Fig. 4) but


the ultrasound velocity is sensitive to particle size and suspension densities. Furthermore, this technique is able to provide information on the metastable zone width and the solubility under different conditions. The 3D ORM gives information on the exposed particle area and from these data allows for calculation of the particle size. Particles with varying shapes deliver signals with different heights in the diagram (Fig. 8). Since needle-like particles can be detected either by their length or width based on their orientations to the focal point, the 3D ORM measurement technique indicates a noticeable fluctuation in the signals of average length. Therefore, a significant effect of the particle shape on the measured signals can be observed in this measurement technique. Consequently, it is possible to detect different morphologies of particles that could enable, e.g., the tracking of polymorphic transformations if it goes in line with a shape change of the particles which is, however, in most cases the point. This is extremely important for a number of applications. The FBRM device provides information related to chord length measurements which can subsequently be converted by the user if desired to a spherical equivalent diameter. Stable signals of the mean chord length for needle-like particles (L-threonine) can be recorded by the FBRM technique. Needle-like particles can be detected due to the flow pattern only by their width. Additionally, this measurement technique was successfully applied as useful tool for detecting nucleation events and characterization of the metastable zone width. By monitoring the change in the chord length distribution, mean chord length, or total counts of the particles, the FBRM technique can be employed to detect polymorphic transformations [17–23]. However, an absolute conclusion concerning the particle shape is not possible with this technique. The 3D ORM measurement technique can supply an average length of particles generated from the exposed particle surface area of the particle. The FBRM measurement technique can give the mean chord length of particles derived from the straight line signal from one edge to the other edge of a particle. However, these measured lengths of particles in 3D ORM and FBRM techniques are different from sieved particle size fractions since the measurement principles are not identical. For example, the principle of the FBRM and 3D ORM measurement techniques are based on the physical effect of the reflection of light. The extent of back-reflected light in the 3D ORM and FBRM measurement technique depends on the optical properties of particles and solution, e.g., refractive index, morphology, roughness, and purity of the particles. Since these mentioned properties are not the same for the various substances, accordingly the measured lengths are different, but reproducible.

Acknowledgment The authors would like to thank Mr. Schwartz from Sequip Company for his kind help in providing information. The authors have declared no conflict of interest.


www.cet-journal.com

Research Article

1728

[15]

References [1]

[2] [3] [4] [5] [6]

[7] [8] [9]

[10] [11] [12] [13] [14]

P. Barrett, B. Smith, J. Worlitschek, V. Bracken, B. O’Sullivan, D. O. Grady, Org. Process Res. Dev. 2005, 9 (3), 348–355. DOI: 10.1021/op049783p T. Stelzer, D. Pertig, J. Ulrich, Cryst. Growth 2013, 362, 71–76. DOI: 10.1016/j.jcrysgro.2011.11.027 L. Helmdach, F. Schwartz, J. Ulrich, Chem. Eng. Technol. 2014, 37, 1–8. DOI: 10.1002/ceat.201300190 P. Mougin, D. Wilkinson, K. J. Roberts, Cryst. Growth Des. 2003, 3 (1), 67–72. DOI: 10.1021/cg025578u A. S. Ahuja, W. R. Hendee, Acoust. Soc. Am. 1978, 63 (4), 1074–1080. DOI: 10.1121/1.381814 J. Ulrich, M. J. Jones, in Industrial Crystallization Process Monitoring and Control, (Eds: A. Chianese, H. J. M. Kramer), Wiley-VCH, Weinheim 2012, pp. 59–69. Sensotech GmbH, http//www.sensotech.com/ [accessed: 17.09.2003] M. J. W. Povey, Ultrasonic Technique for Fluid Characterization, Academic Press, New York 1997. D. Pertig, R. Buchfink, S. Petersen, T. Stelzer, J. Ulrich, Chem. Eng. Technol. 2011, 34 (4), 639–646. DOI: 10.1002/ ceat.201000558 P. Sayan, J. Ulrich, Chem. Eng. Process. 2002, 41 (3), 281–287. DOI: 10.1016/S0255-2701(01)00143-X J. Heinrich, Ph. D. Thesis, Martin Luther University HalleWittenberg 2008. http://www.chemicalbook.com/ChemicalProductProperty_ EN_CB5702648.htm/ [accessed: 16.09.2013] http://www.chemicalbook.com/ChemicalProductProperty_ EN_CB7129971.htm/ [accessed: 16.09.2013] J. J. Rodrigues Jr., L. Misoguti, F. D. Nunes, C. R. Mendonca, S. C. Zilio, Opt. Mater. 2003, 22 (3), 235–240. DOI: 10.1016/ S0925-3467(02)00270-7

www.cet-journal.com

[16]

[17] [18] [19]

[20] [21]

[22] [23] [24] [25]

[26] [27] [28]

F. Schwartz, J. Schwartz, Proc. of BIWIC 2012, (Eds: Z. Sha, Q. Yin, J. Chen, C. Xie), Tianjin University, Tianjin 2012, 165–170. T. Le-Minh, M. Schwartz, Proc. of BIWIC 2013, 20th Int. Workshop on Industrial Crystallization (Eds: H. Qu, J. Rantanen, C. Malwade), University of Southern Denmark, Odense 2013, 373–379. B. O’Sullivan, B. Glennon, Org. Process Res. Dev. 2005, 9 (6), 884–889. DOI: 10.1021/op0500887 W. Su, H. Hao, M. Barrett, B. Glennon, Org. Process Res. Dev. 2010, 14 (6), 1432–1437. D. Sathe, K. Sawant, H. Mondkar, T. Naik, M. Deshpande, Org. Process Res. Dev. 2010, 14 (6), 1373–1378. DOI: 10.1021/op100177s R. Kobayashi, Y. Fujimaki, T. Ukita, Y. Hiyama, Org. Process Res. Dev. 2006, 10 (6), 1219–1226. DOI: 10.1021/op060046y H. Hao, M. Barrett, Y. Hu, W. Su, S. Ferguson, B. Wood, B. Glennon, Org. Process Res. Dev. 2012, 16 (1), 35–41. DOI: 10.1021/op200141z J. Zhao, M. Wang, B. Dong, Q. Feng, C. Xu, Org. Process Res. Dev. 2013, 17 (3), 375–381. DOI: 10.1021/op300320a W. Liu, H. Wei, J. Zhao, S. Black, C. Sun, Org. Process Res. Dev. 2013, 17 (11), 1406–1412. DOI: 10.1021/op400066u J. Heinrich, J. Ulrich, Chem. Eng. Technol. 2012, 35 (6), 967–979. DOI: 10.1002/ceat.201100344 L. X. Yu, R. A. Lionberger, A. S. Raw, R. D’Costa, H. Wu, A. S. Hussain, Adv. Drug Delivery Rev. 2004, 56 (3), 349–369. DOI: 10.1016/j.addr.2003.10.012 M. Li, D. Wilkinson, K. Patchigolla, Part. Sci. Technol. 2005, 23 (3), 265–284. DOI: 10.1080/02726350590955912 M. R. Abu Bakar, Z. K. Nagy, C. D. Rielly, Cryst. Growth Des. 2010, 10 (9), 3892–3900. DOI: 10.1021/cg1002379 S. Titiz-Sargut, J. Ulrich, Chem. Eng. Process. 2003, 42 (11), 841–846. DOI: 10.1016/S0255-2701(02)00215-5

