Memorandum
`To: Mahmoud Kaba
`From: Stephen Cypes
`Date: January 23, 2003
`Re: Weekly Progress Report, 01/20/03 01/24/03
`
`L-224,715
`Salt Break Step
`
`Loss in the aqueous cut has been optimized by setting the THF:Water volume ratio at 4:1 (this was detailed
`in last week s memo). However, the idea has been presented to crystallize the pure phosphate salt directly
`from the washed THF cut. This would remove a lengthy solvent switch from the overall batch time. Since
`THF is hydrophilic, the weight percent of NaCl in the final brine wash needed to be optimized to minimize
`the amount of water in the THF cut if this will be used in the pure crystallization step. An experiment was
`run to determine how NaCl content affected the amount of water in the THF cut (determined from the KF
`of the cut). The resulting relationship is shown in the graph below.
`
`Water Content in THF Cut versus NaCl Content in
`Brine Wash
`
`5.0%
`
`15.0%
`10.0%
`Weight% NaCl in Water
`
`20.0%
`
`25.0%
`
`180
`160
`140
`120
`100
`80
`60
`40
`20
`0
`0.0%
`
`Water Content of THF Cut
`
`(mg/mL)
`
`By using a nearly saturated brine solution, the water content in the THF cut can be reduced to ~6% by
`volume.
`
`Also, as of last week the total volume needed for the salt break was 20 L/kg (16 L/kg for the THF cut and 4
`L/kg for the water cut). An experiment was run to reduce the total volume to 10 L/kg while remaining at
`4:1 by volume THF:Water. The cut was much less clear, and both cuts were a discolored-yellow due to the
`increased concentration. The final volume for the salt break will most likely be around 15 L/kg after more
`experiments are performed.
`
`L-224,715
`Pure Phospate Salt Crystallization
`
`About 30 g of free base was isolated in THF to prepare for several crystallization studies using different
`solvents. Part of this free base batch was solvent switched to tert-butanol, another part to MIBK, and the
`rest remained in THF. Angela Spartalis has presented the relationship between the polymorphs for the pure
`phosphate salt and how ethanol, the solvent used in the first campaign, interacts with the crystal to form a
`
`Merck Exhibit 2126, Page 1
`Mylan Pharmaceuticals Inc. v. Merck Sharp & Dohme Corp.
`IPR2020-00040
`
`

`

`solvate. The idea, shown in the diagram below, is to try different solvents that may not form a solvate with
`the phosphate salt and instead go directly to Form I.
`
`EtOH solvate
`
`+ EtOH
`
`- EtOH
`
`Form II
`
`>35C
`
`THF
`MIBK
`tert-butanol
`Another
`Solvent???
`???
`
`Form I
`
`>80C
`
`<80C
`
`Type A
`
`To start the crystallization studies, the solubility of phosphate salt in each of the solvents above was
`determined at different temperatures. The results are summarized in the table below.
`
`Solvent
`
`Temperature
`
`THF
`THF
`THF
`
`MIBK
`MIBK
`MIBK
`
`t-butanol
`t-butanol
`
`20C
`40C
`50C
`
`20C
`40C
`60C
`
`40C
`60C
`
`Phosphate Salt
`Solubility (mg/g)
`0.39
`0.42
`0.52
`
`0.28
`0.47
`2.36
`
`0.21
`0.46
`
`Since the solubility is extremely low in all the solvents, they all seem reasonably suitable for the
`crystallization. This week, studies with THF and MIBK have been performed in the lab based on the
`following flowsheets:
`
`THF Initial Crystallization Study
`
`MIBK Initial Crystallization Study
`
`Free Base in THF
`215 mg/g
`6 g FB in 27.9 g Solution
`(~28 mL Solution)
`Flow Rate = 14 mL/hr
`
`1.78 g 85 weight % Aqueous
`H3PO4 Added to THF to
`Make 28 mL of Solution
`Flow Rate = 14 mL/hr
`
`Free Base in MIBK
`235 mg/g
`6 g FB in 25.5 g Solution
`(~27 mL Solution)
`Flow Rate = 14 mL/hr
`
`1.78 g 85 weight % Aqueous
`H3PO4 Added to MIBK to
`Make 27 mL of Solution
`Flow Rate = 14 mL/hr
`
`HEEL:
`1.5 g Phosphate Salt
`Suspended in 20 mL THF
`
`Assume:
`~8.5 L/kg total volume
`8.95 g total (heel+formed) in vessel
`Negligible lost to supernatent
`Seed crystals are ~17% of total
`phosphate salt in vessel
`1.05 eq H3PO4 added
`
`HEEL:
`1.5 g Phosphate Salt
`Suspended in 20 mL MIBK
`
`Assume:
`~8.5 L/kg total volume
`8.95 g total (heel+formed) in vessel
`Negligible lost to supernatent
`Seed crystals are ~17% of total
`phosphate salt in vessel
`1.05 eq H3PO4 added
`
`Merck Exhibit 2126, Page 2
`Mylan Pharmaceuticals Inc. v. Merck Sharp & Dohme Corp.
`IPR2020-00040
`
`

`

` Samples were submitted for: PSD, purity, and XRD analysis to determine crystal polymorph. Microscopy
`results of the crystal structures are shown in the pictures below (most crystals are about 20 microns along
`their longest axis):
`
`Crystallization from MIBK
`
`Crystallization from THF
`
`Next week s experiments include:
`
`Running crystallization experiment for tert-butanol.
`Stability of free base in MIBK.
`Performing salt break at 15 L/kg total volume.
`Complete characterization and comparison of
`crystallization.
`
`three solvents
`
`tested for pure step
`
`Merck Exhibit 2126, Page 3
`Mylan Pharmaceuticals Inc. v. Merck Sharp & Dohme Corp.
`IPR2020-00040
`
`

`

`Memorandum
`To: Mahmoud Kaba
`From: Stephen Cypes
`Date: January 30, 2003
`Re: Weekly Progress Report, 01/27/03 01/31/03
`
`L-224,715
`Salt Break Step
`
`The current CSA Salt Break step has been optimized with regard to losses in the aqueous cuts as well as
`water content in the final THF cut. The process flowsheet for this step is shown below.
`
`2 eq KOH in H2O
`with 5 weight%
`NaCl (3 L/kg
`CSA-salt)
`
`L-715 CSA-salt in
`THF slurry (12 L/kg)
`
`Agitate 30 min.
`
`L-715 free base in
`THF solution
`
`Salts in Aqueous
`Solution
`
`Wash with 20 weight%
`NaCl aqueous solution and
`cut aqueous layer
`
`Aqueous Cut (waste)
`
`The following table summarizes the assay results for each operation in the salt break step.
`
`First Salt Break Cut
`THF Cut
`Aqueous Cut
`64.5
`0.115
`657.5
`109.6
`750
`90
`120
`--
`
`Final Brine Wash
`THF Cut
`Aqueous Cut
`70.8
`0.74
`598.6
`235.9
`650
`210
`63.3
`--
`
`Free Base Concentration
`(mg/g)
`Weight of Cut (g)
`Volume of Cut (mL)
`Water Content (mg/mL)
`Percent of Free Base
`Determined by HPLC Assay
`Mass Balance
`101%
`(HPLC Assay/True Value)
`
`99.97%
`
`0.03%
`
`99.60%
`
`0.41%
`
`102%
`
`L-224,715
`Pure Phospate Salt Crystallization
`
`Solvent screening has continued in order to identify a solvent system that would not form a solvate with the
`phosphate salt and therefore could be crystallized directly as either the Form I or Form III polymorph.
`Directly crystallizing one of these forms would greatly reduce the overall batch cycle time.
`
`40 g of free base was isolated in THF to prepare for crystallization studies using different solvents. MIBK
`was the most promising solvent this week, since it did not form a solvate with the phosphate salt and did
`not result in any amorphous material. During a reactive crystallization procedure with a heel of
`
`Merck Exhibit 2126, Page 4
`Mylan Pharmaceuticals Inc. v. Merck Sharp & Dohme Corp.
`IPR2020-00040
`
`

`

`predominately Form I material, the resulting crystals were all Form I. A second crystallization was run
`with a heel of predominately Form III material at a lower temperature. The resulting polymorph of these
`crystals is still pending. Unfortunately, it was discovered that both the free base and phosphate salt are
`unstable in MIBK, even at low temperatures. It is currently thought that a shift base reaction is occurring
`and may exclude any ketone solvent for use in this crystallization.
`
`The table below lists all solvents that have been screened thus far. Besides not allowing a solvate to form,
`the solvent must not form amorphous material, must be able to dissolve either the free base or phosphate
`salt, and must not degrade the phosphate salt.
`
`Free Base
`Soluble?
`N
`
`Phosphate
`Salt Soluble?
`Y
`
`Degrades
`Product?
`N
`
`Becomes
`Form a
`Solvent
`Amorphous?
`Solvate?
`N
`N
`Water
`N
`Y
`Methanol
`N
`Y
`Ethanol
`N
`Y
`1-Propanol
`N
`Y
`2-Propanol
`Y
`N
`t-Butanol
`Y
`N
`Cyclohexanol
`N
`Y
`Ethyl Acetate
`N
`Y
`Acetone
`N
`N
`MIBK
`N
`Y
`Butyl Ether
`N
`Y
`THF
`N
`N
`n-Hexane
`N
`N
`Cyclohexane
`N
`N
`n-Heptane
`Y
`N
`Toluene
`N
`Y
`Acetonitrile
`N
`Y
`DMF
`N
`Y
`DMAC
`?
`N
`DMSO
`?
`N
`MTBE
`N
`Y
`?
`N
`Methylene Chloride
`From the table above, there is only one solvent thus far that meets all the requirements: water. However,
`since the phosphate salt is soluble in water, this is not a viable choice for crystallization since mother liquor
`losses would be too high.
`
`N
`
`Y
`
`N
`
`Y
`
`Y
`
`Y
`
`Y
`N
`N
`N
`Y
`
`N
`
`N
`
`N
`N
`N
`N
`Gels
`
`With the knowledge in the table above, I have produced the following general process for choosing a
`solvent to screen:
`Dielectric Constant above 3.0.
`Leads to a likely solvent to dissolve either free base and/or phosphate salt.
`Molecular Shape
`Six-membered rings and derivatives thereof tend not to form solvates
`Long-chained molecules tend not to form solvates
`Extremely small molecules (i.e. water) may not form solvates
`Highly branched molecules tend not to form solvates
`Eliminate all ketones due to possible degredation from shift base reaction.
`
`With this process, solvents such as methylene chloride, MTBE, DMSO, methyl benzoate, 1,2-
`dimethoxyethane and a number of other common solvents should be screened.
`
`Merck Exhibit 2126, Page 5
`Mylan Pharmaceuticals Inc. v. Merck Sharp & Dohme Corp.
`IPR2020-00040
`
`

`

`Next week s experiments include:
`
`Continue screening solvents for desired properties for use in pure crystallization
`Continue to update screening process based on real-time experimental results as more is
`learned about the solvent behavior with the phosphate salt
`
`Merck Exhibit 2126, Page 6
`Mylan Pharmaceuticals Inc. v. Merck Sharp & Dohme Corp.
`IPR2020-00040
`
`

`

`Memorandum
`To: Mahmoud Kaba
`From: Stephen Cypes
`Date: February 6, 2003
`Re: Weekly Progress Report, 02/03/03 02/07/03
`
`L-224,715
`Rhodium Removal from Hydrogenation Stream
`
`The use of the Biotage Cartridge System to remove Rh using a fixed bed adsorber is to be studied for use in
`Campaign III, and potentially Campaign II if lab-scale development moves quickly. Ponder s End has
`successfully used Biotage s FLASH systems to remove impurities from processing streams, and this
`ultimately led to less material use and greater operator safety.
`
`Upon discussions with Chris Welch, it has been decided that Ecosorb C941 has the greatest affinity for Rh
`removal from the crude hydrog stream. I have been in contact with John Weaver at Biotage and we will be
`receiving cartridges pre-packed with C941 (as well as other resins that John thinks is appropriate to screen)
`sometime next week.
`
`I have obtained a FLASH 12+M unit from Chris Weaver as a loan. This has a 17 cc working bed volume,
`and comes with 1/16 diameter inlet and outlet tubing with male Luer Lock connections. To reduce the
`cost of ordering the entire FLASH 12+ system (~$3,500), I ordered several additional components in order
`to start lab-scale development.
`FLASH+ Start-up kit ($375) Includes a rack with additional tubing in order to facilitate
`easy collection of effluent fractions
`Male-to-female Luer lock adapter to directly attach 100cc glass syringe to inlet of packed
`bed.
`
`These items should arrive by mid-week. Experiments will begin using a syringe pump to supply constant
`flow to the packed bed and producing a breakthrough curve for Rh in C941.
`
`Also this week, using data from an experiment performed by Chris Welch, the adsorption isotherm for Rh
`on C941 was produced.
`
`L-224,715
`CSA Salt Break
`
`An additional salt break was performed under identical process conditions as described in last week s
`report. Samples were submitted to AR for official concentration results.
`
`Cut
`
`First THF
`First Aqueous
`Second THF
`Second Aqueous Wash
`
`Free Base
`Concentation
`(g/L)
`58.84
`0.06
`61.46
`0.60
`
`Percent of
`Total Free
`Base
`99.986%
`0.014%
`99.736%
`0.264%
`
`Mass
`Balance for
`Cut
`106.0%
`
`103.5%
`
`Merck Exhibit 2126, Page 7
`Mylan Pharmaceuticals Inc. v. Merck Sharp & Dohme Corp.
`IPR2020-00040
`
`

`

`L-224,715
`Pure Crystallization
`
`Solvent screening continued to determine a solvent system that would crystallize either Form I and/or Form
`III directly by not forming a solvate with the 715-phosphate salt. A summary of the most current results
`can be found in the following table.
`
`Solvent
`
`Form a
`Solvate?
`
`Becomes
`Amorphous?
`
`Free Base
`Soluble?
`
`Water
`Methanol
`Ethanol
`1-Propanol
`2-Propanol
`t-Butanol
`Cyclohexanol
`Ethyl Acetate
`Acetone
`MIBK
`Butyl Ether
`THF
`n-Hexane
`Cyclohexane
`n-Heptane
`Toluene
`Acetonitrile
`DMF
`DMAC
`DMSO
`MTBE
`Methylene Chloride
`Diethoxymethane
`Methyl Benzoate
`Isoamyl Alcohol
`
`N
`Y
`Y
`Y
`Y
`N
`N
`Y
`Y
`N
`Y
`Y
`N
`N
`N
`N
`Y
`Y
`Y
`Y
`N
`Y
`Y
`N
`N
`
`N
`N
`N
`N
`N
`Y
`Y
`N
`N
`N
`N
`N
`N
`N
`N
`Y
`N
`N
`N
`N
`N
`N
`N
`
`N
`Y
`Y
`
`Y
`
`Y
`
`Y
`N
`N
`N
`Y
`
`Y
`Y
`Y
`
`Degrades
`Product?
`
`N
`
`N
`
`Y
`
`N
`
`Phosphate
`Salt
`Soluble?
`Y
`
`N
`
`N
`
`N
`N
`N
`N
`Gels
`
`Y
`N
`N
`
`Sparingly
`N
`
`Most notably, DMSO and methylene chloride formed solvates upon crystallization, a result not anticipated
`by original screening experiments run by Physical Measurements. Two new non-solvating solvents have
`been found: methyl benzoate and isoamyl alcohol. There are concerns about methyl benzoate forming
`amorphous material, and this will be confirmed with ssNMR. Isoamyl alcohol was ordered and the
`reactive crystallization will be run upon arrival of the solvent.
`
`Experiments continued with the crystallization of phosphate salt directly from water using a saturated salt
`solution as an antisolvent. Last week, this experiment was performed using a NaCl solution as the
`antisolvent. The wetcake was found to have an XRD pattern unlike any polymorph seen thus far. It was
`thought that this was due to the precipitation of the 715-HCl salt as opposed to the phosphate salt. AR is
`still confirming the identity of the salt that was precipitated in this experiment.
`
`Knowing that the 715-HCl salt may have been formed in the previous reaction, I ran the experiment again
`except this time using NaH2PO4 as the salt solution. XRD patterns of the resulting wetcake revealed a
`nearly identical pattern as the previous brine solution experiment. AR is working to confirm the identity of
`
`Merck Exhibit 2126, Page 8
`Mylan Pharmaceuticals Inc. v. Merck Sharp & Dohme Corp.
`IPR2020-00040
`
`

`

`the salt formed in this experiment through the use of titration, as well as Phosphorous NMR with Physical
`Measurements. Solubility of this salt in water will be run tomorrow to ensure it is still around 100 mg/g.
`
`Experiments for next week:
`Setup and run initial breakthrough experiments for Rh in the FLASH12+ unit.
`Screen isoamyl alcohol as a possible solvent for pure crystallization.
`Resolve issues surrounding the pure crystallization from water.
`Continue to screen for additional non-solvating solvents.
`
`Merck Exhibit 2126, Page 9
`Mylan Pharmaceuticals Inc. v. Merck Sharp & Dohme Corp.
`IPR2020-00040
`
`

`

`Memorandum
`To: Mahmoud Kaba
`From: Stephen Cypes
`Date: February 6, 2003
`Re: Weekly Progress Report, 02/10/03 02/14/03
`
`L-224,715
`Rhodium Removal from Hydrogenation Stream
`
`The use of the Biotage Cartridge System to remove Rh using a fixed bed adsorber is currently being
`studied. Pre-packed cartridges of Ecosorb-C941 were received from Biotage for use in the FLASH 12+M
`unit. Also, a manual fractionation unit was purchased to aid in collecting fractions for studies of Rh
`breakthrough. The experimental setup of the FLASH 12+M unit is shown in the diagram below.
`
`Small Parts, Inc.
`Male-to-Male
`Luer Lock Adaptor
`
`Syringe Pump
`Hydrog Crude
`Stream in 100 cc
`glass syringe
`
`FLASH 12+M
`Tubing Connected via
`Swagelok Fittings
`
`Fraction Collector
`
`The following procedure was performed to test the operation and Rh removal ability of the unit:
`Pre-wash C941 bed with 40 mL MeOH at 2 mL/min
`Pump 200 mL crude hydrog stream through bed at 2 mL/min. Collect fractions every 10 mL.
`Post-wash with 40 mL MeOH at 2 mL/min. Continue to collect fractions every 10 mL.
`
`A flowrate of 4 mL/min was attempted with this setup, however the syringe pump stalled (implying that the
`pressure drop over the bed was on the order of 60 psi, the pressure the syringe pump can supply for a 100
`cc syringe). Therefore, 2 mL/min is near the maximum flowrate obtainable for this experimental setup.
`
`Each of the 24 fractions is currently being screened for Rh and L-224715 freebase concentration. Rh
`sample preparations (i.e. dilution in concentrated nitric acid) is being done in our lab to facilitate less prep
`time in the heavy-metals analysis lab due to the large numbers of samples generated.
`
`Preliminary results show 100% of freebase eluting from the adsorbent bed with Rh concentrations still
`pending.
`
`
`L-224,715
`CSA Salt Break
`
`CSA salt most recently produced by Steve Rodgers was taken forward through the salt break and
`subsequently to the phosphate salt reactive crystallization step. The CSA salt used had a Rh concentration
`of 20 ppm. This salt was taken through the salt break step using procedures discussed in previous weekly
`reports. Samples of each cut from the salt break were submitted to AR for freebase, Rh and Na content
`(results still pending). The final THF cut was solvent-switched to denatured EtOH, and the free base
`simultaneously added with phosphoric acid in EtOH to a heel over 3 hours in a reactive crystallization. Rh
`level in the final phosphate salt was 4 ppm. Na content in the pure phosphate salt was 0.1 weight%.
`
`Merck Exhibit 2126, Page 10
`Mylan Pharmaceuticals Inc. v. Merck Sharp & Dohme Corp.
`IPR2020-00040
`
`

`

`L-224,715
`Pure Crystallization
`
`After further experimentation this week, it appears that isoamyl alcohol (ISAA) may be a good solvent to
`either promote the transition of phosphate salt polymorph from Form II to Form I, or be a good solvent for
`a reactive crystallization process.
`
`The following table summarizes the experiments that have led to the conclusions concerning isoamyl
`alcohol:
`
`Date
`
`Person Department
`
`Experiment
`
`Results
`
`Week of 3-FEB-2003 R. Ferlita
`
`Week of 3-FEB-2003 S. Cypes
`
`Phys. Meas. Slurry pure Form I in ISAA at
`room temperature overnight
`Direct reactive crystallization
`of phosphate salt in ISAA
`
`CERD
`
`XRD show s mainly Form I w ith some
`Form III
`XRD show s mainly Form I w ith some
`Form III
`XRD show s w etcake w as all Form II to
`start (as expected). After 4 hours,
`w etcake w as still mainly Form II w ith a
`little Form I. After 24 hours mainly Form
`I w ith a little Form III.
` The last experiment run shows that the transition from Form II to Form I may be faster in ISAA as
`compared to simply heating a drycake to promote the transition. An experiment is currently being
`performed (in conjunction with CERD and Phys. Meas.) to determine the kinetics of the transition in ISAA
`at elevated temperature. Results will be available early next week.
`
`Week of 10-FEB-2003 S. Cypes
`
`CERD
`
`Re-slurry a w etcake from
`EtOH crystallization in ISAA
`at room temperature
`
`Experiments for next week:
`Continue to run experiments for Rh removal in the FLASH12+ unit.
`Determine kinetics of polymorph transition in isoamyl alcohol at different temperatures.
`Continue to look at isoamyl alcohol as a possible solvent for the reactive crystallization of
`phosphate salt in the pure step.
`
`Merck Exhibit 2126, Page 11
`Mylan Pharmaceuticals Inc. v. Merck Sharp & Dohme Corp.
`IPR2020-00040
`
`

`

`Memorandum
`To: Mahmoud Kaba
`From: Stephen Cypes
`Date: February 20, 2003
`Re: Weekly Progress Report, 02/17/03 02/21/03
`
`L-224,715
`Rhodium Removal from Hydrogenation Stream
`
`The Biotage Cartridge System (a fixed bed adsorber) is continuously being studied for Rh removal from
`the crude Hydrogenation stream. The following procedure was performed to initially test the operation of
`the unit and produce breakthrough curves for L-224715 Freebase and Rh:
`Pre-wash Ecosorb C941 bed with 40 mL MeOH at 2 mL/min
`Pump 200 mL crude hydrog stream (0.3 mol% initial Rh charge) through bed at 2 mL/min.
`Collect fractions every 10 mL.
`Post-wash with 40 mL MeOH at 2 mL/min. Continue to collect fractions every 10 mL.
`
`Freebase concentration was determined in each of the fractions via HPLC. Also, each of the fractions was
`prepared for Rh analysis. The following breakthrough curves were produced based on the assay results.
`
`Breakthrough Curves for Freebase and Rh from Hydrog Crude
`Stream in Packed Column of Ecosorb C-941
`
`Displacement of M eOH
`Pre-Wash from Packed
`Bed by Hydrog Crude
`Stream
`
`Displacement of Hydrog
`Crude Stream by M eOH
`Post-Wash
`
`50
`
`150
`100
`Eluent Volume (mL)
`
`250
`200
`L-224715 Freebase
`Rhodium
`
`1.2
`
`1
`
`0.8
`
`0.6
`
`0.4
`
`0.2
`
`Relative Concentration (Cout/Cin)
`
`0
`
`0
`
`There are several important features to note about these breakthrough curves:
`
`Merck Exhibit 2126, Page 12
`Mylan Pharmaceuticals Inc. v. Merck Sharp & Dohme Corp.
`IPR2020-00040
`
`

`

`The freebase eludes from the bed very quickly. This is a positive result since it shows our
`product has low affinity for the C941 resin.
`The Rh curve shows a gradual increase without breakthrough. This reveals that there are non-
`idealities in the Rh adsoprtion (perhaps slow kinetics for the adsorption?). Once more
`experiments are run at several flowrates, the non-idealities can be identified as either kinetic,
`external, or internal mass transfer related.
`The freebase concentration gradually decreases as more Rh is adsorbed onto the C941. I
`currently believe that this decrease is due to an interaction between the freebase and Rh
`(Freebase chelates with Rh?). Shane Krska is the Reaction Engineering Laboratory has seen
`product inhibition in the hydrog reaction, and this may lend some insight as to the mechanism
`of this inhibition.
`~90% of the Rh was removed from a batch of hydrog crude equal to a 20 weight% loading of
`C-941 (current process uses 25 weight% C-941 in a batch slurry to remove 78% of the Rh).
`
`Without actually seeing Rh breakthrough in the first experiment, it is still possible to predict what volume
`of hydrog crude stream can be fed to the adsorber before Rh breakthrough will occur. Beginning with a
`mass balance on a differential element of the fixed bed:
`
`Superficial Velocity, u0
`Rh Concentration in liquid, CRh
`
`Ab = Cross-Section Area of Bed
`Void Fraction in Bed
`q = Amount of Rh on C-941 (Mass/Vol)
`
`z
`
`z+z
`
`IN = OUT + ACCUMULATION
`dCzA
`Rh
`dt
`
`zRh
`
`z
`
`b
`
`b
`
`CAu
`0
`b
`
`zRh
`
`CAu
`0
`b
`
`)
`
`b
`
`dqzA
`dt
`
`b
`
` (1)
`
`1(
`
`0
`
` (2)
`
`Rearranging and taking the limit as z 0:
`
`dC
`Rh
`dt
`
`dC
`Rh
`dz
`
`u
`
`0
`
`)
`
`b
`
`dq
`dC
`
`Rh
`
`1(
`
`b
`
`,
`This partial differential can be solved using the method of characteristics by introducing a new variable,
`that represents a coordinate of constant liquid Rh concentration in the bed (i.e. the point in the bed in which
`the Rh has saturated the C-941 and is therefore the breakthrough front).
`
` (3)
`
`dd
`
`z
`
`dC
`Rh
`dz
`
`dC
`d
`
`Rh
`
`0
`
`dC
`Rh
`dt
`
`dt
`d
`
`Comparing Equations 2 and 3, it can be seen that:
`
` (4)
`
`u
`
`0
`
`)
`
`b
`
`dq
`dC
`
`Rh
`
`1(
`
`b
`
`dd
`
`z
`
`Equation 4 represents the velocity of the breakthrough front for Rh, and is a function only of the superficial
`velocity of the liquid, the void fraction in the bed, and the slope of the adsorption isotherm. The slope of
`
`Merck Exhibit 2126, Page 13
`Mylan Pharmaceuticals Inc. v. Merck Sharp & Dohme Corp.
`IPR2020-00040
`
`

`

`the adsorption isotherm was determined by slurring different quantities of C-941 in the hydrog crude
`stream at room temperature and obtaining equilibrium Rh concentrations. The following plot shows this
`isotherm.
`
`Adsorption Isotherm for Rh on Ecosorb C941
`Based on data from Chris Welch, 03-JAN-2003
`
`y = 64.2x + 0.8
`
`01234567
`
`Adsorbant Rh Concentration
`
`(mg/mL)
`
`0
`
`0.08
`0.06
`0.04
`0.02
`Liquid Rh Concentration (mg/mL)
`
`0.1
`
`Using Equation 4 and the isotherm above, it can be predicted that breakthrough will occur after ~300 mL of
`the hydrog crude batch has been eluded, in good agreement with preliminary adsorption experiments.
`
`Preliminary experiments have also began on pressure drop calculations over the fixed bed. This is
`especially important for this system since the hydrog crude stream is quite viscous. The standard pump
`that is shipped by Biotage with their FLASH-400 production-scale unit has a max discharge pressure of
`100 psig. Therefore, it may be necessary to purchase a larger pump that can produce a higher discharge
`pressure. The following graph shows experimental pressure drop values for the FLASH 12+M unit as well
`as predicted values based on the Ergun Equation.
`
`Pressure Drop over Fixed Bed Adsorber for pure Methanol
`
`Experimental Values
`
`Predicted Pressure Drop:
`Ergun Equation
`
`30
`
`25
`
`20
`
`15
`
`10
`
`05
`
`Pressure Drop over Bed (PSI)
`
`0.0
`
`0.5
`
`1.0
`
`1.5
`Flow Rate (mL/min)
`
`2.0
`
`2.5
`
`3.0
`
`Merck Exhibit 2126, Page 14
`Mylan Pharmaceuticals Inc. v. Merck Sharp & Dohme Corp.
`IPR2020-00040
`
`

`

`L-224,715
`Pure Crystallization
`
`After further studies this week, it still appears as if isoamyl alcohol (ISAA) may be a good solvent to either
`promote the transition of phosphate salt polymorph from Form II to Form I, or be a good solvent for a
`reactive crystallization process.
`
`The following table summarizes the experiments that have led to the conclusions concerning isoamyl
`alcohol:
`
`Date
`
`Person Department
`
`Experiment
`
`Results
`
`Week of 3-FEB-2003 R. Ferlita
`
`Week of 3-FEB-2003 S. Cypes
`
`Phys. Meas. Slurry pure Form I in ISAA at
`room temperature overnight
`Direct reactive crystallization
`of phosphate salt in ISAA at
`20C
`
`CERD
`
`Week of 10-FEB-2003 S. Cypes
`
`CERD
`
`Week of 17-FEB-2003 S. Cypes
`
`CERD
`
`Re-slurry a w etcake from
`EtOH crystallization in ISAA
`at room temperature
`
`Re-slurry a w etcake from
`EtOH crystallization in ISAA
`at 70C
`
`XRD show s mainly Form I w ith some
`Form III
`XRD show s mainly Form I w ith some
`Form III
`XRD show s w etcake w as all Form II to
`start (as expected). After 4 hours,
`w etcake w as still mainly Form II w ith a
`little Form I. After 24 hours mainly Form
`I w ith a little Form III.
`XRD revealed that turnover to 100%
`Form I occurs w ithin 1 hr. Remained
`100% Form I after 2 hrs. Upon cooling
`to 20C, turned back to 100% Form II do
`to slightly higher EtOH concentrations
`compared to previous experiments.
`Max allow able EtOH ~2 w eight%
`Mean PS original EtOH Wetcake = 40.8
`microns, Mean PS 1 hr ISAA (70C)
`Wetcake = 40.7 microns, PSD Identical
`betw een EtOH and ISAA, Habits
`Identical (Plates), No Amorphous
`Material in ISAA Wetcake (ssNMR)
`It has been shown that re-suspending a wetcake from an EtOH crystallization in ISAA at 70C for 1 hr will
`result in complete polymorph turnover to Form I. The logistics of either solvent switching from EtOH to
`ISAA in a slurry or performing a direct reactive crystallization in ISAA will both be studied.
`
`Week of 17-FEB-2003 S. Cypes
`
`CERD
`
`PSD, Habit and Crystallinity
`of crystals in ISAA
`
`Experiments for next week:
`Obtain a breakthrough curve for Rh removal in the FLASH-12 unit by eluding >300 mL
`hydrog crude stream in single experiment.
`Determine the cause of non-idealities in Rh breakthrough curve.
`Optimize the pure step reactive crystallization process in ISAA
`
`Merck Exhibit 2126, Page 15
`Mylan Pharmaceuticals Inc. v. Merck Sharp & Dohme Corp.
`IPR2020-00040
`
`

`

`Memorandum
`To: Mahmoud Kaba
`From: Stephen Cypes
`Date: February 27, 2003
`Re: Weekly Progress Report, 02/24/03 02/28/03
`
`L-224,715
`Rhodium Removal from Hydrogenation Stream
`
`It was shown last week that the Biotage cartridge system could successfully be used to selectively remove
`Rhodium from the Hydrog crude stream more efficiently compared to a batch slurry of Ecosorb C-941.
`One of the main technological obstacles with using C-941 in the packed-bed cartridge system is the large
`pressure drop that results, even at moderate flowrates. For flowrates used in the pilot plant in the largest
`cartridge system, the pressure drop would be on the order of 200 psi, higher than the pressure rating of the
`cartridge system itself (100psi).
`
`The Ergun Equation can be used to predict pressure drop in a packed-bed, and was successfully used to
`predict the pressure drop in the C-941 cartridge (presented last week). The Ergun Equation states that the
`pressure drop is proportional to 1/Dp
` , where Dp is the mean particle diameter (about 3 microns for C-941).
`2
`Therefore, the large pressure drop (about 30 psi at 2.2 cm/min for a 15 cm long bed) is mainly attributed to
`the small particle size of the C-941.
`
`To alleviate the issue with pressure drop while maintaining high levels of Rhodium removal, another resin
`from Biotage (SN-Bio), which has identical acid-activated carbon as C-941, was screened for use in the
`cartridge. SN-Bio is an attractive alternative due to the fact that the material is classified (i.e. the PSD is
`narrow and the mean size is ~35 microns, significantly larger than C-941). The pressure drop across a
`packed cartridge of SN-Bio is compared to C-941 for methanol below. For any given pressure drop, 5x the
`flowrate results for the SN-Bio packed-cartridge. As a reference, a flowrate of 8mL/min in the lab-scale
`cartridge corresponds to a flowrate of 35L/min in the pilot-scale cartridge system.
`
`Pressure Drop over Fixed Bed Adsorber
`
`C941 - Experimental
`
`SN-Bio - Experimental
`
`0.0
`
`2.0
`
`8.0
`6.0
`4.0
`Flow Rate (mL/min)
`
`10.0
`
`12.0
`
`30
`
`25
`
`20
`
`15
`
`10
`
`05
`
`Pressure Drop over Bed (PSI)
`
`Merck Exhibit 2126, Page 16
`Mylan Pharmaceuticals Inc. v. Merck Sharp & Dohme Corp.
`IPR2020-00040
`
`

`

`Since SN-Bio appears to result in a solution to the pressure drop issue, the Rhodium removal
`characteristics of this resin needed to be evaluated. By producing a slurry of different amounts of SN-Bio
`in a set volume of hydrog crude stream, the adsorption isotherm for this resin can be compared to C-941.
`The results are shown in the graph below. SN-Bio actually does a slightly better job of removing Rhodium
`from the hydrog stream as compared to C-941.
`
`Adsorption Isotherm for Rh on Ecosorb C941
`Based on data from Chris Welch, 03-JAN-2003
`
`Ecosorb C-941
`Biotage SN-Bio
`0.06
`
`0.04
`0.02
`Liquid Rh Concentration (mg/mL)
`
`0
`
`5
`4.5
`4
`3.5
`3
`2.5
`2
`1.5
`1
`0.5
`0
`
`Adsorbant Rh Concentration
`
`(mg/mL)
`
`
`
`L-224,715
`Pure Crystallization
`
`It has been shown in a number of experiments that isoamyl alcohol (ISAA) can be used to facilitate the
`formation of either pure Form I phosphate salt or a mixture of Forms I and III. So far, three methods of
`using ISAA to facilitate the polymorph turnover have been investigated. In the order in which experiments
`were performed, they are:
`Re-slurry a wetcake from an ethanol crystallization in ISAA, allow turnover to occur, and re-
`filter;
`Perform the current crystallization procedure in ethanol, solvent switch the slurry to ISAA,
`allow turnover to occur, and filter;
`Perform a reactive crystallization directly in ISAA to produce the desired polymorph.
`
`Details on the status of each of these methods follows:
`
`Re-slurry a Wetcake from Ethanol
`Experiments revealed that the maximum allowable ethanol concentration in the ISAA slurry must be <2
`weight% to achieve a turnover from Form II to Form I and/or III. Therefore, the wetcake from the ethanol
`concentration must be partially dried to an LOD of ~10 weight%, and this wetcake added to 10 L/kg ISAA.
`The flowsheet for this procedure looks like one of two methods:
`Re-slurry in 10 L/kg
`ISAA at 20C, 24
`hrs, Filter
`
`Method A (Deliver Mix of Form I and III)
`Crystallization
`Filter Wetcake, Form II
`in Ethanol
`(LOD<10%)
`
`Wetcake
`Forms I and III
`
`Merck Exhibit 2126, Page 17
`Mylan Pharmaceuticals Inc. v. Merck Sharp & Dohme Corp.
`IPR2020-00040
`
`

`

`Method B (Deliver 100% Form I)
`
`Crystallization
`in Ethanol
`
`Filter Wetcake, Form II
`(LOD<10%)
`
`Re-slurry in 10 L/kg
`ISAA at 70C, 1 hr,
`Filter
`
`Wetcake
`100% Form I
`
`Solvent Switch a Slurry from Ethanol to ISAA
`
`A slurry of large phosphate salt crystals (mean = 50 microns) was suspended in denatured ethanol. A
`vacuum distillation was performed at 50oC to solvent switch to ISAA and promote a polymorph turnover to
`100% Form I. The table below summarizes the results from this experiment.
`
`ST EP
`
`T emp
`(C )
`
`P ressure
`(to rr)
`
`EtOH
`M ass
`(g)
`
`ISA A
`M ass
`(g)
`
`P ho sphate Salt
`C o ncentratio n
`(mg/ g)
`
`M o le%
`ISA A
`
`Weight%
`ISA A
`
`C rystal
`F o rm
`
`M ean
`P article
`Size
`(micro ns)
`
`C rystal
`H abit
`
`Add 240 g Denatured EtOH to 30 g Phosphate Salt (M ix Form I and Form III). Agitate 2 hrs. to solvate.
`50.00
`Start
`50
`225
`240
`0
`111
`0%
`0%
`Form II
`Finish
`50
`225
`110
`0
`214
`0%
`0%
`Form II Undetermined
`Add 110 g Isoamyl Alcohol
`Start
`50
`155
`Finish
`50
`120
`Add 60 g Isoamyl Alcohol
`Start
`50
`115
`Finish
`50
`45
`Add 50 g Isoamyl Alcohol
`Start
`50
`40
`Finish
`50
`22
`
`34%
`45%
`
`56%
`82%
`
`85%
`96%
`
`50%
`61%
`
`71%
`89%
`
`92%
`98%
`
`Form II Undetermined
`Form II Undetermined
`
`Form II Undetermined
`Form II Undetermined
`
`Form II Undetermined
`Form I
`45.53
`
`1
`
`2
`
`3
`
`4
`
`110
`70
`
`70
`20
`
`20
`4.5
`
`110
`110
`
`170
`170
`
`220
`220
`
`120
`143
`
`111
`136
`
`111
`118
`
`P