Day 14 - Challenging times

I was at CERES from 8:15AM to 4PM; traffic got us off to a later start than usual.

I immediately cooled down the instrument in order to perform the previously discussed injection port maintenance and took steps to equilabrate the instrument. However, equilibration was not able to be achieved. The detector signal was noisy (~ +/- 2-8 Hz), it wandered (~+/- 300 Hz), and displayed random spikes (~ >100Hz). Inspection of chromatograms from last night's run revealed the same pattern. Anna and I checked the system for leaks, verified that gases were flowing, replaced injection port O-rings and septa, and tightened column nuts - but with no improvement. Interestingly, symptoms were amplified whenever the column oven was heated to operating temperature and were reduced when at room temperature. A temporal look back through chromatographic data generated over the course of my time at CERES revealed that these symptoms have been present all along, but have steadily escalated to the current state. Consultation with Glenn Wilson (senior research faculty - OSU) and Kim Anderson (OSU) suggested that these symptoms were different from typical instrument issues that OSU staff have experienced. Just as an example, OSU's GC-ECD resting baseline fluctuates +/- 0.25-0.5 Hz. We suspect that there might be something going on in/around the portion of the instrument where the columns connect to the detector. I will evaluate this portion of the instrument on Monday. Unfortunately, last night's run will not be useable for quantitation due to the chromatographic blemishes described above and it's not clear how long it will take to right the instrument. 

On a high note, Anna completed instrumental analysis of the first batch of PSDs, all required bench sheets, and saved the data in Excel files. YES! I'm also happy to report that her confidence improved substantially over the course of the analysis. Furthermore, Adama and Marie completed the last batch of PSD extractions (acetone arrived Thursday evening); all samples have now been extracted. Similarly to Anna, their attention to detail regarding thorough completion of bench sheets has improved markedly compared to the field sampling bench sheets that were received by OSU.

Day 13 - Sample quantitation, visitors, and swarms of flys

Today I arrived at CERES at 7:45 and left at 6:30.

Anna and I immediately began quntitating samples. Additionally, while she was quantitating samples, I changed out the injection port liners. However, one of the glass liners broke in the back detector injection port and I had to take some extra time righting the situation. However, just as Anna and I were getting going, Markfousse and consultants from the UN showed up for interviews. In total, their visit lasted for ~3 hours and brought sample analysis to a halt. Specifics of this meeting can be discussed in another forum if desired.

Anna and I got back to sample analysis immediately after the UN folks departed CERES. However, I found that the CSV macro writer that Ted Haigh (OSU) sent me for data compilation was not working correctly; it only imported retention time and area information for each sample – it did not import quantitated values!!! Upon this discovery, I had Anna bring up the previously saved window snapshots for each of the samples we had already quantitated and had her save results to Microsoft Excel spreadsheets. This is the best I have at the moment. Argggg AGILENT!!!!!!!     

While Anna was creating Excel files, I analyzed a CCV to verify that the instrument was up and operational… it was not. Peaks on the front column were jagged and noisy. I tightened the injection port retaining nut, saw an improvement in the resting baseline shot a CCV, and found that chromatography improved. Anna then set up the next batch of PSD extracts to run on the instrument. While these were running, she finished analyzing the first batch of samples… All said, Anna evaluated 5 samples today which is a marked improvement over yesterday. The largest gains were made early in the morning and at the end of the day when most CERES staff were absent. I see her progress today as a success and will encourage this type of effort tomorrow.  

Interestingly, the last few storms combined with steady high temperatures have served the CERES fly population well - they were everywhere today; in the lab, on the staff, on the instruments, in the fume hoods, on pipettes, on glassware.... literally everywhere. This was a new challenge in patience for me and a challenge that CERES staff took in stride. 

Day 12

We arrived at the lab at 7:45 and left a 6PM. 

It was found that internal standard area counts for the closing CCV from last night’s run were greater than 25% of expected, but just barely. Evaluation of sample internal standard area counts revealed that area counts all met data quality objectives. Additionally, 8 compounds were above the +/- 25% expected recovery data quality objective. However, Kim Anderson (OSU) and I agreed that due to devotions from OSU’s Standard Operating Procedures (i.e. gas purity and detector deterioration) that it would be appropriate to widen the data quality objectives. At +/- 35% of expected, >90% of compounds in the CCV were within data quality objectives. We will proceed in this fashion. Thus, the first analytical batch of samples met project data quality objectives and are ready to be quantified for pesticides. Anna worked through one sample today due to many distractions (~ a rate of 1 sample/hr). 

In order to conserve the 100ppb calibration standard, which I used as a continuing calibration verification in last night’s run, I had Anna prepare 10mL of a 100ppb CCV for use in the analysis. I then compared the GC-ECD responses of the newly prepared CCV solution to the previously used 100ppb calibration standard and found that they were nearly identical except that the internal standard area counts were a bit higher in the news solution… near the ranges of the +20% expected level determined from calibration standards. Therefore,  I will open the internal standard area count data quality objectives to +/-50% as per OSU’s SOP. 

Anna and I also fussed with getting the instrument back in running condition after last night’s sequence; mainly trying to improve responses for a handful of pesticides. We tried a 1hr and a 2 hr instrument bake-out but saw minimal improvement. Discussion with Ted Haigh (OSU) revealed that he replaced liners after each analytical batch and that seemed to work perfectly. I will adopt this method. However, I will wash liners with acetone and hexane and reuse them because CERES does not have any 4mm injection port liners – the ones needed for this method!    

Some specific points:

1. We found that isodrin co-elutes with PCB100 on the front column… thus it cannot be quantified by CERES.
2. Due to myriad distractions, it took Anna much more time than necessary to analyze a sample.
3. The myriad distractions led to Anna making errors filling out chromatographic bench sheets… even while I was sitting with her.
4. I found that Marie and Adama did not truly understand how to report some of the information needed for sample extraction bench sheets. Specifically, they copied the sample extraction batch number verbatim from a bench sheet that I had helped prepare onto the bench sheet of a brand new batch of samples prepared on a completely different day. I explained this to Marie and had her explain back to me the definition of an extraction batch and how to appropriately label bench sheets.
5. I still find myself reminding CERES staff that verifying pipettes before use is absolutely necessary for this project and any other project that they are involved in.

Day 11 - GC-ECD and more

I arrived at the lab at 7:45AM and left at 5:30PM. My day was spent entirely on the GC-ECD with Anna, while Adama and Marie worked on completing the 5^th batch of PSD extractions. Anna and I discussed how to run samples on the instrument and how to compose sequences. We agreed to run analytical sequences as : instrument blank (IB), continuing calibration verification (CCV), 8-10 samples, closing CCV, and IB; the 8-10 samples include field, trip and reagent blanks. I also introduced Anna to the idea of real-time Quality Control (QC) and got her acquainted with how to fill out and organize chromatography bench sheets. I had her calculate all QC limits (i.e. internal standard area counts, surrogate standard recoveries, and CCV passing limits). 

Unexpectedly, I spent a lot of time instructing Anna on how to integrate peaks, where to set integration baselines, and how to actually interpret chromatographic data. I suggested a workflow similar to OSU chemists, she adopted it, and it seems to be working well so far. To this end, I had her explain these concepts back to me as a demonstration of knowledge retention. Anna and I were able to complete the analysis of an IB, CCV, and 3 types of QC blanks over the course of the day. 

At least one, and no more than four, positive hits were identified in all reagent, trip, and field blanks with levels generally being less than 40 ppb. Overlaying these chromatograms revealed that the chromatographic profiles were quite similar between these samples with differences being mainly in peak amplitude. At first glance, background peak patterns appear to be systematic. The remainder of this sequence is running overnight and will be evaluated tomorrow. Completed data files are currently being saved as CSV files along with a snapshot of the final processed chromatogram.  

It was reported that we have one or two more extraction batches until all PSDs have been extracted. However, CERES staff will not be extracting tomorrow because the lab ran out of acetone; a solvent used in the cleaning of glassware. Additionally, the lab is not well stocked with pens. This results in an unneceesary amount of time being spent hunting for supplied within the laboratory.   

Day 10 - Stormy weather

Dakar, Senegal experienced another heavy storm last night that continued through to this evening. I video recorded some thunder and lightning and saw many cars driving through water deep enough to slosh over car hoods. I was also feeling a bit fatigued today for some reason. I arrived at the lab at 7:45 AM and left at 5:30PM.

Anna and I worked on updating the GC-ECD calibration table for the majority of the day. This process was slowed by my unfamiliarity with this version of Chemstation - the instrument software. Additionally, I was frustrated to learn that manual integrations, a tool that I rely heavily on to ensure data quality, can not be saved in away that links them to calibration tables... they are literally snapshots in time. So Anna and I would go through a chromatogram, correct mis-integrations, immediately update the cal table with this information, save the updated method, and save "chromatographic windows" incapable of being re-processed. But at the very least they provide traceability for the analysis. On the bright side, creating the calibration table in this fashion did force Anna and I to do the best job we could integrating chromatograms the 1st time; the cost is too great to go back through that process.

Regardless of these challenges, we were able to successfully input all necessary data into the calibration table.We found that almost all compounds run on the FRONT column/detector met project data quality objectives producing r^2s greater than or equal to 0.99 and were composed of 4 or more levels. The only exception was captifol. The low sensitivity of this compound resulted in a 3 point cal curve from 100-1000ppb. All other pesticides generally calibrated in a range from 5 - 500ppb with a couple from 5-1000ppb.

For the BACK column/detector, I integrated the flat-topped internal standard peak, averaged it over all cal levels, and used this value to produce average response ratios for the purposes of a semi-quant analysis. All pesticides run on the back column produced r^2s greater than or equal to 0.99 and were composed generally of 4 levels ranging from 5-100ppb.

While Anna and I worked on the GC-ECD, Adama and Marie continued to chip away at PSD extractions. All said, they completed another BIG batch, washed and began baking all necessary glassware for tomorrow, and initiated the extraction of a 5th batch of PSDs. Tomorrow CERES will continue with PSD extractions, while Anna and I begin running samples.

Day 9 - Sunday

Anna and I went into CERES today for a little over an hour. During this time we initiated extraction of a 4th batch of PSDs. This batch has ~14 samples including QC, which will make for a longer day for CERES staff tomorrow, but everyone was on board with this approach. Another batch of glassware is baking out tonight and will be ready by tomorrow morning. I confirmed with Anna that CERES does not currently have access to high purity water like the Anderson lab at OSU. Because of this, I have recommended rinsing glassware x2 with acetone and hexane just prior to bake-out. While Anna was transferring glassware to the oven, I verified retention times for the last 4 singles which were run on Friday night.

Tomorrow, CERES staff will continue with dialyses, prepare another batch for extraction tomorrow night, and another batch of glassware for bake-out. Concurrently, I will work with Anna on updating the GC-ECD cal table and assessing the quality of the calibration.