GPC Analysis of Polymers with an On-Line Viscometer Detector

GPCAnalysisof Polymers with an On-LineViscometerDetector

JurisL. Ekmanis- Waters Chromatography Division, Millipore Corporation

I I I / I I I I

Waters DMsionof MILLIPORE

FirstPresentedat the 1989 InternationalGPC Symposium,October 1-4, 1989 wf

" : GPC Analysis of Polymers with an On-Line Viscometer Detector

Juris L. Ekmanis

Waters Chromatography Division Millipore Corporation 34 Maple Street Milford, Massachusetts 01757

ABSTRACT

An on-line capillary viscometer detector for GPC, when used in conjunc- tion with a concentration detector (e.g. refractive index) and appropriate software, can be used to generate calculations of absolute molecular weight averages and intrinsic viscosity and provide data on the extent of any long chain branching in a polymer sample• This paper describes the chromatographic performance of a single-capillary viscometer detector that has been built into the Waters 150C high temperature GPC system• The complete viscometer detector system consists of hardware (capillary, transducer, electronics, etc.) incorporated within the new 150CV instrument along with data reduction software that is in the final stages of development (for Waters 845 and 860 data systems).

This paper describes the performance characteristics of the viscometer detector and demonstrates the use of this 150CV integrated GPC/viscometer system for the analysis of polymers in tetrahydrofuran (THF) at 35oC (polystyrene, polyvinyl chloride, polycarbonate) as well as in 1,2,4-trichlorobenzene (TCB) at 135oC (polyethylene and polypropylene). High temperature operation has been optimized by locating the viscometer components totally within the 150CV system. This eliminates the need for additional temperature-controlled zones that would be required if the viscometer were located in an external module•

Introduction

The technique of Gel Permeation Chromatography (GPC) was developed by John Moore at the Dow Chemical Co. (Freeport, TX) in the early 1960's 1 and the proprie- tary column technology developed by Moore was licensed to Waters Associates by Dow. Waters subsequently commercialized the GPC technique by introducing the first GPC instrument (GPC-100) to the market in 1964. The GPC-100 was a large, floor-standing system (6 feet tall) that could be used to determine molecular weight distributions of polymers in 3 to 4 hours, depending on column selection. This was a

-1- i. m'ajor, advance over the classical fractionation procedures that typically required many days, even weeks, to fully characterize the molecular weight distribution of a polymer sample before the advent of GPC. In 1967, the GPC-100 was replaced by the GPC-200 which was externally similar to its predecessor but included an updated pumping system and an improved RI detector (better temperature control and optics).

In 1979, the GPC-200 was replaced by the Waters 150C high temperature GPC system. The 150C is a fully integrated benchtop instrument which is used with high

efficiency, microparticulate GPC columns, e.g. Waters i_StyragelTM and Ultrastyra- gel® columns, to generate molecular weight distributions of polymers in approximately 30 to 60 min. with much higher resolution than was possible with the larger particle size columns (Waters Styragel®, 37 - 75 p.m) that were used with the GPC-200.

The 150C system includes a refractive index (RI) detector that monitors the difference in RI between the sample side of the RI cell through which the column effluent is flowing and the reference side of the RI cell which is filled with pure solvent. The refractometer is a concentration detector that can be used to calculate true molecular weight averages when an absolute calibration curve is available.

An absolute calibration curve for a GPC system can be generated by using narrow distribution polymer standards, if available. For organic solvents, only well characterized, narrow distribution polystyrene standards are readily available for calibration purposes. For polymers other than polystyrene, molecular weight averages based on polystyrene standards are often used for comparison among samples. Alternately, computerized techniques have been developed to generate an absolute GPC calibration curve from a single well characterized, broad distribution sample of the polymer of interest. The broad standard technique has its limits and must be used with care to generate good data.

In 1989, Waters added an on-line, single-capillary viscometer detector to the 150C. The new, dual-detector (viscometer, RI) 150CV system affords the following benefits.

1) Since the viscometer detector responds relatively more to high molecular weight materials and relatively less to low molecular weight fractions than does a concentration detector (e.g. RI), the viscometer will magnify small differences in the high molecular weight regions of the molecular weight distributions of polymers. This is important since small differences in this region often have a significant influence on the physical properties of a polymer.

2) A single-capillary viscometer is also a very sensitive flow meter and, as such, the viscometer is an excellent system diagnostic tool. Flow control in any GPC system is a critical element in generating accurate molecular weight averages and molecular weight distributions. With this viscometer, the operator can be aware of any flow discrepancy as it occurs so that appropriate corrective action can immediately be taken in order to restore the system to proper operation. Conversely, examination of the viscometer baseline for a series of overnight runs can be used to confirm that the GPC system was operating properly and that the chromatography is worthy of subsequent data reduction to calculate molecular parameters.

3) When used with appropriate data reduction software, which is in the final stages of development for the Waters 845 and 860 DEC-based data systems, the 150CV viscometer and RI detector signals can be used to calculate:

• absolute molecular weight averages (via Universal Calibration 2)

• intrinsic viscosity (T1)of the whole polymer • branching information across the molecular weight distribution via the branching parameter (g')

g' = 1] branched/ 1] linear

This paper describes the design and performance characteristics of the single- capillary viscometer detector that has been built into the existing Waters 150C GPC system and demonstrates the use of the new 150CV system for the analysis of polymers in tetrahydrofuran (THF) at 35oC (polystyrene, polyvinyl chloride, polycarbonate) as well as in 1,2,4-trichlorobenzene (TCB) at 135oC (polyethylene and polypropylene).

-3- Description of the 150CV and the Viscometer Detector

The 150CV GPC/Viscometer system is shown in Figure 1 along with the Waters 845 data station that is used for viscometrycalculations. The viscometer detector has been integrated withinthe chromatographicsystem and, except for the 150CV nameplate, the only external evidence that a viscometer detector is included in this unit is the viscometer electronics control panel in the lower left corner on the front of the instrument.

The block diagram (Figure 2) of the 150CV system indicates the key elements that constitute the viscometer detector. The internal solvent reservoir (rarely used) has been eliminated and solvent is supplied from an external source to the main pumping system (pump compartment) and the total system pressure transducer (PT) in standard fashion. In a regular 150C, the outlet of this pressure transducer is routed to the injector compartment. However, the viscometer detector is sensitive enough to detect the minor flow fluctuations due to the piston cross-over in a properly operating dual-piston pump. In order to successfully use the detector as a viscometer, we have included a pulse dampening system, or baseline optimization box (BOB), to reduce the fluctuations by a factor of more than 100. Since there is not enough space to locate this dampening system in the injector compartment, the dampeners are located in the pump compartment on the drawer just in front of the main pumping system. The baseline optimization box consists of a series of eight dampening and eight restrictive elements in alternating order and enclosed in a dual-wall container so as to minimize the effects of any temperature variations. The flow from the system pressure transducer is routed back into the pump compartment, through the BOB, and then to the injector.

From the injector, the solvent flows into the columns that are located in the column compartment. In the 150CV, the outlet of the last column, which is connected to the inlet of the RI detector in a regular 150C system, is instead connected first to the viscometer capillary and then to the inlet of the RI detector. The viscometer capillary is a 6 inch length of stainless steel tubing (1/16" O.D., 0.014" I.D.). The columns and viscometer capillary are both enclosed in a dual-wall container to isolate them from any temperature fluctuations in the column compartment. From tees at either end of the capillary, connecting tubing (stainless steel, 1/16" O.D., 0.020" I.D.) is routed into the pump compartment and connected to each side of a variable

-4- ,. r_luctance differential pressure transducer (Visc PT). Under normal operation, the solvent in these connectors is stationary. By opening outlet fittings at each side of the transducer, these lines can be flushed as needed to remove air or to change solvents. This viscometer transducer is also enclosed in a dual-wall container to insulate it from the effects of any temperature variations and the complete transducer assembly is located on the top of the dampening system (BOB).

The outlet of the RI detector is directed into a modified internal waste container in the pump compartment, just below the column compartment, and is continually drained to external waste• The output from the viscometer and RI detectors is displayed on the electronics control panel for each detector. The viscometer electronics panel is located in the lower left corner of the front of the 150CV in place of the blank panel (150C) that had been reserved for a future detector. The chromatographic output of each detector can be plotted on a chart recorder to monitor system performance and collected on a data system (Waters 845 or 860) which can be used to calculate sample parameters.

The full-scale range of the differential pressure transducer used in the viscometer detector is 5 kilopascals (KPA) or 0.73 psi. At a flow rate of 1 ml/min with THF at 35oC, the pressure drop across the capillary due to pure solvent is ~50% of full scale, i.e. -- 2.50 KPA. During elution of a polymer under typical chromatographic conditions, this signal increases by a maximum of --1% of the background signal (Po) and then returns to its initial value after sample elution is complete. The transducer is activated by a signal from the viscometer electronics panel and generates a 0 - 5 volt output signal which corresponds to a differential pressure across the viscometer capillary of 0 - 5 KPA. This transducer output is displayed on the viscometer electronics panel. A safety circuit has been incorporated into the electronics. The system is designed to trigger alarms (audible and visual) and stop the pump if a maximum pressure drop (operator selectable) is reached or exceeded.

The location of the key elements of the viscometer detector within the 150CV are shown in the photos in Figures 3 - 5. In Figure 3, the large box at the front of the pump compartment (door open) is the pulse dampening system (BOB) and the viscometer pressure transducer (Visc PT) is mounted on top of the BOB. In Figure 4, the box which encloses the columns and viscometer capillary is located in the column compartment at the top of the 150CV. The container is located between the

-5 - ( ref.ractometer (right, front) and the heating elements (rear) and can easily accom- modate up to six (6) 30 cm long GPC columns. Finally, in Figure 5, the viscometer electronics panel is located in the lower left corner at the front of the system.

Viscometer Detector - Principle of Operation

Poiseuille's Law

When fluid is pumped through a capillary tube, Poiseuille's Law relates the pressure drop across the capillary to capillary dimensions, fluid viscosity and flow rate:

P = (L/r4)(q) (F)

where:

P = pressure drop across the capillary

L = capillary length

r = internal radius of the capillary

- viscosity of fluid (GPC column effluent)

F = flow rate

This relationship indicates that:

a) Pressure drop is directly proportional to capillary length. This should evoke no surprise since it should be intuitively obvious that doubling the length of the capillary will double the measured pressure drop.

b) Pressure drop is inversely proportional to r4. Again, it is very reasonable that P decreases as the internal radius (or internal diameter, I.D.) of the tube is increased. However, it is very important to note that since the relationship is to the 4th power of the radius, a small increase in capillary I.D. will result in a significant decrease in the pressure drop. For example, increasing the I.D. from 0.012" to 0.014" for a 1/16" O.D.S.S. tube will reduce the pressure drop to 54% of its original value. Doubling the I.D. reduces the pressure drop by a factor of 16.

-6- ' c) P'ressure drop is directly proportional to fluid viscosity (Tl). AS fluid viscosity increases, so does the measured pressure drop across the capillary.

d) Pressure drop is directly proportional to flow rate (F). As flow rate increases, so does the measured pressure drop across the capillary.

Consider two situations when the capillary dimensions (length, internal radius) are fixed.

At Constant Viscosity, Capillary is a Flow Meter.

If we pump a solvent (e.g. THF) at a constant temperature (e.g. 35oC) through the viscometer capillary of the 150CV system, the viscosity of the fluid is constant. In this case, Poiseuille's law can be simplified. The pressure drop is then directly proportional to the flow rate and the viscometer detector becomes a very sensitive flow meter.

P = K' F (Flow Meter Function)

In order to use a single-capillary viscometer to measure viscosity in GPC, very precise control of flow rate is required. For a system operated at 1 ml/min, maximum response due to a viscosity change of the column effluent during elution of a typical polymer sample is similar to that due to a flow rate change of -101_l/min (1% of total flow rate). For a signal/noise ratio of at least 100/1, flow rate fluctuations must be controlled to less than 0.1_l/min.

At Constant Flow Rate, Capillary is a Viscometer

In order to use the system as a viscometer, the flow rate of the 150CV is controlled to the required level by the pulse dampening system in the baseline optimization box. With flow rate controlled towithin required limits, Poiseuille's law is again simplified and the pressure drop across the capillary is directly proportional to

fluid viscosity (T1).

P = K"_ (Viscometer Function)

-7- During the course of sample elution from the outlet of the GPC columns in the 150CV, the viscometer baseline (pressure drop, P) is constant as long as the solvent flow rate and capillary temperature are controlled to constant values. After sample injection, as sample fractions elute from the GPC columns and pass through the viscometer capillary, the viscosity of the sample/solvent mixture increases

beyond that of pure solvent. The increase in viscosity (q) is very small because the sample concentration is very low. At its maximum, the pressure drop (P) typically increases only ~1% above Po, the pressure drop with pure solvent. The viscometer detector chromatogram is a record of this small change in overall pressure drop as sample fractions elute from the columns.

As discussed earlier, the full-scale range of the differential pressure transducer in the 150CV is calibrated to 5 KPA (~0.73 psi). The design of this transducer is such that, although the maximum pressure differential (P) that can be measured is small, the actual "line" pressure can be as high as 3200 psi. In the 150CV system, only the refractive index detector is downstream from the viscometer capillary and thus the actual line pressure is relatively low (typically ~20-25 psi).

The viscometer detector responds relatively more to high molecular weight materials and relatively less to low molecular weight fractions than does a concentration detector, e.g. RI, which responds to sample concentration for a given polymer. This is demonstrated in Figure 6 which shows a series of five individual injections of narrow distribution polystyrene standards. Analysis conditions, viscometer detector sensitivity, and injection volumes are identical for each sample. Notice that, even though the concentration of the polymer samples is decreased for the higher molecular weight rnaterials, the response of the viscometer detector still increases with molecular weight. For the two standards injected at the same concentration (422 K, 107 K), the increased response can be observed directly.

150CV Refractometer and Viscometer Detector Responses

The 150CV system is equipped with two detectors, a differential refractive index detector and a single-capillary viscometer detector. Each detector responds to eluting sample in a different way and the combination of both detectors is a powerful tool for polymer characterization.

-8- I:HDetector

The differential refractive index detector is a concentration detector and affords a signal directly proportional to the concentration of polymer in the eluent. For a given polymer/solvent system, the RI response is independent of the molecular weight of the polymer for molecular weights above ~2,000.

AM time

Assume the following definitions:

Ai = area of slice i

A = total area of RI response for polymer sample

C = concentration of polymer in solution injected

V =injection volume

AV = volume of slice i

Then, the concentration (Ci) of sample in any slice in the polymer region of the RI response can be calculated using the RI data as follows:

C, = (C. V. A,)/(z&V. A)

Viscometer Detector

The viscometer generates a signal equal to the pressure drop across the small viscometer capillary. When no sample is eluting from the columns, the viscometer measures the pressure drop (Po) of the pure solvent flowing through the capillary. This background pressure is typically at 40 to 70% of the full scale range of the differential pressure transducer (5 KPA). For example, in THF (35oC) at a flow rate

-0- of,l" ml/min, Po is -2.5 KPA with the standard 6" x 0.014" I.D. capillary. The polymer signal is a very small change in this background signal. During elution of the polymer, the maximum pressure drop increases only ~1% above the value of Po.

_,= Viscometersignal time-,

Po Pi =pressuredropdudng elutionofslicei

-0 'r

A computer data system can easily plot the response of the viscometer detector by subtracting most of the baseline signal. In order to plot this polymer signal on a standard chart recorder, the viscometer electronics controls on the 150CV front panel are used to offset the background signal (Po) so that the small, incremental polymer signal can be expanded -50 to 100x to properly display the chromatogram. For GPC analyses in THF, the 5 volt output signal of the viscometer transducer is passed through an auto zero circuit to offset the background pressure drop (Po) and then displayed on a 50 mv chart recorder (100x amplification) to plot the chromatographic output of the viscometer detector.

Using the relationship for reduced viscosity, the intrinsic viscosity of any slice (i) in the viscometer signal can be calculated as follows.

[Tim=] (P=- Po)/ Po)

where, [Tli] = intrinsicviscosityof slicei

Ci = concentration of sample in slice i

Pi - total viscometer signal at slice i

Po = viscometer background signal at slice i

-10- Experimental

All chromatographic data was generated with a Waters 150CV GPC/Viscometer system which has been described in detail in a earlier section. The GPC columns consisted of a set of 10 3, 104, 10 5, and 10 6 ,_ Ultrastyragelcolumns (each column, 7.8 mm I.D. x 30 cm). Sample concentrationswere 0.1% (w/v), except as indicated (see Figure 6), and injectionvolumeswere 400t_1.The chromatographicoutput was monitoredwitha Waters 845 data system (DEC-based) as well as a 2-pen stripchart recorder. Molecular weights listed for typical samples are as reported by the suppliers.

GPC analyses in THF were performed at a flow rate of 1 ml/min with unstabilized THF that had been filtered by vacuum through a 0.45 _m fluorocarbon membrane. All temperature-controlled compartments (pump, injector, column/detector) were operated at 35oC, Under these conditions, total system backpressure was 100 Bars, with 70 Bars due to the pulse dampening system (BOB) and 30 Bars due to the columns. The pressure drop (Po) across the viscometer capillary due to the THF solvent was 2.50 KPA.

GPC analyses of polyolefins in TCB were performed at a flow rate of 1 ml/min with 1,2,4-trichlorobenzene that had been filtered by vacuum through a 3-cm depth of dry silica (55 to 105 lu.m)and a 0.45 I_m fluorocarbon membrane. The silica was placed above the membrane in the upper portion of the filtration assembly so that the solvent purification process could be done in one step. Polyolefin samples were dissolved in TCB containing antioxidant (Santonox R 3) at 250 ppm by heating in an external oven at 165oC for 2 hours. The solutions were then transferred to the

t, 150CV filter assemblies, equilibrated at 135oC in the injection compartment, and filtered prior to injection.

In TCB, the pump compartment was operated at 60oC (maximum) and the injector and column/detector compartments were both operated at 135oC. Under these conditions, the pressure drop across the column set was still 30 Bars since the viscosity of TCB at 135oC is similar to that of THF at 35oC. However, since the pulse dampening system is located in the pump compartment which was operated at 60°C, the pressure drop across the BOB increased (vs. THF at 35oC) to 200 Bars. The total system backpressure at high temperature in TCB was 230 Bars, well within the

-11- 408, Bar (6000 psi) limit for the system. The pressure drop (Po) across the viscometer capillary due to the TCB solvent was 2.75 KPA.

Applications in THF (35°C)

In order to demonstrate the need for the pulse dampening system (BOB) in the 150CV if the single-capillary viscometer is to be used as a viscosity detector in GPC, consider the chromatogram (Figure 7) from a 400 ILlinjectionof a 0.1% solution of Dow 1683 polystyrene(Mw = 250K, Mn = 100K) onto a set of four Ultrastyra- gel columns in a 150CV system from which the BOB had been removed. The signal-to-noise (S/N) of the viscometer output was approximately one, unaccept- able for use as a chromatographicdetector. When the analysiswas repeated after addition of the 150CV pulse dampening system (Figure 8), the flow-related pulsa- tions were reduced by a factor of ~150. The resultingbaseline noise was ~0.2% and the S/N of the viscometerdetectorwas excellent (> 100).

Analysis of the first batch of ARCO 1085 polystyrene 4 (Mw ~480K) that was supplied to Committee D20 of ASTM for possible use as a round robin sample is shown in Figure 9. The chromatographic output from the Waters 845 data system reveals that the RI output has a shoulder on the high molecular weight side of the peak apex. The viscometer responds relatively more to high molecular weight components than does the RI detector. For this reason, the shoulder on the RI has increased to become the apex of the viscometer output and the apex of the RI peak has decreased to a shoulder on the low molecular weight side of the viscometer peak. The physical offset between the two detectors is minimal (~1001J.I)and the apparently large "offset" between the maximum responses of these two detectors is due to the different response characteristics of the viscometer and the RI. The apex of the viscometer response is at the maximum value of concentration times intrinsic viscosity and the apex of the RI detector is at maximum concentration. For the gaussian chromatogram in Figure 8, the effect is smaller (~1.2 ml) and less dramatic than with the ARCO 1085 polystyrene but the "offset" between the maximum responses of the two detectors is due to same effect.

A typical polyvinyl chloride sample (Mw ~132K) is shown in Figure 10 and the viscometer response is excellent. The viscometer detector in the 150CV system also affords good signal-to-noise for a low molecular weight (Mw ~28K) pol_,car-

-12- •. ' bonate (Figure 11) that was injected as a 0.1% (w/v) solution, as were the other polymers described above. The viscometer and RI detector sensitivities are the same in Figures 8, 10, and 11 so that a direct comparison of detector response from sample to sample is valid.

Since the polycarbonate is a lower molecular weight sample than the Dow 1683 polystyrene (Figure 8), the viscometer response is lower for the polycarbonate but the chromatogram is still good. The concentration of a sample in GPC is typically limited by the need to avoid undesirable viscosity effects ("viscous fingering") which occur upon injection if the solution is too viscous. As viscometer detector response decreases for lower molecular weight materials, sample concentration can be increased to the point where the area of the polymer peak on the viscometer trace is the same as for a higher molecular weight polymer which has been ana- lyzed under conditions (concentration) for which "viscous fingering" effects were not a problem. For this reason, the polycarbonate could easily be injected at a concentration of 0.2% to improve the detector signals still further with no undesirable chromatographic effects.

The analysis of three samples of polyvinyl chloride (PVC) in Figures 12 and 13 demonstrates how the increased response of the viscometer detector to the higher molecular weight fractions can reveal differences among samples which would be difficult to confirm with only an RI detector. A polymer fabricator was informed that his normal source of PVC used for fabrication of a medical filter device would no longer be available. This device was molded with Luer-tip connections at both ends to facilitate its use in filtering intravenous liquids that were being administered to hospital patients. QC testing of the molded parts required wetting of the Luer tip with isopropanol ("rubbing alcohol") and twisting it into a Luer connector to a specified torque without stress cracking.

Figure 12 shows the RI detector chromatograms of the original/good material that was to be discontinued (Sample A) along with two PVC samples (B, C) that were to be evaluated as potential replacements. Physical testing indicated that Alternate #1 (Sample B) was acceptable but that Alternate #2 (Sample C) exhibited stress cracking and could not be used. Comparison of the RI curves indicates that Sample B is a somewhat higher molecular weight material than Sample A (original) and that Sample C is even higher in molecular weight than Sample B. Since Sample B works

- 13- well with this molding process, one might conclude that a slight increase in molecular weight is acceptable but that the greater increase with Sample C is not.

Notice the very small response on the baseline in the very high molecular weight region of Sample C. This might be a very small concentration of some very high molecular weight material which does not seem to be present in Samples A or B and which might have adversely affected the performance of Sample C in this application. Even small amounts of high molecular weight material can radically affect processing characteristics and/or the quality of fabricated parts. Unfortunately, the response is so small that it would be difficult to convince a vendor that this is any- thing other than a slight problem in the RI baseline.

Since the viscometer responds more to higher molecular fractions than does the RI detector, the viscometer trace of Sample C should afford a more significant response for the high molecular weight material if this small FII response is actually part of the sample. Inspection of the three viscometer traces in Figure 13 indicates that there is indeed some real material in the high molecular weight region of Sample C. A smaller amount is also detectable in Sample B. In addition to the overall shift of the molecular weight distribution of the PVC to higher molecular weight, the viscometer detector has uncovered another dimension that can be investigated in order to gain a more thorough understanding of these PVC samples.

Applications in TCB (135oC)

The 150CV system was converted from THF (35oC) to TCB (135oC) using standard procedures. The analysis of NBS 1475 polyethylene (MW -52K) under standard conditions (400 til, 0.1%) is shown in Figure 14. The noise on the viscometer baseline has increased by a factor of approximatelyfive (5) on changingfrom THF to TCB, principallydue to thermal turbulencein the injectioncompartment. The baseline noise on the 845 computer display is increased further by expanding the size of the viscometerpeak for this low molecularweight sample to be the same size as the RI peak. The peak at ~41 min on the RI trace is solvent-relatedand is notpart of the sample. Using identicalconditionsfor the analysisof a sample of highermole- cular weight polypropylene(Mw ~300K) affords a greater viscometer signalthan for the NBS 1475 polyethylene (PE). Less amplification is required to expand this larger signal to full scale on the computer display and the resultingbaseline noise

- 14- appears to be smaller than before.

The GPC analysis of three polyolefins in Figures 16 -18 is another example of how the viscometer detector amplifies differences among samples. Based on the RI detector traces (Figure 16), there are only very small differences between the two polypropylene samples (A, B) and both PP samples are much lower in molecular weight than the polypropylene/polyethylene (PP/PE) block copolymer. The chromatogram for Sample A lies below that of Sample B in both the high and low molecular weight regions. Since the same amount is injected for each sample, the areas of the three RI peaks are essentially the same. The data for intrinsic viscosity (IV) and melt index (MI) were supplied by the customer who submitted these samples for analysis.

The viscometer response for the three polyolefins is shown in Figure 17. The chromatograms are displayed at the same detector sensitivity. As expected, the higher molecular weight sample (C) affords much more response than the other two materials (A, B). In comparing Samples A and B, the difference between these two samples in the high molecular weight region has been magnified (vs. RI data). A more direct comparison between Samples A and B can be made in Figure 18 where the viscometer chromatograms are scaled to the same size as the RI data. The viscometer output shows a greater difference between these two samples than does the RI detector. In this way, the viscometer can be very beneficial in a QC environ- ment where chromatograms are to be compared without calculating molecular weight averages.

Conclusions

The 150CV GPC/Viscometer system is a completely integrated GPC system equipped with viscometer and refractive index detectors. Since the viscometer responds to the product of sample concentration and intrinsic viscosity, it responds relatively more to higher molecular weight materials than does a concentration detector (e.g. RI). In this way, the viscometer detector can magnify small differences that may be observed in the RI chromatograms of polymer samples. The viscometer detector is also a very sensitive flow meter and, as such, is a very sensitive and powerful diagnostic tool which continually monitors the performance of the GPC system and indicates when instrument service is required. From an operational

15- point of view, the viscometer equilibrates quickly (10 - 15 rain) and is ready much sooner than the time required to equilibrate the RI detector in a typical GPC system.

The 150CV system has been optimized for routine operation over a wide range of GPC conditions and the viscometer detector components (pulse dampeners, capillary, transducer) have been designed to allow operation in THF at 35oC as well as at high temperature in TCB (e.g. 135oC) without the need for any system modifica- tions. When used with the GPC/viscometry software that is in the final stages of development for the Waters 845 and 860 (DEC-based) data systems, the 150CV can be used to calculate absolute molecular weight averages (via Universal Calibration), intrinsic viscosity, and branching information for polymers. High temperature operation has been optimized by locating the viscometer components totally within the 150CV system. This eliminates the need for additional temperature-controlled zones that would be required if the viscometer were located in an external module.

References:

1) Moore, J. C., J. Polym. Sci., A2, 835 (1964)

2) Grubisic,Z., Rempp, P. A., and Benoit,H., J. Polym. Sci., B5, 753 (1967)

3) SantonoxR is a trademartk of the MonsantoCompany

4) Thissample was obtainedfrom Dr. G. Hagnauer(U. S. Army MaterialsTesting Laboratory,Watertown,MA)

- I6- Figure 1. 150CV GPC/Viscometer System with 845 Data Station

Pump _ Solvent

Dampeners VlscometerCapillary (,,

Ylsc P T

Waste

150CV Panel Recorder Electronics I Vlsc°rneter _ Chart I R' [ RI I Data System l .,-]_,,,°o.,r I

Figure 2. Block diagram of the major components within the 150CV system

-17- , j

F_gure3. 150CV system with pump compartment door open to show the location of the pulse dampening system (BOB) and the viscometer pressure transducer (Visc PT)

Figure 4. 150CV system with column compartment cover open to show box which encloses the GPC columns and the viscometer capillary

-[8- Figure 5. The viscometer electronics control panel is located at the lower-left corner on the front of the 150CV.

1.26 M 0.01% THF, 35°C

400_1PS standard solutions 422 K Ultrastyragel columns 0.02% (103, 10_, 10s, 106_) 1 ml/min

107 K

16.7 K 0.05% 2.8 K 0.15%

Figure 6. Viscometer response (narrow distribution polystyrene standards) increases with molecular weight.

-:19- I I Without 150CVPulse Dampening System

THF,35°C 400t_01.1%(w/v)solution Ultrastyragelcolumns (103,104, 10s,10sA) 1 ml/min

25 30 35 40 min

Figure7. GPCanalysisof Dow1683polystyrenwithe 150CVsystemfrom whichpulsedampeningassembly(BOB)hasbeenremoved.

-- _- ¢- _- _.... -- ___/_!_j_ = -_...... ,...... - _-i_i_.-_yA_- THF, 35°C

=T ;.------=_L-- +-_I_\'_ +:,:, l _ 40u01.d,0.1%t_t(w/ycv)solution_o,._um.s , Z-+-.+-,++t--+--:.+J_f+_= : '_ --:- J --_--.i'--:::-__k_-_- =-=_-- visc_. i =vmi. _-----=- :=---"= i (10a,104, 10s, 10sA) +-_- I -,L_ ...... _:-_- --_1 -_+-i-_1....I:+7--I -"+"l-__. <.....-! ! .... 1 :\t----L ' i . .;...+_ ._-:_.-_iH;+_-Ri __+I+- ../. • .---.-;_.,.---.>_._:_'_,---_:-I--_.---,....._,_.- +"...... ,--i, • ,-- _...... -.- _--r=--_--_, ...... --I -i ;--- -

25 30 35 40 min

Figure8. GPC analysisof Dow1683polystyrene(Mw ~250K)withcomplete 150CVsystem(includingpulse dampeningassembly)

- 2O - THF, 35°C

4001_O.I I% (w/v) solution LJltrastyragelcolumns (103, 104, 105, 106A) 1 ml/min

Visc RI

24 28 32 36 40 min

Figure 9. GPC analysis of ARCO 1085 polystyrene (Mw ~480K) with 150CV

THF, 35°C

400p.I0.1% (w/v) solution Ultrastyragelcolumns (103, 104, lOS,lo6A) 1 ml/min

25 30 35 40 min

Figure 10. GPC analysis of polyvinyl chloride (Mw -132K) with 150CV

-21 - !-- - -...... I '......

.... -_=_--...... " -_ -::: 400_ 0.1% (w/v) solution _ _ t_ r---//- _"--E---_-__t- " - Ultrastyragelcolumns ...... I:-/_I _---_:- - . i '- -t'_ ! -1-_,-- i " (103, 104, 10s, 10sA). I 'i--,: •- _\ RI - ._: __ / i _-. _. ---'-- _ --i _.°__:---L-:I:-._.,..._Z'. -

25 30 35 40 min

Figure 11. GPC analysis of polycarbonate (Mw ~28K) with 150CV

Sample PerformanceProduct THF, 35o0 400pl 0.1% (w/v) solution A Good. Standard Ultrastyragelcolumns B Good - Alternate #1 ! (103, 104, 10s, 10eA) 1 ml/min C Bad. Alternate #2

C /_ RI Detector

time --,

Figure 12. GPC analysis (RI detector) of three samples of PVC used to fabricate filter housings for a medical application

- 22 - 4

Product THF, 35°C Sample Performance 400_10.1% (w/v) solution A Good - Standard Ultrastyragelcolumns B Good " Alternate #1 (103' 1014m' l105'/min106A) C Bad - Alternate #2 eter

B A time --,

Figure 13. GPC analysis (viscometer detector) of three samples of PVC used to fabricate filter housings for a medical application

TCB, 135°C

4001_10.1% (w/v) solution Ultrastyragel columns \ (103, 104, 105, 106/_) '\ i ml/min

\ \

Vise

25 30 35 40 min

Figure 14. GPC analysis of NBS 1475 polyethylene (Mw ~52K) with 150CV

- 23 - TCB, 135°C

4001_10.1% (w/v) solution Ultrastyragelcolumns (103, 104, 10s, 106A) 1 ml/min

Visc

24 28 32 36 40 min

Figure 15. GPC analysis of polypropylene (Mw ".300K) with 150CV

TCB, 135oC

a 400pJ0.1% (w/v) solution Uitrastyragelcolumns C (1o3,1o4,1os,lO6A) 1 ml/min

A RI Detector

Reported Sample IV MI

A (PP) 2.00 12.6 RI B (PP) 1.97 13.8 RI C (PP/PE) 2.97 2.1

25 30 35 40 min

Figure 16. GPC analysis (RI detector) of three polyolefin samples using 150CV

- 24 - TCB, 135°C

C 4001110.1% (w/v) solution Ultrastyragel columns (103, 104, lOS,106A) 1 mr/rain Viscometer B \ Reported Sample IV MI

A A (PP) 2.00 12.6 B (PP) 1.97 13.8 C (PP/PE) 2.97 2.1

Visc Visc

25 30 35 40 min

Figure 17. GPC analysis (viscometer detector) of three polyolefin samples using 150CV

B TCB, 135oC

400pl 0.1% (w/v) solution Ultrastyragelcolumns A (103, 104, lOS, 106A) 1 ml/min B

Reported Sample IV MI

A (PP) 2.00 12.6 RI B (PP) 1.97 13.8

25 30 35 40 min

Figure 18. GPC analysis of two polyolefins using 150CV (viscometer chromatograms amplified for direct comparison with RI chromatograms)

- 25 -

I!mBIIIIBIB_T :: ' - ' Waters GPCProducts. " - Maximizing the informationyou get from a single injection.

DataManagement 150CVGPC/ViscometrySystem advanced data managementproducts.

l When you buy a modularsystemfrom ...... t Waters, you get top performancefor manyyears and the greatestlong-term I' , valueof any GPC system.

•,,:I_! i'::= ,2 _ ! • ChemicalProducts A complete line of columns,stan- Waters 150CV GPC/Viscometry dards, and samplepreparation de- Systemgivesyou moreinformation vicesfor small molecule,large tool- about a polymerthanany singleGPC ecule, high temperature,aqueous, Singleand multi-systemdata acqui- system.Forthefirsttime, polymer and organic polymer separations. sition and control.ProprietaryMaxima chemistscan get accuratemolecular and ExpertEaseGPC Softwarefor weight and molecularweight distribu- Supportand Service obtaining accuratemolecularweight tJoninformationon polymersfromone Our sales,service,and applica- distributionand molecularweight injection. The 150CV is temperature- tionspersonnelall speak thesame values. PCand VAX-basedcomputers, controlledto 150°C for analyzing language-chromatographic polymer Customreportformatting.Waters high-temperature-solubleas well as characterization.No matterwhere softwarebringsa new level of so- room-temperature-solublepolymers, you are, whenyou need advice about phisticationto GPC calibration, data Therefractiveindex/viscometrydetec- systemmaintenanceor operation, reductionand systemcontrol, tor combinationenablesaccurate troubleshooting,or methodsdevelop- molecularweights, intrinsicviscosity, ment,you're assuredof a timelyand Detection and branchinginformationto be effectiveresponse. calculated.

Modular GPC Systemsfor Polymer i Characterization • ' ----' Waters - ] DMsionof MILLIPORE if ii,il ,_I- .... JJ II_H Waters ChromatographyDivision Millipore Corporation Photodiodearray, fixed, variable 34 Maple Street and programmableUV/Vis. Refractive Milford, MA 01757 index. Conductivity.Electrochemical. 508-478-2000 Fixedand programmable/scanning wavelengthfluorescence.Fromprod- uctqualil'_control; to polymer research and dew.,Iopment,Waters offersthe polymer_hemistthe bestchoice of Whetheryou are interestedin polymeror additive analysis,or doing detectors, qualitycontrolor researchand development,Waters hasa modularGPC systemto meetevery budget. Engi- neeredto be compatible with all other Waters products,our modularGPC systemscan be enhancedwith mul- tiple detectors,gradient solventdeliv- ery systems,autoinjectors,and

@1989MilliporeCorporation,PrinledintheU.S.A.,8/90, T132