This technical note discusses the Expert Advice System which is present in 5.00 and later versions of the Zetasizer Nano software. This provides information on the quality of the data and results obtained from both live and recorded measurements.
Previous versions of the Zetasizer Nano software contain various reports and parameters to help users with the interpretation of the data and results obtained from measurements. These include:
These reports and parameters are available for post measurement analysis i.e. a measurement has to be completed before the quality of the data can be assessed. The Expert Advice system is available for both live and post measurement analysis.
The Expert Advice system is present in the Zetasizer Nano software at various locations.
This is available during a live measurement in the Measurement Display Window (Figure 1). The Expert Tab is a live report of the quality of the data being generated for both a single measurement and a series of repeat measurements.
The advice is available for size, zeta potential and molecular weight measurements.
|
The single measurements use the same criteria used in the quality reports, with the difference being that the comments can be viewed during the measurement. During a series of repeat measurements, the system looks for trends in the results. Comments are given about variability in the results from changes in the sample as a result of, for example, aggregation or dissolution, or where the first measurement is different because the sample has not had time to thermally equilibrate.
This is accessed post measurement by right mouse clicking on selected records, or by selecting View, Expert Advice (Figure 2). This will display the expert advice system's tips dialogue indicting the quality of the measurements selected. Figure 3 shows an example dialogue which could be obtained for size measurements.
|
|
The Expert Advice report is available in the appropriate workspace for post measurement analysis of the data quality (Figure 4). It contains the same information which is displayed in the Expert Advice system discussed in the previous section.
|
For sizing measurements, the Expert Advice system uses the various tests found in the Size Quality Report to test data quality and these are summarized in Table 1.
Test Number | Test Description | Test Details | Warning Message |
|---|---|---|---|
1 | Check appropriate lower size analysis limit used in the distribution analysis | Is the z-average diameter < lower analysis limit? | z-average is smaller than lower size analysis limit |
2 | Check appropriate upper size analysis limit used in the distribution analysis | Is the z-average diameter > upper analysis limit? | z-average is larger than upper size analysis limit |
3 | Check appropriate lower size display limit used | Is the z-average diameter < lower display limit? | z-average is smaller than lower display limit |
4 | Check appropriate upper size display limit used | Is the z-average diameter > upper display limit? | z-average is larger than upper display limit |
5 | Check polydispersity index value | Is the polydispersity index value > 1? | Polydispersity index is very high |
6 | Check correlation function intercept value | Is the correlation function intercept <0.1 or >1.0? | Correlation function intercept out of range |
7 | Check first point selection for multimodal analysis | Is the first correlation point >3 and the z-average diameter <5nm? | Check first correlation point selection for multimodal analysis |
8 | Check the in range value | Is the in range value less than 95% | The in range figure is low |
9a | Check for low mean count rate | Is the mean count rate <20 kcps? | Count rate is out of range |
9b | Check for high mean count rate | Is the mean count rate >1000 kcps? | Count rate is out of range |
10 | Check sufficient data collected | Is the total number of photons collected <10000K? | Insufficient signal collected |
11 | Check the cumulants fit error | Is the cumulants fit error >0.005? | Cumulant fit error high |
12 | Check the distribution fit error | Is the distribution fit error >0.005? | Multimodal fit error high |
13 | Check if the cumulants and distribution fit errors | Is the cumulants fit error >0.005 and the distribution fit error <0.0051? | Sample too polydisperse for cumulant analysis - suggest rely on distribution analysis |
More information on these tests can be found in the technical note discussing the Size Quality Report (MRK693-xx).
During a measurement, if any of the tests fall outside specified limits, warning messages are displayed in the Expert Advice Tab in the Live Measurement Window (Figure 1).
If none of the tests fail, a "Data Meets Quality Criteria" message is displayed in the Expert Advice Tab.
If repeat sizing measurements are performed (3 or more), the Expert Advice system will also look for trends in the results. The tests which are performed for looking at these trends are listed in Table 2.
| Test Number | Test Description | Test Details | Warning Message |
|---|---|---|---|
1 | For dual angle measurements, calculate an aggregation index value and check whether there is multiple scattering present | (1) An aggregation index is calculated from ((ZD (13°)/ZD(173°) - 1) where ZD (13°) and ZD (173°) are the z-average diameters determined at forward and back scatter angles.
| (1) Aggregation index = "X" (see (1) below more detail)
|
2 | Check if the first measurement deviates from the rest of the set | Statistical tests are performed on the mean obtained from (1) all of the records and (2) with the first record omitted to determine whether the first record deviates significantly from the rest | First measurement deviates from rest of set - exclude it or repeat measurement |
3 | Check if the mean size is trending from repeat measurements | A model is fitted to the z-average diameters of the records to determine whether there is a statistically significant trend present | Mean size is trending by (+/-) "Y"%(see (2) below more detail) |
4 | Check if the count rate and size are both decreasing from repeat measurements | A model is fitted to the mean count rates and z-average diameters of the records to determine whether there is a statistically significant decreasing trend present | Count rate and size are both decreasing - sample may be sedimenting or dissolving |
5 | Check if the count rate and size are both increasing from repeat measurements | A model is fitted to the mean count rates and z-average diameters of the records to determine whether there is a statistically significant increasing trend present | Count rate and size are both increasing - sample may be aggregating |
6 | Check if the count rate varies significantly from repeat measurements | A model is fitted to the mean count rates of the records to determine whether there is a statistically significant variation present | Count rate varies significantly - sample not stable or not clean |
(1) where "X" = the aggregation index value determined from dual angle measurements
(2) where the direction (+/-) and "Y" percentage of any trend in the mean size is calculated and displayed in the warning message
For zeta potential measurements, the Expert Advice system uses the various tests found in the Zeta Potential Quality Report to test data quality and these are summarized in Table 3.
| Test Number | Test Description | Test Details | Warning Message |
|---|---|---|---|
1 | Check the intensity of the reference beam | Is the reference beam intensity < 500kcps | Reference beam count rate too low. If you see this warning for a number of zeta potential results then please contact your nearest Malvern representative |
2 | Check the quality of the phase plot data | Is zeta quality factor < 1.0? | Phase data poor - signal to noise ratio low - increase number of sub runs per measurements |
3 | Check the quality of the distribution plot data | Is the quality factor obtained from the Slow Field Reversal measurement < 1.0? | Distribution data poor - increase number of sub runs per measurements |
4 | Check the limits of the zeta potential distribution are within the analyzed range | Does the zeta potential spectrum fit in to the selected zeta range? | Increase zeta potential analysis range |
5 | Check the conductivity and appropriate measurement protocol has been used | Is the conductivity >10mS/cm and has General Purpose been used? | Suggest use of Monomodal analysis as conductivity is >10mS/cm |
6 | Check for flare originating from the cell wall | Is spectral quality factor >0.0 and <0.9? | Flare from cell wall - check for bubbles, increase sample concentration |
7 | Check attenuator position used | Is attenuator position selected 11? | Sample concentration may be inappropriate (too high or low) |
If repeat zeta potential measurements are performed (3 or more), the Expert Advice system will also look for trends in the results. The tests which are performed for looking at these trends are summarized in Table 4. These tests are similar to those employed for sizing measurements.
| Test Number | Test Description | Test Details | Warning Message |
|---|---|---|---|
1 | Check if the first measurement deviates from the rest of the set | Statistical tests are performed on the mean obtained from (1) all of the records and (2) with the first record omitted to determine whether the first record deviates significantly from the rest | First measurement deviates from rest of set - exclude it or repeat measurement |
2 | Check if the mean zeta potential is trending from repeat measurements | A model is fitted to the mean zeta potential values of the records to determine whether there is a statistically significant trend present | Mean zeta potential is trending by (+/-) "Y"%(see (1) below for more details) |
3 | Check if the count rate and zeta potential are both decreasing from repeat measurements | A model is fitted to the mean count rates and zeta potential values of the records to determine whether there is a statistically significant decreasing trend present | Count rate and zeta potential are both decreasing - sample may be sedimenting or dissolving |
4 | Check if the count rate and zeta potential are both increasing from repeat measurements | A model is fitted to the mean count rates and zeta potential values of the records to determine whether there is a statistically significant increasing trend present | Count rate and zeta potential are both increasing - sample may be aggregating |
5 | Check if the count rate varies significantly from repeat measurements | A model is fitted to the mean count rates of the records to determine whether there is a statistically significant variation present | Count rate varies significantly - sample not stable or not clean |
The Expert Advice system for molecular weight measurements only analyses the data from single measurements i.e. no information on trending from repeat measurements is provided. The tests used for molecular weight measurements are listed in table 5.
| Test Number | Test Description | Test Details | Warning Message |
|---|---|---|---|
1 | Check that the scattering from the various sample concentrations are appropriate | Is the count rate obtained from the various sample concentrations at least 10% more than the scattering obtained from the solvent? | The "A" g/ml concentration is producing less excess scattering ("B" kcps) than is advised. Ideally the scattering should be more than "C" kcps. Use a higher concentration of sample (see (1) below for more details) |
2 | Check that the scattering from the various sample concentrations are not too high | Is the count rate obtained from the various sample concentrations greater than 1000kcps? | The "D" g/ml concentration is producing more excess scattering ("E" kcps) than is advised. Ideally the scattering should be less than 1000 kcps. Use a lower concentration of sample (see (2) below for more details) |
3 | Check the value of the correlation coefficient (R2) | Is the correlation coefficient (R2) < 0.9 | Low correlation coefficient ("F" < 0.9). Either remove outlying points or re-measure the samples at those concentrations. Use the Debye plot to identify outlying points (see (3) below for more details) |
(1) where "A" is the concentration of the sample and "B" is the mean count rate in kcps of that sample and "C" is the minimum ideal scattering in kpcs calculated from ((Scattering of solvent) + 10%)
(2) where "D" is the concentration of the sample and "E" is the mean count rate in kcps of that sample
(3) where "F" is the correlation coefficient (R2) calculated for the selected record
Each of these tests is now looked at in more detail.
A common mistake that users make during molecular weight measurements is to prepare samples for which are too low in concentration. In these cases, the excess scattering produced by the analyte above that of the solvent will be too low to obtain good quality data. This test is designed to highlight inappropriate scattering levels for the sample concentrations prepared.
During a molecular weight measurement on the Zetasizer Nano, the mean count rate from the solvent is measured. The mean count rates obtained for the various sample concentrations (i.e. analyte dissolved in the solvent) should be greater than that for the solvent i.e. there is excess scattered light produced. A reasonable minimum excess level of scattering is considered to be 10%. This test therefore ensures that the mean count rates for each sample concentration are at least 10% more than the mean count rate of the solvent.
The avalanche photodiode detector (APD) used in the Zetasizer Nano series has an operating range which is highly linear up to 1000kcps. This is much more important for total intensity, i.e. molecular weight measurements than size measurements. Therefore, test 2 highlights any sample concentrations which produce a mean count rate of greater than 1000kcps and advises that such samples are diluted before being re-measured.
R squared (R2), also known as the correlation coefficient or Pearson's coefficient, is a calculated parameter that is used to describe how well graphical data fit an applied model. Values for R2 range from 0 to 1, with a value of 1 indicating that there is zero deviation of the data from the model or best fit line (or curve).
R2 is calculated for the Debye plot and a value less than 0.9 is considered to be a poor fit. The advice given is to either remove outlying points or re-measure the samples at those concentrations. Any obvious outliers in the Debye plot can be identified in the Molecular Weight report and can be removed by editing the record.
