Method and apparatus to improve analytical method development and sample preparation for reproducible particle size measurement

Abstract

A method and an apparatus to improve the precision and reproducibility of particle size analysis by laser diffraction is presented. Powder particles are typically prepared for laser diffraction testing using an ultra-sound bath which will disperse particle agglomerates and allow a precise measurement. However, the precision and reproducibility of agglomerate dispersion is affected by ultra-sound probe wear, corrosion and age. Differences in sonication performance can be compensated by voltage adjustments to the ultra-sound probe, leading to substantial improvements in the precision and reproducibility of particle size determination.

Claims

1. A method to develop and validate an analytical method relating to an ultra-sound probe to identify correct and reproducible operating parameters for de-agglomerating particles of a product of interest, comprising the steps of: i. applying a first power setting of the ultra-sound probe for sonicating a first sample of the product of interest, measuring and recording a first voltage produced by the ultra-sound probe at the first power setting; ii. sonicating the first sample at the first power setting whilst taking particle size measurements at timed intervals to track the progress of the de-agglomeration of the first sample; iii. selecting a second power setting of the ultra-sound probe which is different from the first power setting, measuring and recording a second voltage produced by the ultra-sound probe at the second power setting, sonicating a second sample of the product of interest at the second power setting whilst taking particle size measurements at timed intervals to track the progress of the de-agglomeration of the second sample; iv. choosing a power level and time for the particle type for optimum de-agglomeration based on results from the de-agglomeration of the first sample and the second sample; v. determining a validated voltage measured on the ultra-sound probe corresponding to the selected power level; and vi. establishing the validated voltage and selected time as the settings for sonication of the particle of interest.

2. The method according to claim 1, comprising repeating the selection of the power settings for further samples of the product of interest and taking further particle size measurement at timed intervals to track the progress of the de-agglomeration of the further, and selecting the power level and the time for the particle type for optimum de-agglomeration being selected based on the particle size measurement data collected for the power settings.

3. The method of claim 1, wherein the product of interest is a pharmaceutical powder.

4. A method for producing ultra-sound energy to disperse particle agglomerates suspended in a liquid medium using the settings for validated voltage and selected time established using the method of claim 1.

5. The use method of claim 4 further comprising a preparation of a suspension of a sample of a powder of interest in the liquid medium.

6. The method of claim 5, further comprising measurement of the particle size of the suspension by laser diffraction.

7. The method of claim 4, wherein the further comprising adjusting and setting of the settings to compensates for wear and corrosion of the probe.

Description

DESCRIPTION OF THE FIGURES

(1) FIG. 1 shows the laser diffraction test equipment of the art.

(2) FIG. 2 shows the laser diffraction test equipment of the invention.

(3) FIG. 3 is a flow diagram of the method of the current invention, in routine testing.

(4) FIG. 4 is a flow diagram of the method of the current invention in method development and validation.

(5) FIG. 5 shows the distribution of particle sizes for a first compound for 100% power in different equipment.

(6) FIG. 6 shows the distribution of particles sizes of the first compound when the power is adjusted to maintain a constant ultra-sound probe energy.

(7) FIG. 7 shows the distribution of particle sizes for a second compound for 20% power in different equipment.

(8) FIG. 8 shows the distribution of particles sizes of the second compound when the power is adjusted to maintain a constant ultra-sound probe energy.

DETAILED DESCRIPTION OF THE INVENTION

(9) The invention will now be described on the basis of the drawings. It will be understood that the embodiments and aspects of the invention described herein are only examples and do not limit the protective scope of the claims in any way. The invention is defined by the claims and their equivalents. It will be understood that features of one aspect or embodiment of the invention can be combined with a feature of a different aspect or aspects and/or embodiments of the invention.

(10) FIG. 1 shows a representation of the laser diffraction test equipment according to the prior art. A power controller command 1a comprises a percent power scale 1b which is used to set a given power level to be transmitted to an electronic motherboard 2 which calculates a desired voltage level from a user input. This calculated voltage level is communicated to an electronic power controller 4 via data channels 3 (which contain additional information about alarms and filters). The functions of the electronic motherboard 2 and of the electronic power controller 4 are shown here as being located in different hardware components, but need not be. In other embodiments, the functions of the electronic motherboard 2 and of the electronic power controller 4 could be supported in single component.

(11) The electronic power controller 4 is connected to a power supply 5 which the electronic power controller 4 uses to generate the calculated voltage which is then sent to an ultra-sound probe 6 of an ultra-sound bath 12 containing a suspension 7 of particles 8 for a product. The particles 8 are shown in FIG. 1 to be of variable sizes to represent imperfect sonication. The sonicated particles 8 are then channeled or taken to a laser diffraction equipment 9 for particle size measurements, which appear recorded as particle size distributions on histograms 10. Widely spaced histograms 10 represent the variability in size measurements taken of the same sample of product in different laser diffraction equipment and the poor precision and reproducibility of the prior art method and apparatus.

(12) FIG. 2 shows a representation of the laser diffraction test equipment according to the present description. The same reference numerals are used in FIG. 2 and in FIG. 1 to illustrate the same or similar elements. In the power controller command 1a the percent power scale 1b has been set to give a higher power level than the one used in the prior art apparatus of FIG. 1, to compensate for the poor sonication and results variability seen there. The difference between the prior art apparatus shown in FIG. 2 and the apparatus of this document is the presence of the voltmeter 11 which is connected to ultra-sound probe 6.

(13) In certain equipment, a voltage measurement function may be included in the existing hardware, obviating the need to install a voltmeter. Similarly, a probe controller comprising those elements needed to control the ultra-sound probe 6 can be constructed. The probe controller could include the voltmeter 11, the power controller 4 as well as a comparator for comparing the actual (measured) voltage with the validate voltage and adjust the power controller 4, as will be explained below. The probe controller could be integrated with the ultra-sound probe 6.

(14) In use, an analyst will check the voltage displayed in voltmeter 11 and set the percent power scale 1b until the voltmeter 11 displays the validated voltage level determined during a method development and validation process. This adjustment could also be carried out automatically, using a feedback loop in a comparator. The power level calculation, signal transmission and power generation are effected using elements 2, 3, 4, and 5 in the same manner as that described in FIG. 1. The voltage is then transmitted to the ultra-sound probe 6 which sonicates the product particles 8 contained in suspension 7. The particles 8 are shown of regular sizes to represent improved sonication. The sonicated particles 8 are then channeled or taken to a laser diffraction equipment 9 for particle size measurements, which appear recorded as particle size distributions on histograms 10. Closely spaced histograms 10 represent the improved precision and reproducibility of the method and apparatus of the present document.

(15) FIG. 3 shows a flow diagram of the method to maintain correct sonication power. In step 100, a power value is applied using the power controller command 1a to the ultra-sound probe 6 in the ultra-sound bath 12. The actual voltage produced by the ultra-sound probe 6 is measured in step 110 using the voltmeter 11 connected to the ultra-sound probe 6. The voltage of the ultra-sound probe 6 is compared with validated voltage in step 120. Should a difference be established then the voltage at the ultra-sound probe 6 is maintained at a desired level by reading the measured actual voltage and setting the current delivered to the ultra-sound probe 6 in step 130 using the electronic power controller 4 until the validated voltage is reached.

(16) FIG. 4 shows a flow diagram of the manner in which an analyst can develop and validate an analytical method relating to the ultra-sound probe 6 to identify correct and reproducible operating parameters for de-agglomerating particles of a product of interest. The product of interest includes, but is not limited to a pharmaceutical product, such as a pulmonary pharmaceutical.

(17) In a first step 200, a first power setting of the ultra-sound probe 6 for sonicating a first sample of the product of interest is chosen. A first voltage produced by the ultra-sound probe 6 at the first power setting is measured and recorded in step 210. The first sample is sonicated in step 220 at the first power setting whilst taking in step 230 particle size measurements at timed intervals to track the progress of the de-agglomeration of the first sample.

(18) The process is repeated several times from step 240 using a further sample from the same batch at further power settings until a sufficient number of measurements have been made.

(19) In step 250, a power level and time for the particle type for optimum de-agglomeration based is chosen based on results from the de-agglomeration of the samples. The validated voltage measured on the ultra-sound probe 6 corresponding to the selected power level is determined in step 260 these values are established in step 270 as the validated voltage and selected time as the settings for sonication of the particle of interest.

(20) The invention will now be described further with reference to two examples, which also contain explanations about FIGS. 5, 6, 7, and 8.

Example 1

(21) The method of the present invention was tested in five different laser diffraction machines, being Malvern Mastersizer 2000 and Mastersizer 3000 machines manufactured by Malvern Instruments Ltd (Malvern, United Kingdom). The Malvern Mastersizer 3000 was also used in Mastersizer 2000 mode, so six sets of data were obtained. Each of the machines had its own built-in ultra-sound bath, with ultra-sound probes of varying ages, and they were used to sonicate the samples prior to testing.

(22) The same batch of a pulmonary inhalation drug product was used for the testing, and this batch was obtained using a validated manufacturing process of wet polishing yielding the highest particle size precision and reproducibility. When observed by optical microscopy, these particles had similar physical and size characteristics.

(23) Table 1 below contains the results obtained using the method of the prior art, with the power level of the ultra-sound probe being set conventionally with a percent scale of maximum power. Table 2 contains the results obtained using the improved method of the present invention.

(24) The drug product was tested in all machines following the conventional method where power is set at the same value in the percentage power scale 1b of the controller. (For the purpose of obtaining information about the voltage actually transmitted to the ultra-sound probe 6, a voltmeter was connected to the ultra-sound probe 6. However, the voltmeter was not used to adjust the percent power scale on the power controller command 1a.)

(25) Samples of the same drug product batch were suspended in water with the addition of an appropriate dispersant in the built-in ultra-sound bath of each of the machines (four Malvern Mastersizer 2000 units, identified as CM06, CM09, CMOS and CM10 and one Malvern Mastersizer 3000 identified as CM10), sonicated and then size-tested by laser diffraction.

(26) The test measured the size of the particles in the sample and then computed its distribution expressed as D10, D50 and D90. When the particles 8 are arranged on an ascending mass basis, these values are representative of the particle size, in microns, of the diameter of the largest particle found in 10%, 50% and 90% of the mass of particles and they are highly characteristic of the particle size distribution and thus of the quality of the product. The data are shown in Table 1.

(27) TABLE-US-00001 TABLE 1 % Probe D10 D50 D90 Equipment Power Energy [V] (m) (m) (m) CM06 100 130 0.782 1.639 3.375 CM09 100 150 0.761 1.628 3.324 CM02 100 155 0.851 1.79 3.641 CM05 100 152 0.814 1.703 3.308 CM10 (2000 mode) 100 151 1.439 3.026 5.783 CM10 (3000 mode) 100 151 1.079 2.631 4.953 Mean 0.9543 2.0695 4.0640 Standard deviation 0.2638 0.6038 1.0505 Relative std. dev. 27.64% 29.18% 25.85%

(28) The power was set at 100% of maximum power, as per the original method development data. The ultra-sound probe energy level was measured using a voltmeter and ranged from 130 to 155 V. Columns D10, D50 and D90 indicate the size data in each of the three size classes, given by each of the six series of tests in the five laser diffraction machines. Processing of the D10, D50 and D90 data yielded a mean, a standard deviation and a relative standard deviation (standard deviation/mean).

(29) Significantly, the relative standard deviation ranged from 26% to 29%. These are values indicating high variability, although the particles 8 tested all came from the same batch, of known homogenous physical characteristics. These data therefore suggest that the laser diffraction test method was imprecise, had low reproducibility and therefore was not reliable.

(30) The same batch of product was then retested in the same laser diffraction machines after the sonication power level was appropriately adjusted for each one of them as per the method of the present invention.

(31) Table 2 shows the results obtained from following the method of the invention whereby the power is set by measuring the actual voltage produced by the probe and adjusting it using the percent power setting until it is at the value prescribed by the method development and validation process.

(32) TABLE-US-00002 TABLE 2 % Probe D10 D50 D90 Equipment Power Energy [V] (m) (m) (m) CM06 100 130 0.782 1.639 3.375 CM09 80 130 0.83 1.781 3.581 CM02 75 130 0.758 1.693 3.847 CM05 85 130 0.688 1.498 3.215 CM10 (2000 mode) 61 130 0.784 1.869 3.831 CM10 (3000 mode) 61 130 0.651 1.826 3.731 Mean 0.7488 1.7177 3.5967 Standard deviation 0.0668 0.1369 0.2572 Relative std. dev. 8.91% 7.97% 7.15%

(33) The data were obtained under the method of the invention and were again computed in the same manner as for Table 1.

(34) The percent power scale was adjusted in each of the ultra-sound baths, so that the ultra-sound probe energy level measured by a voltmeter would read a constant 130 V. In order to set this validated voltage, the percent power scale 1b had to be set at values ranging from 61% to 100%. (It is interesting to note that the method was originally developed in machine CM06, at maximum power, suggesting CM06 was operated with a worn or corroded ultra-sound probe 6. This appears to be confirmed by the fact that the ultra-sound voltages used for the other machines in table 2 were set at a much lower percent power level to achieve the same voltage of 130 V).

(35) In this second series of tests, the mean particle for each of the three size classes of D10, D50 and D90 were consistently smaller, (21.5%, 17% and 11.5% smaller, respectively). More significantly, the relative standard deviation in the tests measurements in each of the D10, D50 and D90 size classes ranged from 7% to 9%, approximately a 3-fold improvement in reproducibility over the data of table 1 obtained with the known method.

(36) The fact that the reproducibility in measurement data from five different machine is three times better when using the inventive method is an indication of its higher precision.

(37) The improvement can also be seen in FIGS. 5 and 6, where the data of Tables 1 and 2 have been plotted, respectively. The vertical bars represent particle size for each of the three size classes D10, D50 and D90 and the error bars represent the distance of the maximum data point to the mean and the distance of the minimum data point to the mean. FIG. 4 illustrating the present invention shows an evident improvement over FIG. 3, with a smaller amplitude of the error bars.

(38) The data indicate a remarkable and long overdue improvement in particle size analysis by laser diffraction. They give proof to the benefits of the present invention by demonstrating the improved precision and reproducibility of particle size measurements of the same product across different ultra-sound baths and laser diffraction machines.

Example 2

(39) The experiments of Example 1 were repeated to test a different compound, using four laser diffraction machines. The methods were the sameprior art and inventive methods.

(40) The same batch of an undisclosed drug product was used for the testing, and this batch was obtained using a validated manufacturing process of a size reduction process yielding known particle size precision and reproducibility. When observed by optical microscopy, these particles had similar physical and size characteristics.

(41) Table 3 below contains the results obtained using the method of the prior art, with the power level of the ultra-sound probe being set conventionally with a percent scale of maximum power. Table 4 contains the results obtained using the improved method of the present invention.

(42) The drug product was tested in all machines following the conventional method where power is set at the same value in the percentage power scale of the controller. Samples of the same drug product batch were suspended in an appropriate anti-solvent with the addition of an appropriate dispersant in the built-in ultra-sound bath of each of the machines (Malvern Mastersizer 2000 units, identified as CM06, CM09, CM02 and CMOS), sonicated and then size-tested by laser diffraction.

(43) The data are shown in Table 3.

(44) TABLE-US-00003 TABLE 3 % Probe D10 D50 D90 Equipment Power Energy [V] (m) (m) (m) CM06 20 39 12.271 46.846 92.688 CM09 20 60 9.787 37.519 75.988 CM02 20 41.6 12.667 44.53 90.043 CM05 20 42.8 12.095 42.69 86.662 Mean 11.7050 42.8963 86.3453 Standard deviation 1.3008 3.9677 7.3320 Relative std. dev. 11.11% 9.25% 8.49%

(45) The power was set at 20% of maximum power, as per the original method development data. The energy level was measured using a voltmeter and ranged from 39 to 42.8 V. Columns D10, D50 and D90 indicate the size data in each of the three size classes, given by each of the series of tests in the four laser diffraction machines. Processing of the D10, D50 and D90 data yielded a mean, a standard deviation and a relative standard deviation (standard deviation/mean).

(46) In this example, the relative standard deviation ranged from 8% to 11%. These are values indicating medium variability of the size determination method.

(47) The same batch of product was then retested in the same laser diffraction machines after the sonication power level was appropriately adjusted for each one of them as per the method of the present invention.

(48) Table 4 shows the results obtained from following the method of the invention whereby the power is set by measuring the actual voltage produced by the probe and adjusting it using the power setting until it is at the validated, prescribed value.

(49) TABLE-US-00004 TABLE 4 % Probe D10 D50 D90 Equipment Power Energy [V] (m) (m) (m) CM06 22 43 10.927 41.448 85.625 CM09 8 43 10.725 41.135 84.004 CM02 21 43 12.377 43.738 87.992 CM05 19 43 12.089 42.76 87.096 Mean 11.5295 42.2703 86.1793 Standard deviation 0.8249 1.2054 1.7479 Relative std. dev. 7.15% 2.85% 2.03%

(50) The data were obtained under the method of the invention and were again computed in the same manner as for Table 3.

(51) The percent power scale was adjusted in each of the ultra-sound baths, so that the energy measured by a voltmeter would read the validated voltage value of 43 V. In order to set this voltage, the percent power scale had to be set at values ranging from 19% to 22%.

(52) In this second series of tests, the mean particle for each of the three size classes of D10, D50 and D90 were of the same size as with the method of the prior art, but the relative standard deviation in the tests measurements in the D10 size class showed a slight improvement (from 11.11% to 7.15%), while in the D50 size class it showed a 3-fold improvement and in the D90 size class a 4-fold improvement over the data of table 3 obtained with the known method.

(53) The improvement can also be seen in FIGS. 7 and 8, where the data of Tables 3 and 4 have been plotted, respectively, in the same manner as the figures illustrating the data of the previous example. FIG. 6 illustrating the present invention shows an evident improvement over FIG. 5, with a smaller amplitude of the error bars.

(54) The data for this example indicate that the same improvement in method precision and reproducibility as seen in Example 1 could be achieved with a different drug.

Example 3

(55) The sonication performance of a new ultra-sound probe was compared with the sonication performance of an ultra-sound probe of undetermined age.

(56) Two different batches of the same pharmaceutical drug product were tested. First, the samples were sonicated using an ultra-sound probe of undetermined age and tested by laser diffraction using a Malvern Mastersizer 2000 machine (CM02). Then, samples of the same two batches were sonicated using a new ultra-sound probe, in a different Malvern Mastersizer 2000 (CMOS) but no power adjustment was applied. Finally, the samples were tested again in CMOS, but this time the power of the new ultra-sound probe was adjusted so that the resulting voltage would match the voltage obtained in the first sonication and test of each of the batches. The particle size (PS) data is in tables 5 and 6.

(57) TABLE-US-00005 TABLE 5 % Probe D10 D50 D90 Equipment Power Energy [V] (m) (m) (m) Mastersizer 2000 - CM02 20% 44 V 9.862 35.561 76.775 with probe of undetermined old age Mastersizer 2000 - CM05 20% 54 V 8.031 31.262 68.291 with new probe, no power adjustment % difference between old 18.57% 12.09% 11.05% and new probe PS data Mastersizer 2000 - CM05 7% 44 V 8.837 32.873 71.429 with new probe, with power adjustment % difference between old 10.39% 7.56% 6.96% and new probe PS data

(58) Compared to the first test carried out after sonication using an ultra-sound probe of undetermined age in CM02, particle size in the second test carried out after sonication using a new ultra-sound probe in CMOS was found to be smaller by 11% to 18.6% in each of the three size classes.

(59) When the third test was carried out again in CM5 with a new ultra-sound probe but this time resorting to power adjustment, the particle size was found to be smaller but by a lower margin, 7% to 10.4% in each of the three size classescloser to the original measurement in CM02.

(60) Note that the new adjusted ultra-sound probe is producing 44V but only requires an input of 7% on the power scale to do so, as opposed to the CM02 sonicator which required 20% power to produce 44V. This confirms that the power output difference between the old probe and the new probe is real.

(61) TABLE-US-00006 TABLE 6 Probe % Energy D10 D50 D90 Equipment Power [V] (m) (m) (m) Mastersizer 2000 - 20% 44 V 6.675 24.841 59.38 CM02 with probe of undetermined old age Mastersizer 2000 - 20% 54 V 6.178 22.589 52.878 CM05 with new probe, no power adjustment % difference between 7.45% 9.07% 10.95% old and new probe PS data Mastersizer 2000 - 7% 44 V 6.39 23.73 56.803 CM05 with new probe, with power adjustment % difference between 4.27% 4.47% 4.34% old and new probe PS data

(62) Table 6 shows the test data of the second batch of the same product, following the same test protocol.

(63) Compared to the first test carried out after sonication using an ultra-sound probe of undetermined age in CM02, particle size in the second test carried out after sonication using a new ultra-sound probe in CMOS was found to be smaller by 7.4% to 11% in each of the three size classes.

(64) When the third test was carried out again in CM5 with a new ultra-sound probe but this time resorting to power adjustment, the particle size was found to be smaller but by a lower margin, around 4% in each of the three size classescloser to the original measurement in CM02.

(65) The data in tables 5 and 6 indicates that reproducibility error was reduced by adjusting percent power so as to ensure the use of a constant voltage.

REFERENCE NUMERALS

(66) 1a Power controller command 1b Power scale 2 Motherboard 3 Data channels 4 Electronic power controller 5 Power supply 6 Ultra-sound probe 7 Suspension 8 Particles 9 Laser diffraction equipment 10 Histograms 11 Voltmeter 12 Ultra-sound bath

BIBLIOGRAPHY

(67) K. Mindgard, R. Morrell, P. Jackson, S. Patel and R. BuxtonMeasurement Good Practices Guide No 111Good Practice Guide for Improving the Consistency of Particle Size MeasurementISSN 1368-6550 Amy SabianThe Particle Experts, Problems in particle size: Laser Diffraction ObservationsAutumn 2011, Vol 15/No. 04) F. Storti and F. BalsamoParticle size distributions by laser diffraction: sensitivity of granular matter strength to analytical operating procedures. Published in Solid Earth Discussions, 19 Apr. 2010). Van der GraafSonocrystallization, Nucleation of ammonium sulfate and alfa-lactose monohydrate due to ultrasonic irradiationMaster of Science thesis, Delft University of Technology, January 2011). Lawrence C. LynnworthIndustrial Application of Ultra-soundA Review. II. Measurements, Tests, and Process Control Using Low Intensity Ultra-soundIEEE Transactions on Sonics and Ultrasonics, Vol SU-22, no. 2, March 1975 Patrick DunneNonthermal Processing Technology for foodISBN: 978-0-8138-1668-5, October 2010)