`
`Syringe Siliconization Process Investigation and Optimization
`
`EDWIN CHAN, AARON HUBBARD, SAMIR SANE, and YUH-FUN MAA*
`
`Pharmaceutical Processing and Technology Development, Genentech, a member of Roche Group ©PDA, Inc. 2012
`
`ABSTRACT: The interior barrel of the prefilled syringe is often lubricated/siliconized by the syringe supplier or at the
`syringe filling site. Syringe siliconization is a complex process demanding automation with a high degree of precision;
`this information is often deemed “know-how” and is rarely published. The purpose of this study is to give a detailed
`account of developing and optimizing a bench-top siliconization unit with nozzle diving capabilities. This unit
`comprises a liquid dispense pump unit and a nozzle integrated with a Robo-cylinder linear actuator. The amount of
`coated silicone was determined by weighing the syringe before and after siliconization, and silicone distribution was
`visually inspected by glass powder coating or characterized by glide force testing. Nozzle spray range, nozzle
`retraction speed, silicone-coated amount, and air-to-nozzle pressure were found to be the key parameters affecting
`silicone distribution uniformity. Distribution uniformity is particularly sensitive to low-target silicone amount where
`the lack of silicone coating on the barrel near the needle side often caused the syringes to fail the glide force test or
`stall when using an autoinjector. In this bench-top unit we identified optimum coating conditions for a low silicone
`dose, which were also applicable to a pilot-scale siliconization system. The pilot unit outperformed the bench-top unit
`in a tighter control (standard deviation) in coated silicone amount due to the elimination of tubing flex. Tubing flex caused
`random nozzle mis-sprays and was prominent in the bench-top unit, while the inherent design of the pilot system
`substantially limited tubing flux. In summary, this bench-top coating unit demonstrated successful siliconization of the 1
`mL long syringe with ⬃0.2 mg of silicone oil using a spraying cycle also applicable to larger-scale siliconization.
`
`KEYWORDS: Prefilled syringe, Siliconization, Dive-in nozzle spray, Silicone amount, Silicone distribution, Glide
`force, Glass powder testing
`
`LAY ABSTRACT: Syringe siliconization can be considered a well-established manufacturing process and has been
`implemented by numerous syringe providers. However, its technical details and associated critical process parameters
`are rarely published. The purpose of this study is three-fold: (1) to reveal design details of a bench-top siliconization
`unit, (2) to identify critical process parameters and determine their optimum range to provide consistent and even
`silicone coating, and (3) to demonstrate the applicability of the optimum process condition derived from the bench-top
`unit to a pilot siliconization unit. The outcomes of this study will benefit scientists and engineers developing pre-filled
`syringe products by helping them to better understanding silicone spray coating principles and their relationship to
`siliconization processes in a large-scale manufacturing setting.
`
`Introduction
`
`Pre-filled syringes are now the primary container of
`choice for most parenteral drug delivery systems,
`mainly because they are safe and user-friendly (1).
`Manufacturing a pre-filled syringe product is a com-
`plex process
`(2– 4),
`including liquid formulation
`
`* Corresponding author: Genentech, 1 DNA Way,
`South San Francisco, CA 94080. TEL: 650-225-3499.
`FAX: 650-742-1504. E-mail: maay@gene.com.
`doi: 10.5731/pdajpst.2012.00856
`
`preparation (thawing, compounding, sterile-filtration,
`etc.), component assembly (syringe, stopper, needle,
`and needle shield), syringe fill, stopper placement,
`labelling, packaging, and so on. Some device com-
`ponents are lubricated (or siliconized, as silicone
`oil
`is the industry standard and a Food and Drug
`Administration–approved lubricant), particularly the
`interior barrel of the syringe to ensure ease of syringe
`performance and consistency of the injection force
`(3–5). Siliconization can be performed by the syringe
`manufacturer in the ready-to-fill (or nest or tub) format
`or in the syringe fill facility prior to filling in the bulk
`configuration (3).
`
`136
`
`PDA Journal of Pharmaceutical Science and Technology
`
`Novartis Exhibit 2006.001
`Regeneron v. Novartis, IPR2020-01317
`
`
`
`There are three types of silicone fluid, or polydimeth-
`ylsiloxane (PDMS), available for syringe/cartridge lu-
`brication: non-reactive silicone oil (e.g., Dow Corning
`[DC] 360 Medical Fluid available in five viscosities),
`non-reactive silicone emulsion (e.g., DC 365 35%
`Dimethicone NF Emulsion), and reactive (curable)
`silicone fluid (e.g., DC MDX5-4159 Medical Grade
`Dispersion). After silicone application, a high-temper-
`ature “baking” process is required for silicone emul-
`sion and reactive silicone fluid for depyrogenation, or
`curing. The non-reactive silicone oil doesn’t require
`the post-application baking process as DC 360 are
`tested for bacterial endotoxins and certified to meet
`National Formulary/European Pharmacoepeia specifi-
`cations (6). Regardless of the silicone fluid type, sili-
`cone
`applications
`are
`commonly performed by
`wipe-on or spray coating (5), where the fluid is atom-
`ized into a mist via a nozzle and deposited on the
`coated surface. Thus, siliconization relies on three
`mechanisms: accurate dosing, controlled atomization,
`and precise nozzle movement. Two-fluid atomization
`of viscous silicone fluid via high-pressure air is cur-
`rently the best option for producing fine droplets. A
`precision pump is used to accurately deliver a minute
`amount of fluid for atomization. The nozzle needs
`to travel at a controlled speed inside the syringe
`barrel to coat its inner surface evenly. These three
`mechanisms need to coordinate perfectly to provide a
`fixed dose of fluid distributed uniformly across the
`internal surface of the syringe. Despite the fact that
`automated, high-speed manufacturing processes have
`been established for many years to produce siliconized
`syringes, our literature search failed to find publica-
`tions detailing siliconization process development. In-
`stead, most publications focused on formulation and
`stability considerations as the result of silicone-protein
`interactions (4, 7–9).
`
`The objective of this study is to develop and optimize
`a bench-top siliconization system with a focus of
`assessing the effect of critical parameters on coating
`amount and distribution uniformity. This bench-top
`unit was also compared with a semi-automated, pilot-
`scale siliconization system to understand how system
`design affects siliconization performance.
`
`Materials and Methods
`
`All experiments in this study employed 1.0 mL long
`27G 1⁄2⬙ staked needle syringes to be coated with
`silicone oil (DC 360 Medical Fluid, 1000 cSt) using a
`bench-top siliconization system (Figure 1) or a semi-
`
`automated pilot scale unit. Equipment and materials
`used in this study are tabulated in Table I.
`
`Bench-Top Siliconization System—Spraying Unit
`
`The siliconization bench-top unit was assembled and
`tested by Volo Technologies (Roseville, CA). This
`setup (Figure 1a) integrates a spraying unit with a
`robot system controlling nozzle movement. The spray-
`ing unit (Figure 1b) consists of an IVEK Sonicair
`nozzle, a piston pump, a pump linear actuator, a heater
`module, and a Digisonic controller, while the robot
`system is composed of a Robo-cylinder linear actuator
`and its controller. A high-precision, ceramic positive
`displacement pump controlled by a linear actuator for
`dosing delivers silicone oil to the Sonicair nozzle set.
`The nozzle set includes a nozzle head and a nozzle
`(enlarged section of Figure 1b). The head contains
`ports to receive silicone oil, air, and a heat cartridge.
`The nozzle itself, based on a two-fluid atomization
`mechanism, features two concentric tubes with sili-
`cone oil flowing through the inner tube and high-
`pressure air flowing between the two tubes. The oil
`and the air meet at the tip of the nozzle for mixing, and
`the mixture is sprayed via a 0.5 mm orifice into a
`cone-shaped mist (Figure 1c). Nozzle air is supplied
`via in-house compressed air and controlled by a pres-
`sure gauge. The same air source controlled by a sep-
`arate pressure gauge is fed to the silicone oil reservoir
`(5 psi was used throughout this study), which precedes
`the piston pump module. The nozzle head can be
`equipped with a heating element as an optional fea-
`ture. A heat cartridge (Firerod威, Watlow, St. Louis,
`MO), when inserted into the port on the nozzle head,
`can heat the nozzle head to a pre-set temperature (in
`the range of 25 to 125 °C). The temperature of the
`nozzle will rise due to heat transferred from the nozzle
`head. Nozzle air can be simultaneously heated by a
`heating tube inserted between the pressure gauge and
`the nozzle head and controlled by a separate heat
`controller.
`
`The pump module is a piston/cylinder arrangement
`providing positive displacement. The actuator module
`selectively rotates and reciprocates the ceramic piston
`via a coupling at one end, and the piston incorporates
`a slit that provides a valving function at the other end.
`Initially the piston flat aligns with the silicone oil
`intake port and then retracts to fill the cylinder with
`liquid. The piston rotates 180° with the flat facing the
`discharge port and then pushes forward to force the
`
`Vol. 66, No. 2, March–April 2012
`
`137
`
`Novartis Exhibit 2006.002
`Regeneron v. Novartis, IPR2020-01317
`
`
`
`TABLE I
`Equipment and Materials Used in the Study
`
`Equipment
`
`Balance
`
`Two fluid nozzle (5 mm OD and 90 mm length)
`
`Ceramic positive displacement piston pump and
`controller
`
`Pump controller module (Micro linear actuator)
`
`Robo-cylinder linear actuator
`PLC/HMI controller
`Nozzle heater
`Nozzle heater controller
`
`Instron force measuring device
`Vent wire stoppering tool
`
`Semi-automatic syringe washer
`
`Materials
`1 mL-Long syringe with 27G ⫻ 1⁄2⬙ staked needle
`
`Tefzel tubing (1/8⬙ OD, 1/16⬙ ID)
`Plunger stopper
`
`Polystyrene plunger rod
`
`Silicone fluid
`
`Supplier/Model
`
`Mettler Toledo AG 204 (Columbus,
`OH)/AG135
`IVEK Corp. (North Springfield,
`Vermont)/Sonicair
`IVEK Corp. (North Springfield,
`Vermont)/Micro linear pump
`module & Linear actuator module
`IVEK Corp. (North Springfield,
`Vermont)/Digispense 3020
`IAI America (Torrence, CA)
`Omron American (Schaumburg, IL)
`Watlow (St. Louis, MO)/Firerod威
`Red Lion Controls (York, PA)/Model
`T48
`Instron (Grove City, PA)/Model 5542
`Becton Dickinson (Franklin Lakes,
`NJ)/Custom tool
`Bausch Advanced Technology
`(Clinton, CT) Type 303 Washer
`Supplier/Model
`Becton Dickinson (Swedesboro, NJ)/
`Hypak Type 1 glass syringes (Lot #
`082111)
`Upchurch Scientic (Oak Harbor, WA)
`West Pharmaceuticals (Lionville, PA)
`W4023/FLT (Lot # 0000122330)
`Becton Dickinson (Bridgeport, NJ)/
`Plunger rod for 1mL Long (Lot #
`44386)
`Dow Corning (Midland, MI)/Dow 360
`Medical Fluid, 1,000 cSt (Lot #
`0005938393)
`
`required amount of oil through the discharge port.
`The piston rotates 180° again back to the initial posi-
`tion to complete the dispense cycle. The height that
`the piston travels determines the volume of the oil to
`be sprayed.
`
`Robotic Movement Control
`
`The Robo-cylinder linear actuator provides the diving
`action for the nozzle set (Figure 1a & b). The whole
`nozzle set moves up and down per the commands from
`the Volo Controller. This controller combines the
`function of a programmable logic controller (PLC) and
`
`a human machine interface (HMI). The siliconization
`bench-top unit can be run manually or programmed to
`run an automatic siliconization cycle where the nozzle
`dives upward into the syringe and sprays silicone
`while retracting. The Volo Controller is interfaced
`with the Digispense Controller to send signals to begin
`and end pumping the oil. To enable the function of
`controlling nozzle movement such as the dive-in po-
`sition, the spray position, the spray rate, and the ac-
`celeration/deceleration, various setpoints need to be
`entered on Volo HMI and Digispense. The procedure
`of preparing a worksheet for setpoint entry is detailed
`in the Appendix.
`
`138
`
`PDA Journal of Pharmaceutical Science and Technology
`
`Novartis Exhibit 2006.003
`Regeneron v. Novartis, IPR2020-01317
`
`
`
`Figure 1
`
`(a) The entire bench-top siliconization system; (b) detailed representation of the pump module, the nozzle set,
`and syringe fixture and station; (c) detailed graphic representation of the nozzle design and the two-fluid
`atomization mechanism; and (d) syringe fixture station without the extension.
`
`Pilot-Scale Siliconization Station
`
`The semi-automated syringe washing/siliconization
`equipment was designed and assembled by Bausch
`
`Advanced Technology (Clinton, CT). This pilot-scale
`equipment is equipped with an Allen Bradley touch
`screen HMI that provides operators with full control of
`machine settings to wash, air-dry, and siliconize sy-
`
`Vol. 66, No. 2, March–April 2012
`
`139
`
`Novartis Exhibit 2006.004
`Regeneron v. Novartis, IPR2020-01317
`
`
`
`ringes. Syringes are fed manually at the in-feed station
`with needle pointing downward. The in-feed scroll
`indexes and transports the syringes into individual
`grippers. The gripper then rotates and inverts the sy-
`ringe 180°, with needle pointing upward. The syringe
`is then transported through 16 gripper stations that
`include three washing stations, five drying stations,
`one barrel siliconization station, and one needle sili-
`conization station (a feature not relevant to this study).
`The mechanism of siliconization is very similar to the
`bench-top silicone oil spraying unit described above.
`
`Both pilot-scale and bench-top units utilize the same
`IVEK Sonicair spraying nozzle combined with a high-
`precision, ceramic positive displacement pump and
`heating elements. Customized setpoints, such as sili-
`cone dosing amount, nozzle/air
`temperatures, and
`spray nozzle movement/positions can be controlled
`via the washer HMI. A dive-in siliconization motion is
`achieved by the use of a cam fixture, which works
`similarly to the linear actuator in the bench-top unit.
`However, the pilot-scale siliconization station features
`a synchronized movement of the pump and the nozzle,
`as both are physically attached to the cam fixture. This
`feature enhances siliconization performance and will
`be discussed in the Discussion Section.
`
`Silicone Amount
`
`The amount of silicone oil applied on each syringe was
`determined by using a high-precision digital balance
`(Mettler Toledo AG 204). Empty syringes were
`weighed before and after the siliconization process.
`The silicone oil quantities were measured by calculat-
`ing the differences between the two values.
`
`Fixed Nozzle Siliconization Testing
`
`Syringes (n ⫽ 5) were sprayed with 0.5 mg of silicone
`oil using an air-to-nozzle pressure of 15 psi at three
`fixed positions, 5 mm, 20 mm, and 35 mm past the
`flange. Each siliconized syringe was then coated with
`glass powder and visually inspected.
`
`Silicone Distribution (Glass Powder Method)
`
`⫽
`A Schott glass powder 8250 (Grain Size K1, d50
`30 m ⫾ 10 m, d99
`ⱕ 150 m) was used to visualize
`and assess silicone distribution uniformity in the inner
`barrel of a syringe. This specific glass powder was
`designed to only stick to glass surfaces coated with
`silicone oil. Approximately 150 mg of glass powder
`
`was poured into the open end of the syringe, which
`was subsequently covered with parafilm. The syringe
`was then tilted, tapped, and shaken manually for 30 s
`while in a horizontal position to distribute powder
`along the axis of the barrel. Subsequently, the used
`powder was discarded. With the open end pointing
`downward, the syringe was tapped gently against a flat
`surface for up to 25 times to remove any excess
`powder. Coated syringes were then inspected visually
`for homogeneous powder distribution. Any empty gap
`greater than 5 mm was defined as unacceptable.
`
`Silicone Distribution (Glide Force Method)
`
`High-speed force testing was performed as a quanti-
`tative assay to gauge silicone distribution uniformity
`within the inner barrel of the syringe. A plunger stop-
`per
`(W4023/FLT, West Pharmaceutical, Lionville,
`PA) was inserted into the empty, post-siliconized sy-
`ringe approximately 8 mm inward (measured from the
`back of the syringe flange to the back of a stopper)
`using a vent wire stoppering tool (Becton Dickinson,
`Franklin Lakes, NJ). Next, a polystyrene plunger rod
`was carefully threaded into the plunger stopper. A
`material testing system (model 5542, Instron, Grove
`City, PA) with a load cell was used to apply a steady
`compression rate, and the gliding force profile was
`then analyzed for silicone coating consistency and
`variation per an internal protocol.
`
`Results and Discussion
`
`Two coating experiments were initially performed to
`confirm the effect of fixed-position spray coating on
`distribution uniformity and coated silicone amount.
`The experimental set-up follows Figure 1d, where the
`syringe fixture extension was not used.
`
`Effect of Fixed-Position Spray on Coating Distribution
`Uniformity
`
`Figure 2a– c shows distribution uniformity results for
`the fixed-position spray coating at 5, 20, and 35 mm
`past the flange, respectively. Spraying near the flange
`(5 mm) resulted in silicone deposition mostly on the
`flange side of the barrel and leaving the barrel on the
`needle side uncoated. When the nozzle was deeper
`into the syringe (20 mm), the middle section of the
`barrel was coated. Further into the syringe at 35 mm
`past the flange, coating occurred primarily at the nee-
`dle side. This finding suggests that fixed-position
`spray yields uneven distribution and that the nozzle
`
`140
`
`PDA Journal of Pharmaceutical Science and Technology
`
`Novartis Exhibit 2006.005
`Regeneron v. Novartis, IPR2020-01317
`
`
`
`Figure 2
`
`Glass powder testing on syringes coated with 0.5 mg of silicone oil at three fixed positions, (a) 5 mm, (b) 20 mm,
`and (c) 35 mm past the flange, using 5 psi air-to-nozzle pressure.
`
`needs to spray and retract simultaneously. Spraying
`over the range of 5–35 mm past the flange along with
`nozzle movements may result in uniform coating over
`the entire barrel. This will be studied using the pilot-
`scale system and results will be discussed later.
`
`Effect of Spraying Nozzle Position, Nozzle Pressure,
`and Spray Rates on Coating Amount
`
`The design of this experiment involved the variation
`of three process parameters—three nozzle spraying
`positions (10, 27, and 44 mm past the flange), two
`air-to-nozzle pressures (5 and 15 psi), and three spray
`rates (0.420, 4.316, and 8.210 mg/s). The selection of
`these spray rates was intended to cover a wide nozzle
`movement range, that is, deep dive with slow retrac-
`tion and shallow dive with quick retraction. A full
`factorial experimental design set with these variables
`gives 18 sets of conditions, that is, a total of 180 sam-
`ples at n ⫽ 10. The target dose is 0.3 mg silicone for
`all conditions.
`
`The results for coated silicone amount (Figure 3)
`reveal an obvious problem: a significant number of
`under-coated samples (⬍0.1 mg), including no coating
`(zero silicone-coated). It appears that the siliconiza-
`tion bench-top unit occasionally mis-sprayed, or
`
`sprayed prematurely. As the nozzle dived in and out of
`the syringe, it pinched the Teflon tubing between the
`pump and the nozzle to flex. The flexing movement
`could squeeze a minute amount of silicone oil out of
`the nozzle and spray out of the cycle (not inside the
`syringe). When the tubing returned to a relaxed posi-
`tion, the silicone level retracted and left a void volume
`at the nozzle tip. Thus, the syringe in the next spraying
`cycle would be undercoated. The mis-spray problem
`confounded the data and allowed for few conclusions
`to be drawn. This problem occurred randomly and
`needed to be resolved or minimized before other ex-
`periments could be performed.
`
`Approaches to Resolve Mis-Spray/Premature Spray
`Issues due to Tubing Flex
`
`A straightforward approach is to shorten the distance
`that nozzle travels as well as the length of tubing
`between the pump and the nozzle. To test this concept,
`an extension (95 mm) was inserted at the top of the
`syringe fixture (Figure 1d) to minimize the distance
`that the nozzle travels (only inside the syringe), and
`the tubing length was reduced to only loosely hang in
`the air. This setup was used to spray 0.3 mg of silicone
`during nozzle retraction over the range of 30 to 10 mm
`past the flange. Two parameters were varied at three
`
`Vol. 66, No. 2, March–April 2012
`
`141
`
`Novartis Exhibit 2006.006
`Regeneron v. Novartis, IPR2020-01317
`
`
`
`Figure 3
`
`Result summary of coated silicone amount for syringes coated at three fixed positions (10, 27, and 44 mm past
`the flange), two air-to-nozzle air pressures (5 and 15 psi), and three spray rates (0.420, 4.316, and 8.210 L/s)
`with a target silicone dose of 0.3 mg.
`
`levels, air-to-nozzle pressure (5, 10, and 15 psi) and
`retraction speed (50, 100, and 150 mm/s). Thus, this
`study involved nine conditions (n ⫽ 10) with a total
`sample size of 90 syringes.
`
`The silicone amount results in Figure 4 show a great
`improvement in repeatability of the silicone dose, with
`
`Figure 4
`
`Result summary of coated silicone amount with a
`modified experimental setup using a syringe exten-
`sion (95 mm) and a shortest tubing length (⬃25 cm).
`Syringes were coated with a moving nozzle that dived
`in 40 mm past the flange and sprayed across the
`range of 30 –10 mm past the flange during retraction
`using three air-to-nozzle pressures (5, 10, and 15 psi)
`and retraction speeds (50, 100, and 150 mm/s) with a
`target silicone dose of 0.3 mg.
`
`only two samples coated with ⬍0.1 mg silicone, in-
`cluding one with zero coating (5 psi/100 mm/s). This
`suggests that tubing flex was much reduced but not
`completely eliminated. Overall,
`the average coated
`silicone weight is less than the target dose (0.3 mg) for
`most of the conditions, suggesting that some silicone
`droplets may be lost to depositing on the nozzle and/or
`being blown out of the syringe. The effect of the
`air-to-nozzle pressure and the retraction rate on sili-
`cone loss is not clear from this study. Conceptually,
`faster retraction may result in greater silicone loss
`(less coating) because the downward movement of the
`nozzle can offset the upward velocity of the spray so
`that some droplets may flow with air out of the syringe
`instead of depositing on the barrel.
`
`There are other approaches that may be able to com-
`pletely eliminate the tubing flex issue. One is to mo-
`bilize the nozzle and the pump (including the tubing in
`between) as one unit. Unfortunately, the linear actua-
`tor in the current setup was not capable of carrying the
`weight of the pump module. However, the pilot-scale
`siliconization unit had such a feature (Figure 8) and
`was tested as discussed later.
`
`Another approach is to keep the nozzle stationary but
`to move the syringe instead. The nozzle still sprays
`upwards but stays motionless during syringe fixture
`movement. With both nozzle and pump being motion-
`less, there should be no tubing flex. However, this
`approach is not within the scope of this investigation
`and will present a different configuration to the pilot
`unit to be evaluated later.
`
`142
`
`PDA Journal of Pharmaceutical Science and Technology
`
`Novartis Exhibit 2006.007
`Regeneron v. Novartis, IPR2020-01317
`
`
`
`Figure 5
`
`Glass powder testing and Instron glide force profile for syringes coated with 0.2 mg silicone and sprayed in
`range of 10 to 44 mm past the flange using 15 psi air-to-nozzle pressure at the retraction rate of (a) 200 mm/s
`or (b) 50 mm/s (for the three syringes on the left).
`
`Finding Key Process Parameters on Coating
`Uniformity
`
`As determined earlier, for silicone droplets to be
`evenly distributed across the whole barrel surface, the
`spraying nozzle has to dive in the syringe and spray
`during retraction. Proper siliconization at the needle
`side of the barrel is important. For spring-based auto-
`injection devices, the most severe challenge will be at
`end of the stroke, when the available spring energy
`will be at its lowest level (11). To reach this area, the
`nozzle has to dive in the syringe barrel deep enough or
`spray silicone oil droplets far enough to cover the
`needle side. The upward velocity of the droplets is
`dictated by the change in air-to-nozzle pressure but
`
`can be offset by nozzle retraction velocity and gravity.
`Therefore,
`to evaluate the effect of changing the
`spraying process on silicone distribution, four process
`parameters were varied: nozzle retraction rate (50 or
`200 mm/s), nozzle spray range (22–10 or 44 –10 mm
`past the flange), air pressure to the nozzle (5 or 15 psi),
`and target silicone quantity (0.2, 0.5, or 1.0 mg).
`Please note that nozzle retraction rate is important for
`production speed. A normal production-speed sili-
`conization spray should be under 1 s per cycle and
`50 mm/s is near that mark, depending on dive-in
`depth. Samples created for this test are listed in Table
`II with n ⫽ 10, where three samples were used for
`glass powder testing and three for Instron glide force
`testing. The glide force testing results are also sum-
`
`Vol. 66, No. 2, March–April 2012
`
`143
`
`Novartis Exhibit 2006.008
`Regeneron v. Novartis, IPR2020-01317
`
`
`
`TABLE II
`Study Design for Silicone Distribution Test (n ⴝ 10 Total Produced; n ⴝ 3 Used for Glass Powder Testing,
`and n ⴝ 3 Used for Instron Glide Force Testing)
`
`Air
`Pressure
`(psi)
`
`Speed
`(mm/s)
`
`Si
`Quantity
`(mg)
`
`Position
`(mm)
`
`# Passing
`on
`Instron
`Test
`
`15
`5
`15
`15
`5
`15
`15
`15
`15
`5
`5
`15
`15
`5
`15
`5
`5
`5
`5
`15
`5
`5
`5
`15
`
`50
`200
`200
`200
`50
`50
`200
`200
`50
`50
`200
`200
`50
`200
`50
`50
`50
`200
`50
`50
`200
`200
`50
`200
`
`0.5
`0.2
`0.5
`0.2
`0.2
`0.5
`1.0
`0.5
`0.2
`0.2
`1.0
`1.0
`0.2
`0.2
`1.0
`0.5
`1.0
`0.5
`1.0
`1.0
`0.5
`1.0
`0.5
`0.2
`
`22
`44
`44
`44
`22
`44
`22
`22
`22
`44
`44
`44
`44
`22
`22
`44
`22
`22
`44
`44
`44
`22
`22
`22
`
`3
`1
`1
`0
`1
`3
`3
`0
`1
`1
`3
`3
`3
`0
`3
`3
`3
`0
`3
`3
`1
`3
`3
`0
`
`Pattern
`⫹⫺2⫺
`⫺⫹1⫹
`⫹⫹2⫹
`⫹⫹1⫹
`⫺⫺1⫺
`⫹⫺2⫹
`⫹⫹3⫺
`⫹⫹2⫺
`⫹⫺1⫺
`⫺⫺1⫹
`⫺⫹3⫹
`⫹⫹3⫹
`⫹⫺1⫹
`⫺⫹1⫺
`⫹⫺3⫺
`⫺⫺2⫹
`⫺⫺3⫺
`⫺⫹2⫺
`⫺⫺3⫹
`⫹⫺3⫹
`⫺⫹2⫹
`⫺⫹3⫺
`⫺⫺2⫺
`⫹⫹1⫺
`
`marized in Table II, where the number of passing
`samples is listed out of three. Several observations can
`be drawn from these results.
`
`I. Nozzle Retraction Speed: Nozzle retraction rate
`plays an important role for distribution uniformity; the
`faster the retraction, the less uniform the coating. For
`example, glass powder testing for syringes coated with
`0.2 mg silicone and sprayed in range of 10 – 44 mm
`past
`the flange using 15 psi air-to-nozzle pressure
`shows less uniform coating at the retraction rate of
`50 mm/s (Figure 5a) than 200 mm/s (Figure 5b). Fast
`downward retraction offsets droplet velocity so that
`the droplets failed to reach the needle-end of the barrel
`despite a deep dive. This visual observation was con-
`firmed by the glide force measurement. All three sy-
`
`ringes coated at the 50 mm/s retraction rate show a
`constantly low glide force (⬍2 N) over the barrel
`length of 0 –38 mm (Figure 5a). At the high retraction
`rate (200 mm/s) all three syringes display an increas-
`ing glide force profile, initially below 2 N (between
`0 to 15 mm) and escalating dramatically for the rest of
`the barrel (Figure 5b), indicating the lack of lubrica-
`tion on the barrel near the needle.
`
`II. Nozzle Spraying Range: Intuitively, the move-
`ment and coverage range of nozzle spray within the
`barrel are critical to uniform distribution. The range
`should be wide to cover both ends of the barrel.
`However, a wide spraying range means that the nozzle
`has to travel a longer distance, which will prolong the
`cycle time. Therefore, to balance distribution unifor-
`
`144
`
`PDA Journal of Pharmaceutical Science and Technology
`
`Novartis Exhibit 2006.009
`Regeneron v. Novartis, IPR2020-01317
`
`
`
`cantly lower than the air pressure set by the pressure
`gauge due to pressure drop along the path between the
`pressure gauge and the orifice. The true pressure at the
`orifice is difficult to measure or predict. Another effect
`of air-to-nozzle pressure is on the size of silicone
`droplets; on the same photo it appears that 15 psi
`produced smaller droplets than 5 psi, which facilitated
`more even distribution of silicone coating.
`
`IV. Coated Silicone Amount: There is a clear trend
`that, regardless of the spraying condition, the higher
`the amount of coated silicone, the easier the syringe
`passes the glide force test. In a worst-case scenario
`where the siliconization parameters involved the high-
`est retraction rate (200 mm/s) and a shallow spray
`range (22–10 mm past the flange), and a low air-to-
`nozzle pressure (5 psi), syringes did not pass the glide
`force test until the coated amount reached 1.0 mg. It
`can be envisioned that even though silicone distribu-
`tion is not uniform (the lack of coating near the
`syringe end), the excessive silicone from the high-
`silicone-amount group may be pushed by the plunger
`rod during injection toward the needle side of the
`barrel. Despite this advantage, using a high amount of
`silicone is not desirable for biopharmaceutical prod-
`ucts where excessive silicone may be incompatible
`with the protein and jeopardize product quality (7–9)
`due to visible/sub-visible silicone particle formation
`(4, 10). The preferred silicone amount for the 1 mL
`long syringe is in the range of 0.2 to 0.5 mg per
`syringe (internal standard).
`
`Optimization Considerations
`
`The parameter investigation above was used to support
`the selection of optimum siliconization conditions. In
`the optimization test, we targeted a 0.3 mg silicone
`coating dose, which is near the low range of the
`preferred silicone amount. The selection of nozzle
`retraction rate was based on two criteria: (1) the re-
`traction rate should be as fast as possible to shorten the
`coating cycle and increase production speed, and
`(2) the retraction rate should not be too fast to com-
`promise distribution uniformity. Thus, the nozzle re-
`traction of 100 mm/s was chosen from the range of
`50 to 200 mm/s. Ideally, slow and long spray range
`should promote distribution uniformity. Since previ-
`ously the spray range of 44 to 10 mm past the flange
`displayed acceptable distribution uniformity, we tight-
`ened the nozzle spray range to 35–10 mm past the
`flange to further benefit cycle time reduction. For
`air-to-nozzle pressure, 10 psi represents the mid-point
`
`Figure 6
`
`Glass powder testing on syringes coated using the
`condition of 22–10 mm spraying range, 50 mm/s
`retraction rate, and 0.2 mg target silicone amount
`with two air-to-nozzle pressures, 15 psi (three sy-
`ringes on the right) and 5 psi (three syringes on the
`left).
`
`mity and cycle time, the nozzle spraying range and
`nozzle retraction rate should be considered at the same
`time. For example, for syringes coated using the noz-
`zle spraying range of 44 –10 mm or 22–10 mm past the
`flange (other coating parameters: 0.5 mg target sili-
`cone quantity, 15 psi air-to-nozzle pressure, 200 mm/s
`retraction rate), syringes with the deeper spraying
`range successfully passed glide force testing while
`syringes with the shallow spraying range failed the
`test. Therefore, with the fastest retraction rate, a pre-
`ferred position to begin spraying should be between
`22 mm and 44 mm past the flange. Ending spraying at
`10 mm past the flange appears to be acceptable.
`
`III. Air-to-Nozzle Pressure: By the law of conser-
`vation of energy, higher nozzle pressure can disperse
`droplets faster and farther. In Figure 6 for glass pow-
`der testing, despite the significant pressure difference,
`silicone coating is only marginally deeper using 15 psi
`air-to-nozzle pressure (three syringes on the right)
`than 5 psi (three syringes on the left). These syringes
`were coated under the condition of 22–10 mm spray-
`ing range, 50 mm/s retraction rate, and 0.2 mg target
`silicone amount. It should be noted that the air pres-
`sure coming out of the nozzle orifice may be signifi-
`
`Vol. 66, No. 2, March–April 2012
`
`145
`
`Novartis Exhibit 2006.010
`Regeneron v. Novartis, IPR2020-01317
`
`
`
`Figure 7
`
`Glass powder distribution of (a) syringes coated under the optimum condition (0.3 mg target silicone amount,
`100 mm/s nozzle retraction rate, 10 psi air-to-nozzle pressure, and 35–10 mm past the flange for the spraying
`range) and glide force testing of (b) five syringes above and of (c) another set of five syringes coated under the
`same optimum condition but with heating at 85 °C.
`
`of the tested range (5–15 psi) and appears to be
`appropriate because 5 psi might be too low to generate
`small droplets and 15 psi might be so high that it might
`disturb droplet flow for smooth deposition. In addi-
`tion, to further improve the atomization condition to
`produce finer droplets without using higher air-to-
`nozzle pressure, silicone was heated to a higher tem-
`
`perature, in the range of 85–100 °C (see Figure 1b for
`the heating mechanism) to reduce viscosity prior to
`atomization.
`
`Five syringes coated under the optimum condition
`were tested by the glass powder meth