Trace Level Total Sulfur determination according to ASTM D5453

A comprehensive study on the 0.1 to 1 mg/kg range

In recent developments, the need for trace level Sulfur analysis by elemental combustion analyzers is growing. The method appears to be being used more in several Hydrocarbon products such as Benzene, Toluene, or recently developed renewable fuels like VGO or HEFA diesel(blends). The current ASTM D5453 however, in its current form, doesn’t allow reporting Total Sulfur below 1.0 mg/kg and misses directions on how to analyze at trace levels down to 0.1 mg/kg.

This article describes an extensive test setup to venture the requirements for our analytical combustion equipment to reach low-level quantification levels in the range of 0.1 to 10 mg/kg for Total Sulfur according to the ASTM D5453 method.

Detection of Total Sulfur – UV-Fluorescence

When organic Sulfur components are combusted in the high-temperature furnace, the following reaction takes place:

R – S + O₂ → SO₂ + H₂O + CO₂

After complete sample oxidation, the combustion gas is conditioned by removing water vapor and particles. The conditioned gas stream containing the Sulfur Dioxide (SO₂) molecules is transferred to the reaction chamber. The Xenon flashed UV lamp will excite the Sulfur Dioxide molecules to SO₂^* at specific wavelengths and due to the unstable character of SO₂^* it will relax back to SO₂ instantly. The released energy will be emitted in the form of light and will be detected by the Photomultiplier Tube (PMT). The amount of light emitted equals the total amount of SO₂ (Total Sulfur) present in the sample.

Detection:

SO₂ + hv₁ → SO₂*

SO₂* → SO₂ + hv₂

Trace level study setup

For this study, a variety of different configurations was used. All configurations will run the same standards and samples. The following analyzers have been used to perform the analysis of TS and TN at trace level:

Xplorer TN/TS – Horizontal furnace setup

 Sample introduction:

• Boat introduction
• Liquid module: direct injection

Autosamplers:  

• Archie
• XLS-30

Xplorer-V TN/TS – Vertical furnace setup

 Sample introduction:

• Xpro-V Liquid inlet: direct injection

Autosamplers:

• Integrated LS-26

Experimental setup and system parameters

For the horizontal sample introduction, all samples and standards were introduced by either direct liquid injection (Archie and XLS-30) or boat introduction (Archie). For the vertical injections, the integrated liquid autosampler (LS-26) of the Xplorer-V was utilized.

For calibration of the analyzers, two different solvents were used based on Xylene or Iso-Octane, to verify the applicability of these solvents. After each calibration, a series of different matrices were injected to check the variation in injection volumes and solvent variations on the recovery. The following table shows the variation of the several injection configurations tested:

Table 1: Overview of variation of different injection configurations and samples

System parameters applied:

The following method parameters were applied to the variations of calibration lines and sample injections:

Table 2: Overview of method parameters per configuration

Evaluation of results:

Depending on the injection volume and injection speed, for each configuration, the analyzer was calibrated with a set of Xylene standards and a set of Iso-octane standards from 0.1 mg/kg to 10 mg/kg. Resulting in a total of 21 calibration lines. Evaluation of the calibration lines and the sample results were based on:

• The coefficient for determination of the calibration lines (how well a fit is applicable).
• The RSD values on the lowest injection point of the calibration curve at 0.1 mg/kg.
• Recovery of the sample CRM.
• The RSD values of the sample matrices.

Calibration

Standards used for calibration were made from Dibutyl Sulfide (S) in Iso-Octane or Xylene. The Xplorer analyzers are calibrated in the range of 0,1-10 mg/kg. All calibration points have been corrected for the average blank area count. After running all the individual calibration lines and corresponding samples the following overview was extracted:

Table 3: Overview of various calibration lines for TS 0,1 – 10 mg/kg

In order to evaluate the results from the different configurations on their performance in the trace range, the results were rewarded with a point-based system. Points were rewarded based on a set of criteria. The more points that were earned the better the performance was in the trace range. Each independent end result on variations was rewarded based on the following criteria:

• R² of designated calibration line >0.999 = 1 point
• RSD on 0.1 mg/ kg calibration injection (n=5) <5.0 = 1 point
• Recovery of the sample CRM between 95% and 105% = 1 point per sample matrix
  • Gasoil CRM limits: Expected 4,80 mg/kg (limits are 4,56 to 5,04 mg/kg)
  • Naphtha CRM limits: Expected 0.53 mg/kg (limits are 0.50 to 56 mg/kg)
  • Biodiesel CRM limits: Expected 0.25 mg/kg (limits are 0.24 to 0.26 mg/kg)
  • Gasoline CRM limits: Expected 5.20 mg/kg (limits are 4.94 to 5.46 mg/kg)
• RSD on the end value of sample matrix <5.0 = 1 point per sample matrix

A maximum of 10 points could be earned in total. A score of 8 points was set as the minimum to apply for applicability for trace level determination of Total Sulfur at levels ranging from 0.1 – 1.0 mg/kg. The complete overview of all the results including the total earned points can be viewed in the table results.

Results

In table 4 an overview of the results can be found, this table shows per analyzer depending on both the sample introduction type and injection volume the results of the analyzed CRM samples using a calibration line made in Iso-octane or Xylene matrix. The results are rounded according to ASTM D5453 including the RSD value measured over five injections. The last column shows the sum of total points earned for that individual configuration.

Table 4: Sample matrix injections and final results overview

Summary of the application

 Sample volume:

Looking at the complete overview there is a clear downtrend in the variation of RSD results noticeable with the increase of the sample volume. This applies to both sample injections and the low-level calibration level of 0.1 mg/kg for the Total Sulfur standard.

Iso-Octane VS Xylene:

The effect of using Iso-Octane or Xylene as solvent results in the following noticed effects:

• Injections of Xylene-based standards give a more stable baseline view and show higher response factors at the lower level concentration injections compared to standards that are made in Iso-Octane solvent.
• Using Iso-Octane as a solvent basis for the calibration line standards results in peak splitting when the boat application is used and higher concentration levels are injected.
• Xylene gives a higher blank injection for Total Sulfur compared to Iso-Octane but still within acceptable ranges.
• Iso-Octane gives far higher Total Nitrogen blanks, this is not lined out in the scope of this application note but could contribute when Total N is also valuable.

Most effects mentioned above are probably caused due to the fact that Iso-Octane contains far more CH₃ molecules compared to Xylene, which can form more H₂O during combustion and an excess of O_2. The lower boiling point of Iso-Octane results in the peak splitting with boat application which makes it harder to integrate.

Boat introduction VS direct liquid injection:

Both a direct liquid injection and boat introduction setup are applicable for trace level determinations. The usage of a boat introduction reaches acceptable results but results in longer analysis times and lower response factors. The best results are achieved by injection of high sample volumes, the boat application is limited to sample size injections of a max 30 mL.