1 Introduction Aqueous based engine coolants have long been used due to the ability of water to efficiently conduct heat away from engines. However due to the relatively high freezing point of water and the extremely cold conditions engines sometimes operate in, chemical modification is necessary to ensure the coolant doesn’t freeze when exposed to low temperatures. When mixed with water, ethylene glycol and propylene glycol can alter both the freezing point and the boiling point of a coolant relative to the proportion of glycol in the mixture. [1] Hence the term ‘antifreeze’ commonly used to refer to glycol based engine coolant. ASTM standard specifications for engine coolants prescribe that a glycol concentration range of between 40 - 60 %v/ v in water of a suitable quality, will function effictively during both winter and summer to provide protection against corrosion, cavitation, freezing and boiling. [2,3] The glycol concentration of a coolant indicates the effective temperature range that a coolant will be suitable to operate in and can thus confirm whether the coolant being tested is fit for a given operating environment. Should the glycol content have been depleted during a coolant’s lifetime or if it becomes diluted by an increase in water content, the coolant can suffer from an increased freeze point or even cause increased corrosion and cavitation. Should the glycol levels be too high due to too much coolant concentrate being added, the coolant will suffer from a reduced ability to transfer heat. Commonly, glycol content is determined manually using a refractometer (ASTM D3321 [4]), either in the field or in a laboratory environment. Testing by GC allows for large sample sets to be analysed sequentially using an autosampler without the need for constant analyst attention. In high throughput testing labs, simple sample preparation procedures with the potential for automation are highly favorable. However, given the high concentrations of water and glycol in engine coolant, the sample can be quite harmful to the lifetime of the GC injection liner and column. As such two sample preparations are investigated in this paper, a 10 - fold aqueous sample dilution and a 100-fold dilution in isopropanol. The columns used for both sample preparation methods shared a common stationary phase but differed in their internal dimensions to best optimize the method for the concentration of glycol injected. 2 Methods and Materials 2.1 Instrumental A PerkinElmer Clarus 690 Gas Chromatograph was used for the analysis. The GC was configured with a liquid autosampler, programmable split/ splitless injector (PSSI), flame ionization detector (FID) and an Elite WAX ETR analytical column. Two methods were evaluated to compare the longevity of the column used. Method 1 consisted of a 10-fold dilution in deionised water Research 19 Tribologie + Schmierungstechnik · 70. Jahrgang · eOnly Sonderausgabe 1/ 2023 DOI 10.24053/ TuS-2023-0029 Rapid identification and quantification of ethylene and propylene glycol in engine coolant by gas chromatography Nicholas Lancaster, Keith Cory Schomburg* Determining the glycol type and concentration of an engine coolant is a useful analytical test for in-service testing laboratories to both identify the coolant type and determine whether the coolant is fit for its intended purpose. Too much glycol can result in poor heat transfer within the engine whereas too little glycol can increase the freeze point of the coolant. As such repeat testing over the lifespan of the coolant is required to avoid potential problems and ensure the overall health of the cooling system. This paper details a GC method to rapidly separate ethylene and propylene glycol in under five minutes and determine the concentration of these glycols in percentage. Two preparation methods are detailed to demonstrate a simple ‘dilute & shoot’ preparation by diluting in deionised water (10-fold dilution) or isopropanol (100-fold dilution). Keywords Glycol, Antifreeze, Engine Coolant, Condition Monitoring Abstract * Nicholas Lancaster PerkinElmer, Seer Green, United Kingdom, HP9 2FX Keith Cory Schomburg, Ph. D. PerkinElmer, Downers Grove, IL, USA, 60515 signs of column overload. For method 2 (100-fold dilution in isopropanol), a narrower bore ID and thinner df more suitable to the calibration range of 0.1 - 0.9 % ethylene and propylene glycol was selected. For both methods a 5 m long deactivated silica guard column was used with the same internal dimensions as the analytical column. 2.2 Experimental The analysis was designed to separate and quantify the glycols in the shortest possible time to allow increased sample throughput. An isothermal oven program was chosen to avoid oven cooldown times in between sample measurements. A high split ratio was necessary to deal with the large volume of water vapour formed in the injector due to the aqueous matrix of the samples. The high split ratio also helps to prevent backflash and split peaks which are common issues associated with aqueous injections. The high split ratio also contributes to improved peak shape and helps prevent column overload which is vital when measuring such high analyte concentrations. The GC parameters for both methods are shown in table 1. Both Helium and Hydrogen were tested for suitability with both showing similar results in peak retention and resolution. Hydrogen was used as carrier for the results represented in this paper. Before analysis each day the column was conditioned for an hour at 230 °C. 2.3 Standard Preparation Combined ethylene glycol and propylene glycol standards were made up in deionized water (Method 1, 10fold dilution) and Isopropanol (Method 2, 100-fold dilution). The calibration ranges for each method are show in tables 2 and 3. Research 20 Tribologie + Schmierungstechnik · 70. Jahrgang · eOnly Sonderausgabe 1/ 2023 DOI 10.24053/ TuS-2023-0029 and wide bore WAX column. Method 2 consisted of a 100-fold dilution in isopropanol and a narrower bore WAX column. The Elite WAX ETR column makes use of a cross bound Carbowax Polyethylene Glycol (PEG) stationary phase. The bonding mechanism of this stationary phases produces a very stable polar retention. [4] The bonding mechanism is also responsible for the column’s ability to withstand repeat water injections which is vital given the aqueous matrix of the coolant sample. The dimensions were chosen to be suitable for the concentration of glycol injected. For method 1 (10-fold aqueous dilution), the calibration ranged from 3 - 7 % of ethylene and propylene glycol. The wide bore internal diameter (ID) of 0.53 mm and dispersion film thickness (df) of 1 µm provides the most capacity for the analyte before showing GC Parameters Instrument PerkinElmer Clarus 690 Column (Method 1) Elite WAX ETR 30 m x 0.53 mm x 1 µm Column (Method 2) Elite WAX ETR 3 0 m x 0.32 mm x 0.5 µm GC Oven Parameters (Method 1) Isothermal at 150 °C for 5 minutes. GC Oven Parameters (Method 2) Isothermal at 130 °C for 5 minutes. Liquid Autosampler Parameters Syringe Size (Method 1) 5 µL Injection Vol. (Method 1) 0.5 µL Syringe Size (Method 2) 0.5 µL Injection Vol. (Method 2) 0.3 µL Injection Speed Normal Number of Plunges 6 Sample Washes 2 Pre-Injection solvent wash 1 Post-injection solvent wash 3 Wash Solvent (Method 1) Methanol: Water (50: 50) Wash Solvent (Method 2) Isopropanol Viscosity Delay 2 Injector Parameters Type Programmable Split/ Splitless Temperature 250 °C Carrier/ mode Helium/ Hydrogen flow Flow rate (mL/ min) 6 mL/ min Split Ratio 50: 1 Liner Deactivated glass liner with deactivated glass wool Detector Parameters Type FID Temperature 250°C Hydrogen 45 mL/ min Air 450 mL/ min Attenuation -3 Table 1: GC Method Parameters Table 2: Calibration range, Method 1 Table 3: Calibration range, Method 2 Ethylene and Propylene Glycol concentration (% v/ v) Standard 1 3 Standard 2 4 Standard 3 5 Standard 4 6 Standard 5 7 Ethylene and Propylene Glycol concentration (% v/ v) Standard 1 0.1 Standard 2 0.3 Standard 3 0.4 Standard 4 0.5 Standard 5 0.6 Standard 6 0.9 2.4 Sample Preparation Method 1 involves a 10-fold dilution in deionised water. 1 mL of sample is added to 5 mL of deionised water before being made up to 10 mL total volume. Method 2 involves a 100-fold dilution in isopropanol. 100 µL of sample is added to 5 mL of isopropanol before being made up to 10 mL total volume. To determine the accuracy of the method a set of ten samples with ethylene glycol concentrations determined by refractometer were used as reference. 3 Results and Discussion 3.1 Method 1: 10-fold aqueous dilution Baseline resolution was acheived in under 4 minutes with propylene glycol eluting first at 2.88 minutes and ethylene glycol eluting second at 3.19 minutes. Retention times were confirmed by injecting individual glycol standards under the same conditions. Figure 1 shows an example chromatogram of the 7 % calibration standard. Figures 2 and 3 show the linear calibration curves for the two glycols under the method 1 analysis conditions. Both curves have a coefficient of determination (R 2 ) greater than 0.99 indicating an good level of data correlation within the specified calibration range. Ten ethylene glycolbased reference sam- Research 21 Tribologie + Schmierungstechnik · 70. Jahrgang · eOnly Sonderausgabe 1/ 2023 DOI 10.24053/ TuS-2023-0029 Figure 4: Method 1 - example sample chromatogram y = 21275x + 751 R² = 0.994 61000 81000 101000 121000 141000 161000 3 4 5 6 7 Peak Area Concentration (%v/ v) Propylene Glycol Figure 2: Method 1 - propylene glycol calibration curve y = 18039x + 452.6 R² = 0.995 52000 62000 72000 82000 92000 102000 112000 122000 132000 3 4 5 6 7 8 Peak Area Concentration (%v/ v) Ethylene Glycol ples with known glycol concentrations (determined by refractometer reference method, ASTM D3321) were measured on the GC after method calibration. Figure 4 shows an example chromatogram clearly confirming that the sample is an ethylene glycol-based coolant with no propylene glycol present. Figure 3: Method 1 - ethylene glycol calibration curve Figure 1: Method 1 - example chromatogram (7 % propylene and ethylene glycol) 3.2 Method 2: 100-fold dilution in Isopropanol Figure 5 shows an example chromatogram using method 2. Baseline resolution is achieved for method two in under three minutes with propylene glycol eluting first at 1.99 minutes and ethylene glycol eluting second at 2.3 minutes. To further optimize method 2 an isothermal oven temperature of 130 °C was used due to the decreased retention on this column. A smaller injection volume of 0.3 µL was also used to compensate for the decreased sample capacity on the narrower bore column used for method 2. Figures 6 and 7 show the linear calibration curves of propylene and ethylene glycol collected using GC method 2. The calibration range was modified for additional accuracy given the increased dilution for method 2. The R 2 value for both analytes is greater the 0.999 indicating excellent correlation within the calibration range. The same sample set evaluated with method 1 was used with method 2 to determine the accuracy of the method. Figure 8 is an example sample chromatogram from method 2 confirming that the sample measured was an ethylene glycol-based coolant. Research 22 Tribologie + Schmierungstechnik · 70. Jahrgang · eOnly Sonderausgabe 1/ 2023 DOI 10.24053/ TuS-2023-0029 Table 4 shows the results for these coolant samples collected using GC method 1 with the results from the GC in the %v/ v column and the results from the reference method in the ‘Reference (%v/ v)’ column. When compared, the results collected using the two methods agreed well showing a high level of accuracy from the GC method. Table 4: Method 1 - reference sample results Propylene Glycol Ethylene Glycol Unknown Sample RT(min) %v/ v RT(min) %v/ v Reference (%v/ v) 1 2.88 ND 3.19 32.8 35 2 2.88 ND 3.19 30.1 32 3 2.88 ND 3.19 38.3 40 4 2.88 ND 3.19 46.4 45 5 2.88 ND 3.19 9.1 3 6 2.88 ND 3.19 50.4 49 7 2.88 ND 3.19 43.5 44 8 2.88 ND 3.18 40.1 42 9 2.88 ND 3.19 54.4 52 10 2.88 ND 3.19 52.4 49 Figure 6: Method 2 - propylene glycol calibration curve y = 19883x - 185.62 R² = 0.999 1700 3700 5700 7700 9700 11700 13700 15700 17700 19700 0.1 0.3 0.5 0.7 0.9 Peak Area Concentration (%v/ v) Ethylene glycol Figure 7: Method 2 - ethylene glycol calibration curve y = 24547x - 259.87 R² = 0.999 2100 7100 12100 17100 22100 27100 0.1 0.3 0.5 0.7 0.9 Peak Area Concentration (%v/ v) Propylene glycol Figure 5: Method 2 - example chromatogram (0.9 % ethylene and propylene glycol) The results of the reference sample set measured using method 2 yielded a similar level of agreement between the calculated values and reference values for the samples further indicating a high level of accuracy of the GC methods. 3.3 Method Robustness Over the course of several weeks, 500 samples were run using both methods to determine the robustness of the methods and the resilience of the column to repeat sample injections. The size of the sample set varied between 50 and 100 samples per day and the column was conditioned before measuring each set (230 °C for one hour). After the final sample set was tested a set of 10 control standards were run to evaluate the repeatability of the results. For method 1 the control standard was made up to 5 %vol/ vol of both propylene and ethylene glycol. For method 2 the standard was 0.5 %vol/ vol in isopropanol. Figure 9 shows an overlay of the ten standard chromatograms measured using method 1. Table 6 shows the retention times and calculated concentrations for both glycols after taking the dilution factor into account for each control standard (method 1). The percentage relative standard deviation (%RSD) for both propylene and ethylene glycol are below 2.5 % indicating a good degree of repeatability despite the concentration of the standards and the large volume of aqueous sample measurements that preceded the repeatability test. Retention times for both glycols showed excellent consistency. Research 23 Tribologie + Schmierungstechnik · 70. Jahrgang · eOnly Sonderausgabe 1/ 2023 DOI 10.24053/ TuS-2023-0029 Figure 8: Method 2 - example sample chromatogram Figure 9: Method 1 - overlaid chromatograms of ten standard replicates (5 %) after 500 aqueous injections Propylene Glycol Ethylene Glycol Unknown Sample RT(min) %v/ v RT(min) %v/ v Reference (%v/ v) 1 1.99 ND 2.30 35.2 35 2 1.99 ND 2.30 32.1 32 3 1.99 ND 2.30 38.8 40 4 1.99 ND 2.30 47.0 45 5 1.99 ND 2.30 11.0 3 6 1.99 ND 2.30 51.9 49 7 1.99 ND 2.30 44.7 44 8 1.99 ND 2.29 41.7 42 9 1.99 ND 2.29 55.9 52 10 1.99 ND 2.30 51.1 49 Table 5: Method 2 - reference sample results col by GC. The carbowax stationary phase in combination with the split injection method allow for repeat analysis of aqueous coolant samples with good accuracy and precision despite the high glycol concentration of the anlaytes present in both the samples and the standards. The repeatability of the methods tested after the injection of 500 samples display very consistent results. However the increased tailing of the glycol peaks after approximately 700 samples indicate that the methods will likely not be suitable where very high sample throughput is a requirement. Further testing is required to determine a more suitable stationary phase where sample volumes in excess of 100 per day are routine. References [1] J.D. Laukkonen, The history of antifreeze, Crankshift, March 1, 2017. http: / / www.crankshift.com/ history-of-antifreeze/ [2 ] ASTM D3306-21, “Standard Specification for Glycol Base Engine Coolant for Automobile and Light-Duty Service,” ASTM International, West Conshohocken, PA, 2021, DOI: 10.1520/ D3306-21, www.astm.org [3] ASTM D6210-2017, “Standard Specification for Fully- Formulated Glycol Base Engine Coolant for Heavy Duty Engines,” ASTM International, West Conshohocken, PA, 2017, DOI: 10.1520/ D6210-17, www.astm.org [4] ASTM D3321-19, “Standard Test Method for Use of the Refractometer for Field Test Determination of the Freezing Point of Aqueous Engine Coolants,” ASTM International, West Conshohocken, PA, 2019, DOI: 10.1520/ D3321-19, www.astm.org [5] https: / / www.perkinelmer.com/ uk/ product/ pe-wax-etr-30m-53-mm-1-0-m-n9316569 Research 24 Tribologie + Schmierungstechnik · 70. Jahrgang · eOnly Sonderausgabe 1/ 2023 DOI 10.24053/ TuS-2023-0029 The results from method 2 were comparable to those measured using method 1. The methods were both used to measure more samples to determine whether the methods would be suitable for large sample volumes. After approximately 700 samples both methods began to show severe peak tailing that could not be improved without both exchanging the guard column and trimming the inlet of the analytical column. 4 Conclusion This paper presents two simple and rapid methods for separating and quantifying ethylene and propylene gly- Propylene Glycol Ethylene Glycol RT %v/ v RT %v/ v 5% std 1 2.88 47.9 3.22 47.7 5% std 2 2.88 49.0 3.22 48.8 5% std 3 2.88 48.8 3.22 48.7 5% std 4 2.88 48.8 3.22 48.5 5% std 5 2.88 48.8 3.22 48.6 5% std 6 2.88 48.8 3.22 48.6 5% std 7 2.88 48.9 3.22 48.7 5% std 8 2.88 52.4 3.22 52.1 5% std 9 2.88 49.0 3.22 48.7 5% std 10 2.88 49.0 3.22 48.8 %RSD 0.00 2.42 0.00 2.38 Table 6: Method 1 - replicate analysis of ten standards (5 %) after 500 aqueous injections