WO1995008777A1

WO1995008777A1 - Nmr on-line viscometer for liquids or polymer melts

Info

Publication number: WO1995008777A1
Application number: PCT/US1994/010068
Authority: WO
Inventors: Christian I. Tanzer
Original assignee: Auburn International Inc
Current assignee: Auburn International Inc
Priority date: 1993-09-23
Filing date: 1994-09-09
Publication date: 1995-03-30
Anticipated expiration: 1996-03-23
Also published as: AU7722194A

Abstract

A device (Figures 1 and 2) and method for measuring on-line viscosity of sample liquids (2) and polymer melts (Figure 2) using a Carr-Purcell-Meiboom-Gill (CPMG) (Figure 3) or other suitable pulse sequence wherein a free induction decay (FID) (A and B) echo train occurs, and where the FID's are analyzed to yield the time constants (T2's) and other information. The T2's are calibrated via known standards to yield viscosity of the sample material. Temperature control (24) is provided since viscosity is dependent on temperature.

Description

NMR ON-LINE VISCOMETER FOR LIQUIDS OR POLYMER MELTS

CROSS REFERENCE TO RELATED APPLICATIONS

This application is closely related to U.S. Patent Application, S.N. 07/961,264, filed Oct 15,1992, entitled "Apparatus to obtain Flow Rates in Plastics via Fixed Frequency, Pulsed NMR" and to U.S. Patent Application, S.N. 08/075,799, filed June 14, 1993, entitled "Thermal Control for Sample Under Test in an On-Line NMR System". These patent applications are of common assignment with this application, and the disclosures of all are hereby incorporated herein by reference, as though set out at length herein.

FIELD OF THE INVENTION The present invention relates generally to pulsed NMR spectroscopy and spin-spin relaxation process theory as applied for determining viscosity of liquids and polymer melts .

BACKGROUND OF THE INVENTION Pulsed NMR technology produces free induction decay (FID) signals generated in samples under test. These FIDs have decay time constants (T2's) which are dependent upon the viscosity of the liquid samples. As viscosity increases, either through composition changes or temperature changes, the T2 values decrease; and as viscosity decreases, T2 *s increase.

A pulse sequence, known in the art as the Carr-Purcell-Meiboom-Gill (CPMG) Sequence, has been developed for the accurate measurement of Spin-Spin or transverse relaxation times (T2's) . Measuring T2 's by line-width or directly from free induction decay (FID) is inaccurate if the magnet and sample inhomogeneity factors are major contributors to the decay of the NMR signal, i.e. the magnet and sample inhomogeneity factor is comparable to or greater than the actual T2 of the sample. This is particularly the case with low-field time domain instrumentation used for on-line measurement applications. The effect of the CPMG pulse sequence is to remove effects due to magnet inhomogeneities as well as pulse imperfections such that the measured NMR response reflects the true T2 of the spins. Thus relatively low resolution magnets may be employed to measure signals with very long time constants.

There is a continuing need for viscosity measurements in the chemical industry, particularly in the polymer chemical industry where product specifications are often related to the average molecular weight of the desired products, wherein the average molecular weight is reflected in the viscosity of the polymer in liquid form (polymer melts) . Average molecular weight is a measure of polymer quality and viscosity.

The NMR technique has not yet been applied to on-line viscosity measurement, and present on-line viscosity equipment has significant maintenance and calibration problems, particularly when used for polymer melts and when large shifts in viscosity occur.

An object of this invention is to overcome the above illustrated limitations and problems with an on-line NMR approach.

It is another object of the present invention to provide on-line measurements via a free flow chamber requiring little maintenance, and where calibrations are performed over a range of viscosity values without requiring equipment changes and re-calibration.

It is another object of the present invention to maintain temperature control of on¬ line samples. It is another object of the present invention to provide calibration data where T2 values for the sample under test are compared to T2 versus viscosity measurements of standards (of known viscosity) such that sample viscosity can be calculated.

SUMMARY OF THE INVENTION

The oregoing objects are met in an on¬ line NMR system and a corresponding process for measuring viscosity of liquids or polymer melts comprising:

(a) means for separating a sample from a flowing stream or a polymer melt,

(b) means for positioning and holding said sample in a sample chamber,

(c) means for generating a CPMG or other suitable NMR excitation pulse sequence,

(d) means for transferring said NMR pulse sequence into said sample, (e) means for receiving spin echo responses to the NMR pulse sequence from said sample, (f) means for generating viscosity calibration data from echo responses to NMR pulse sequences applied to known samples,

(g) means for comparing said received spin echo responses to said calibration data whereby viscosity of said sample is determined, and (h) means for removing the sample .

In addition means may be provided for temperature control of said sample during NMR -. measuring, wherein the sample temperature is stable and at a known temperature compared to the flowing liquid or the polymer melt. Typically that known temperature is about equal to the temperature of the flowing liquid or the polymer melt. Other objects, features and advantages will be apparent from the following detailed description of preferred embodiments thereof taken in conjunction with the accompanying drawings in which:

BRIEF DESCRIPTION OF THE DRAWINGS FIG. 1 is a functional block diagram/schematic in cross section;

FIG. 2 is a block diagram used with heated samples;

FIG. 3 is a graphical representation of a preferred pulse sequence;

FIG. 4 is a graph of the echo response from motor oils; FIG. 5 is graph of a least squares curve fit;

FIG. 6 is a calibration plot; and

FIG. 7 is a graph of temperature effects. DETAILED DESCRIPTION OF PREFERRED EMBODIMENTS

FIG. 1 shows an on-line, liquid flow- through system highlighting the basic elements of a preferred embodiment. The above incorporated references describe in detail a basic apparatus for an industrial NMR system suitable for carrying out the present invention when combined with the added elements described herein. In FIG. 1, a liquid 2 flows through conduits in an industrial system. The liquid is continuously pumped through the sample measurement chamber 4 located between the poles of a jnagnet 6. The sample chamber is surrounded by a temperature controlled air curtain which maintains the sample in the chamber at the same temperature as that of the flowing liquid 2. The air curtain description is found in detail in the above incorporated patent application references. The sample temperature in the chamber must be stable and equal to that of the flowing liquid since the viscosity is highly temperature dependent . In another preferred embodiment the actual temperature value of the flowing liquid may be used to adjust the viscosity measurements. This is accomplished by making calibration measurements of the specific sample material and developing correlation data between the temperature and t a viscosity of that material.

Referring again to FIG. 1, a control valve 10 is opened allowing the sample chamber to be filled by new liquid material. The valve 10 is closed, and the NMR measurement is made. As noted above during this time the temperature of the sample is maintained equal to that of the flowing liquid. The bypass loop ensures that a fresh sample is available at all times. The control valve 12 is typically operated in a complementary fashion with the valve 10; when one valve is opened the other is closed. In other preferred embodiments valve 12 may be always opened or always closed or partially opened.

In FIG. 1, the RF (radio frequency) coils provide the NMR pulse source and pick up. These are described in more detail in the incorporated patent applications above.

FIG. 2 shows the modifications needed for utilizing the present invention with a polymer melt flow. For products that require melting, the flow circuit of FIG. 1 is replaced by a vertically mounted single screw extruder 14 driven by a motor 16. Multi-stage heating units 18, attached to the walls of the extruder, melt the material and the extruder forces the melt into the measurement chamber 20. Solid samples are introduced into the extruder feed hopper 22 where the screw 14 forces the material through the extruder where it is melted. The final stage of the extruder and the sample chamber are thermally regulated at a predetermined temperature value specific for the material. The air curtain provides the sample chamber temperature required. The sample is driven into the chamber by the screw; the screw stops and the sample in the chamber is measured; then the screw drives out the old sample while introducing a new sample.

FIG. 3 shows the CPMG pulse sequence used in this preferred embodiment, and the sequence is P90(x) " ⁽t -Pl80(y) ~ t⁾ _n.

Still referring to FIG. 3, an excitation pulse P90(x) is generated and transmitted to the sample followed by a Pi80 (y) excitation pulse at an interval of t and a series of Pl80(y) pulses at intervals of 2t . The P90(x) and P180 (y) represent pulses with relative phase orientation. Excitation by the Pi80 (y) pulses causes echo responses to occur. The falling portions of the echoes (B in FIG. 3) have the same general pattern as the original FID (A in FIG. 3) excepting that the intensities at the echo maxima are attenuated according to the time constant T2. Measurement of the intensities of the echoes is taken, and a curve plotted of echo intensity as a function of time is made. The curve is decomposed into component curves, and these curves are fitted to the data by an iterative process based upon a Marquardt- Levenberg (M-L) approximation technique applied automatically through a structured realization in software. This technique is used to determine the magnitude of all the parameters, constants, frequencies, etc. which best fit the FID curve. This is an iterative technique where the entire curve is determined at once. The M-L iteration process performs the curve fitting by attempting to minimize the Chi-Squared error function (the sum of the squared differences between the measured data points and the data points from the derived equation) . The results of the M-L approximation are accepted if t^'he Chi Squared error is small enough, if not, the M-L fitting procedure may be reapplied with a different set of starting guesses. If this process also fails, the sample is discarded and a new sample obtained. The M-L technique is documented in the following references: Ind. Appl . Math . , vol. 11, pp. 431-441 by D. . Marquardt, 1963; Data Reduction and Error Analysis for the Physical Sciences (New York, McGraw Hill) , Chapter 11 by Philip R. Bevington, 1969; and The State of the Art in Numerical Analysis (London: Academic Press, David A.H. Jacobs, ed 1977), chapter III.2 by J.E. Dennis. As applied to the measurement regime of interest herein, in a^: preferred embodiment of the present invention, the selected parameters taken from the derived curves are the y-axis intercept ratios, time constants (T*s), frequency terms and other parameters.

Other known in the art iterative techniques which may be applied instead of or with the Marquardt-Levenberg, include: Gauss- Newton and "steepest descent" (found in the above J.E. Dennis reference), Newton-Raphson

(known in the art) , or like techniques, including combinations of these techniques.

Calibration is accomplished by measuring T2 values for a number of samples with known viscosity values. Regression analysis is then performed to derive the coefficients relating the T2 values to the viscosities . For an unknown sample, the T2 values are determined and the viscosity is calculated from the known coefficients.

FIG. 4 shows the echo response train obtained from four motor oil samples with known relative viscosities (commercial grade 30, 40,. 50 and 60 wt motor oils) . FIG. 5 shows the result of a least square exponential curve fit on one oil sample with the resulting T2 value.

FIG. 6 shows the calibration plot of T2 values with the known relative viscosity. The temperature dependence of viscosity is well established, as shown in FIG. 7. This temperature effect, as seen by NMR, is graphed as performed on one of the oil samples . The effect of a decrease in temperature on the echo train response of one oil sample results in a shorter T2 value consistent with the higher viscosity value.

In another preferred embodiment different groupings of the echoes may be used to generate a plot of intensities versus time.

It will now be apparent to those skilled in the art that other embodiments, improvements, details and uses can be made consistent with the letter and spirit of the foregoing disclosure and within the scope of this patent, which is limited only by the following claims, construed in accordance with the patent law, including the doctrine of equivalents . What is claimed is:

Claims

CLAIMS:

1. An on-line NMR system for measuring viscosity of liquids or polymer melts comprising:

(a) means for separating a sample from a flowing stream or a polymer melt, (b) means for positioning and holding said sample in a sample chamber,

(c) means for generating an NMR excitation pulse sequence,

(d) means for transferring said NMR pulse sequence into said sample,

(e) means for receiving spin echo responses to the NMR pulse sequence from said sample,

(f) means for generating viscosity calibration data of known sample by measuring its responses to an NMR excitation .-pulse sequence,

(g) means for determining sample viscosity from said calibration data, and

(h) means for removing the sample .

2. Apparatus as defined in claim 1 wherein said means for generating viscosity calibration data comprises means for calculating T2 values from NMR echo responses of known sample, and means for establishing calibration coefficients by relating the T2 values to the known viscosities; and wherein said means for determining sample viscosity comprises means for applying said calibration coefficients with the T2 measurements for an unknown sample whereby the viscosity of the sample is determined.

3. Apparatus as defined in claim 1 further comprising means for temperature control of said sample during NMR measuring, wherein the sample temperature is stable and at a known temperature compared to the flowing liquid or the polymer melt.

4. Apparatus as defined in claim 3 wherein said known temperature is about equal to the 11

temperature of the flowing liquid or the polymer melt.

5. A process for on-line NMR measuring of viscosity of liquids or polymer melts comprising the steps of:

(a) separating a sample from a flowing stream,

(b) positioning and holding said sample in a sample chamber, (c) generating an excitation NMR pulse sequence,

(d) transferring said NMR pulse sequence into said sample,

(e) receiving spin echo responses to the NMR excitation pulse from said sample,

(f) calculating T2 values from echo responses to NMR echo responses of known samples, and establishing calibration coefficients by relating the T2 values to the known viscosities, (g) determining sample visc'bsity by using said calibration coefficients with the T2 measurements for an unknown sample, and (h) removing the sample.

6. A process as defined in claim 5 further comprising the step of controlling the sample temperature .