Shear Stress Sensor for Peristaltic Pump Head as an in-line PAT Tool for
Protein Solutions Ajit S. Narang1, Vadim Stepaniuk2, Valery Sheverev2,
Mary E. Krause1, and Smeet Deshmukh1
1Bristol-Myers Squibb, Co., New Brunswick, NJ 2Lenterra, Inc., Newark, NJ
AAPS-NBCMay 2016
M1036
AbstractPurpose: Shear stress during processing can impact protein stability. Closed loop pumping of protein solutions, especially during tangential flow filtration, can impact the stability of sensitive proteins. Although shear stress rates have been modeled in tubular flow systems, direct, experimental measurement of the rate of shear stress on the protein in a peristaltic pump head has beenchallenging. In this study, we demonstrate the application of wall shear stress (WSS) sensor in quantitating shear experienced by viscous aqueous solutions in the peristaltic pump head.
Methods: An F-series Lenterra’s RealShear™ stress sensor was adapted to measure wall shear stress on the wall of plastic tube within and outside Flexicon PF6 peristaltic filling machine (Figure 1). Water (1cP viscosity) and two PEG (20,000 molecular weight) water solutions (6 w% and 10 w%) with viscosities of 7 cP and 13 cP, respectively, were used in tests. Also, some tests were carried out without any fluid, effectively pumping air. Flexicon PF6 peristaltic filling machine motor speed was in the range from 30 RPM to 250 RPM. WSS sensor response was recorded during pumping of three different fluids at a range of RPMs. The WSS measurement rate was held at 500 Samples/s throughout the tests. Flow rates were determined by measuring the weight of the fluid accumulated in a beaker in a set time interval. Pulse amplitude data was analyzed for various RPM for different fluids, both for the setup where the sensor was inside and outside the pump.
Results: All raw data plots showed periodic structure with pulses corresponding to successive roller occurrences. The tests demonstrated that the flow properties inside the peristaltic pump are of complex nature. The measured values of WSS are much greater than those estimated from the fully developed flow model. It was shown that these high values of WSS could not be explained by probe vibration and movement experienced during the tests. The temporal dependence of WSS was found to be different for fluids with different viscosities.
Conclusions: Measured magnitudes of the generated WSS pulses (several hundreds Pa) were significantly higher than the expected wall shear stress in a circular cross-section in fully developed flow (a fraction of Pa for realized flow rates). The shape of the pulse varied for different fluids: for lower viscosity fluids the pulse had a shape of two partly overlapping peaks with earlier peak being higher than the later one, and for higher viscosity fluids the pulse shape showed a single peak with relatively slowly rising front edge and a sharp back edge. In addition, reversal of the flow direction was observed in the pipe that is being compressed by theroller. Understanding of the flow patterns and the wall shear forces inside the plastic tubing in the pump can help understand the stresses experienced by protein solutions during processing.
2
Purpose: Shear stress during processing can impact protein stability
http://www.emdmillipore.com/US/en/product/Amicon-Ultra-0.5%C2%A0mL-Centrifugal-Filters-for-DNA-and-Protein-Purification-and-Concentration,MM_NF-C82301http://www.pall.com/main/laboratory/literature-library-details.page?id=34212
Normal flow filtration (NFF) Tangential Flow Filtration (TFF)
33
Effect of process shear on aggregation and viscosity after accelerated storage
0 10 20 30 40 500
10
20
30
40
storage temp (οC)
Visc
osity
(cP
)
0 10 20 30 40 500
5
10
15TFFNFF
storage temperature (οC)
%H
MW
S
Processing lead to increased aggregation and higher viscosity(TFF > NFF)
Protein was concentrated to 140 mg/L using an NFF process and a TFF process, then placed on station for 3 month
Viscosity Aggregation
4
F i L t R lSh ™ ll h t F-series Lenterra RealShear™ wall shear stress sensorProbe:
Sensor directly measures wall shear stress induced by fluids on construction walls in mixers, pipes, turbines, pumps, extruders and other devices
Probe is mounted flush with an inner surface of the flow channel wall to provide in line measurement with noprovide in-line measurement with no disruption of process flow
Measures both wall shear and temperaturep
Detection principle: Two optical strain gages, Fiber Bragg Gratings (FBG), are affixed to the
opposite sides of the cantileveropposite sides of the cantilever The FBG assembly detects a minute deflection of the tip of the pin (< 10
nm) and which is related to equivalent force on the tip Directional measurement – measures wall shear stress projection on the
5
p jplane formed by FBGs
Setup of stress sensor with peristaltic pump
Cross-section of plastic tubing t-junction and RealShear™ sensor
Sensing surfaceof the sensor
Sensor
Plastic tubing t-junction
Optical fibers connected to optical interrogator
Hermeticallysealed
Movable part of tube bridge
Rollers
T-junction tubing
Wall shear stress sensor
An F-series Lenterra’s RealShear™ stress sensor was adapted to measure wall shear stress (WSS) on the wall of plastic tube within and outside Flexicon PF6 peristaltic
filling machine
6
Sensor calibration: vibration evaluation
Time, s
4.2 4.4 4.6 4.8 5 5.2 5.4 5.6
Sen
sor
resp
onse
, Pa
-100
-80
-60
-40
-20
0
20
40
60
80
100Sensor response vs. time
Air: Movable part of the bridge was pushed up at ~4.3s and released at ~4.6s
Oscillation frequency: 76±1 Hz
Water: Movable part of the bridge was pushed up at ~12.15s and released at ~12.65s
Free mechanical oscillations of the probe floating element
Time, s
12 12.2 12.4 12.6 12.8 13 13.2 13.4
Se
nso
r re
spo
nse
, P
a
-100
-80
-60
-40
-20
0
20
40
60
80
100Sensor response vs. time
Oscillation frequency: 77±1 Hz.
While the sensor oscillation might generate some false signal in the sensor output, the signal shapes and levels observed in tests with PF6 are different
from calibration testing; cannot be explained by sensor shaking alone.
• In the flow path, complex probe motion is observed, with vertical andlateral displacements.
7
Method development: Studies with PEG and water
3 solutions tested at 30, 50, 75, 100, 150, 200,
and 250 rpm:Water
6% PEG (7 cP) 10% PEG (13 cP)
0
5
10
15
20
25
30
35
0 50 100 150 200 250 300
Flow
rate
, ml/s
Pump RPMWater PEG 7cP PEG 13 cP
Flow rate was measured for three different solution at different pump speeds
8
Calculation of force pulse magnitude (FPM)
• FPM - difference between the raw signal at a maximum value and the preceding minimum.• Characterizes passing of one roller under the probe pin.
FPM1 FPM2FPM3 FPM4
Blade 2 Blade 3Blade 1 Blade 4
FPM1 FPM2FPM3 FPM4
Blade 1 Blade 2 Blade 3 Blade 4
9
Pulse shape: sensor in tubing downstream of the pump
Time, s
19.8 19.85 19.9 19.95 20 20.05 20.1 20.15 20.2
WSS
, Pa
-50
-40
-30
-20
-10
0
10
20
30
40WSS vs. time
WaterTime, s
20.1 20.15 20.2 20.25 20.3 20.35 20.4 20.45 20.5
WSS
, Pa
-50
-40
-30
-20
-10
0
10
20
30
40WSS vs. time
PEG 7 cP
Time, s
19.9 19.95 20 20.05 20.1 20.15 20.2 20.25 20.3
WSS
, Pa
-50
-40
-30
-20
-10
0
10
20
30
40WSS vs. time
PEG 13 cP
Sensor was placed in tubing downstream of the pump and WSS was monitored
Periodic structure - pulses correspond to successive roller occurrences.
Pulse shape varies for fluids with different viscosities
10
Pulse shape: sensor inside the pump
Time, s
20.1 20.15 20.2 20.25 20.3 20.35 20.4 20.45 20.5
WSS
, Pa
-350
-300
-250
-200
-150
-100
-50
0
50
100
150
200WSS vs. time
PEG 7cP
Time, s
20 20.05 20.1 20.15 20.2 20.25 20.3 20.35 20.4
WSS
, Pa
-350
-300
-250
-200
-150
-100
-50
0
50
100
150
200WSS vs. time
PEG 13 cP
Time, s
20 20.05 20.1 20.15 20.2 20.25 20.3 20.35 20.4
WSS
, Pa
-350
-300
-250
-200
-150
-100
-50
0
50
100
150
200WSS vs. time
Water
High frequency oscillations observed in water are due to resonant oscillations of
the sensor’s floating element.
Pulse shape varies for fluids with different viscosities
Sensor was placed in the peristaltic pump and WSS was monitored
11
Impact of flow rate and viscosity on magnitude of WSS within the pump
• Magnitudes of WSS minimum and maximum are higher for fluids with lower viscosity (water)
• In general, minimal systematic change with flow rate.
050
100150200250300350
0 5 10 15 20 25 30 35
Min
(abs
val
ue),
Pa
Flow rate (mL/s)
Minimum (plotted as absolute value)
Water PEG 7cP PEG 13cP
0
50
100
150
200
250
300
350
Max
, Pa
Flow rate (mL/s)
Maximum
Water PEG 7cP PEG 13cP
• With the sensor in the peristaltic pump head, fluid was pumped for 1 minute with monitoring (triplicate for each fluid/flow rate)
• Data presented are averages of 10 roller occurrences from each of 3 pumping events.
12
Impact of flow rate and viscosity on FPM within the pumpFP
M -f
orce
pul
se m
agni
tude
FPM is higher for fluids with lower viscosity (water).
FPM decreases (slight) with increasing flow rate.
050
100150200250300350400450
0 5 10 15 20 25 30 35
FPM
, Pa
Flow rate (mL/s)
Water PEG 7cP PEG 13cP
13
Impact of flow rate and viscosity on magnitude of WSS in tubing
Higher WSS with lower viscosity
Higher WSS with increasing flow rate
020406080
100120
0 10 20 30 40
Min
, Pa
(abs
olut
e va
lues
)
Flow rate (mL/s)
Minimum (plotted as absolute value)
PEG 13cP PEG 7cP Water
020406080
100120
0 10 20 30 40M
ax, P
aFlow rate (mL/s)
Maximum
PEG 13cP PEG 7cP Water
• With the sensor in tubing downstream of the pump, fluid was pumped for 1 minute with monitoring (triplicate for each fluid/flow rate)
• Data presented are averages of 10 roller occurrences from each of 3 pumping events.
14
Impact of flow rate and viscosity on FPM in tubing
FPM increases with increasing flow rate
FPM is higher for fluids with lower viscosities
0
50
100
150
200
250
0 5 10 15 20 25 30 35
FPM
, Pa
Flow rate (mL/s)
PEG 13cP PEG 7cP Water
15
Summary of water/PEG studies: Fluid viscosity and flow rate impact shear
• Flow properties inside the peristaltic pump are of complex nature.
• The roller generates a negative pulse in all liquids and in air. • Floating element of the sensor moves against the flow direction in the pipe.• The observed reversal of the flow direction in the pipe that is being
compressed by the roller is not immediately obvious.
• The measured values of WSS are much greater than those estimated from the fully developed flow model.
• In general, WSS is higher for fluids with lower viscosities (water). Further understanding of FPM trends is needed
• WSS dependence on flow rate is different in the pump head (slight decrease with increased flow rate) and in the tubing downstream of the pump head
16
Application to protein system
• Protein was tested in different formulations with different viscosities
• pH 4.0• pH 4.0 + 60% sucrose (for high viscosity)
• Tests were only performed with the probe inside the peristaltic pump
• Measurements were recorded for one minute (in triplicate) at each flow rate
• Maximum, minimum, and FPM values were found for each roller
17
Pulse shape changes with protein formulation
Time, s
14.7 14.75 14.8 14.85 14.9 14.95 15 15.05 15.1
WS
S, P
a
-100
-50
0
50
100
150WSS vs. time
pH 4 sucroseTime, s
14.7 14.75 14.8 14.85 14.9 14.95 15 15.05 15.1
WS
S, P
a
-100
-50
0
50
100
150WSS vs. time
pH 4
• Pulse shape is different for sample with and without sucrose (effect of viscosity)
• Pulse shape is also modified in the presence of protein • pH 4 protein formulation has different pulse shape than water
(similar viscosity)
• Viscosity may not be the only factor impacting shear 18
Impact of protein on how fluid viscosity and flow rate effect shear
0.050.0
100.0150.0200.0250.0300.0350.0
0 10 20 30 40
Min
WS
S, P
a
Flow rate (mL/s)
Minimum (plotted as absolute values)
pH 4 sucrose pH 4
0.050.0
100.0150.0200.0250.0300.0350.0
0 10 20 30 40
Max
WS
S, P
aFlow rate (mL/s)
Maximum
pH 4 sucrose pH 4
• Magnitude of maximum/minimum WSS increases with increasing flow rate.
• Magnitude is higher with higher viscosity.
Data presented are averages of 10 roller occurrences from each of 3 pumping events.
19
Impact of protein on how fluid viscosity and flow rate effect shear
0.0
100.0
200.0
300.0
400.0
500.0
600.0
0 5 10 15 20 25 30 35
FPM
, Pa
pH 4 sucrose pH 4.0
For protein solutions, FPM increases with both increasing flow rate and with higher viscosity
20
Impact of shear on protein: degradation experiments
0
50
100
150
200
250
300
350
400
0:00 1:12 2:24 3:36 4:48 6:00
FPM
, Pa
Elapsed time, h:min
Sucrose pH4
pH4
Continuous pumping (6 mL/s) and monitoring every 15 min over 6 hours
See Poster M1037
• Shear impacts protein conformation (see poster M1307)
• Sucrose has a protective effect against shear stress21
Conclusions: Viscosity may not be the only factor impacting shear
• Fluid viscosity and flow rate impact shear differently in the presence and absence of protein
• Further understanding of FPM trends observed in the absence of protein is needed
• Shear impacts protein conformation
In the absence of protein: • FPM decreases with increasing flow
rate
• FPM is lower with higher viscosity
In the presence of protein:• FPM increases with increasing flow
rate.
• FPM is higher with higher viscosity.
For solutions with similar viscosity:
22