Arizona state university

ARIZONA STATE UNIVERSITY

National Center for Sustainable Water Supply

Department of Civil and Environmental Engineering

Tempe, AZ 85287-5306

480/965-2885 Fax [phone]

email: [e-mail]

TO: City of XXX

FROM: Paul Westerhoff

SUBJECT: Screening PAC Suppliers Bid Submissions

ASU screened eight (8) powder activated carbon (PAC) samples for their ability to remove MIB and Geosmin, with the intent of the City of XXX partially basing the selection of a PAC supplier upon this data. This was a blind testing study; codes (A through H) were used to designate each PAC brand. ASU takes no legal responsibility for the City of XXX decisions for a PAC supplier. Below is the testing and evaluation protocol, test findings, and results. The recommendation from this study is that the City of XXX contract with the supplier of PAC brand “B”.

Testing Protocol

Water Source. Water was collected from the Salt River (the dominant water supply for SRP during peak taste and odor (T&O) episodes). The water will be filtered (Whatman GF/F), and DOC measured. MIB will be spiked into the water for final concentrations of approximately 81 ng/L; Geosmin spiked to 58 ng/L.
PAC Batch Experiments. Activated carbon adsorption studies with MIB and Geosmin will be conducted in the laboratory with commercially available brands of PAC. PAC samples were obtained from a single batch from manufacturers in amounts sufficient to run all experiments. A total of eight (8) different PAC types will be tested. PAC doses will be set at two PAC doses (15 and 25 ppm). This was a blind testing study; codes (A through H) were used to designate each PAC brand. A PAC slurry of each PAC sample was prepared at a concentration of 2500 mg PAC/L ultra-pure water; the slurry was mixed and allowed to hydrate for 24 hours at room temperature. Amber glass bottles (250ml) were used for treatments and were shaken on a wrist shaker (Multi-wrist shaker, Lab-Line, Melrose Park, IL). The duration of shaking was based upon the average hydraulic residence time of PAC in the pre-sedimentation basins plus flocculation basins (conservative HRTs provided by the City of XXX – 5 hours). Activated carbon was removed from the samples by syringe filtering with a 0.2 um nylon filter (Acrodisc 32 Supor 0.2 um syringe filters, Pall Corporation, Ann Arbor, MI). Control treatments containing MIB and Geosmin, but no PAC, were shaken and filtered in a similar manner as the samples containing PAC. Experiments were conducted at room temperature. All experiments were conducted in duplicate.
Measurement of MIB and Geosmin. MIB and Geosmin were measured using Solid-Phase Microextraction/Gas Chromatography Mass Spectroscopy (SPME-GC/MS) (Watson et al., 2000; Lloyd et al., 1998). Twenty-five ml of sample is added to a 40 ml septum capped vial that contains 8 gm desiccated sodium chloride and a magnetic stir bar. An internal standard (10 ng/L IPMP, Aldrich Chemical Co., Milwaukee, WI) is added through the septum and the vial is placed in a water bath on a magnetic stir plate heated to 50  1.5 ⁰C. A SPME fiber (Supelco # 57348 U) is introduced into the head space gas through the septum and the sample is stirred for 30 minutes. The fiber is removed from the vial and inserted into the gas chromatograph injector at 250 ⁰C for 5 minutes. The fiber was then retracted into the holder, removed from the GC inlet and reused for the next sample. Compounds are eluted from the column gas chromatograph to a mass spectrometer set for selective ion storage (selective m/z values: MIB = 95, Geosmin = 112 and IPMP = 124, 136). Calibration curves are generated using MIB and Geosmin standards (mixture standard: Supelco # 47525 U). Method detection limit for SPME is 2 ng/L. An MIB internal standard was run in triplicate, and had excellent reproducibility: 27.50.8 ng/L.
results
The fraction remaining of MIB and Geosmin was calculated from experimental results. The fraction remaining is defined as C/Co, where C is the MIB or Geosmin concentration (ng/L) after contact with the PAC and Co is the initial MIB or Geosmin concentration (ng/L). The PAC brands (A through H) were ranked from best performing (lowest C/Co) to worst performing (highest C/Co) for MIB or Geosmin removal (Figure 1). The values indicated in the bar charts are the average of two separate PAC tests, and the error bar represents the difference between the average and one of the samples. At a PAC dose of 15 ppm the fraction of MIB remaining ranged from 0.59 to 0.88, with the top three performing PAC brands being: B>H>E.
At a PAC dose of 25ppm more MIB was removed than at 15 ppm, and lower fraction remaining values were observed (Figure 1). At a PAC dose of 25 ppm the fraction of MIB remaining ranged from 0.33 to 0.66, with the top three performing PAC brands having essentially equivalent MIB removal capability (brands H, E, and B).
Geosmin was removed more effectively than MIB (Figure 1). Geosmin removal at 25 ppm of PAC was greater than 65% (data not shown). Geosmin removal at a PAC dose of 15 ppm is shown in Figure 1. The fraction of Geosmin remaining ranged from 0.31 to 0.86, with the top three performing (lowest C/Co) PAC brands being: B>H>E.
interpretation of results
After completion of the blind laboratory PAC performance testing, the City of XXX provided unit cost data on each PAC (A through H). The unit costs provided are presented in Table 1. The three least expensive PAC brands were: BIndex Value = [% MIB Remaining]x[Price per pound] Equation 1
In principle, the PAC brand with the lowest Index Value represents the most cost effective brand of PAC. For example, a lower PAC dose could offset a higher PAC price.
Table 1 – PAC Unit Costs

PAC Brand	PAC Unit Cost ($/lb)
A	$0.311
B	$0.275
C	$0.385
D	$0.300
E	$0.500
F	$0.289
G	$0.418
H	$0.478

Index values for each PAC brand at two PAC doses (15 and 25 ppm) for MIB and 15 ppm PAC dose for Geosmin are shown in Figure 2. The Index values for the 25 ppm PAC dose and Geosmin were all quite low, given the high removal efficiency. Based the ranking of Index Values for MIB removal with 15 or 25 ppm of PAC, the same trend is observed: B < A < D. Therefore, PAC brand “B” would be the most cost effective. A summary of the Index Values and removal efficiencies are provided in Appendix A.

Recommendation

Based upon the batch PAC tests, PAC brand “B” had the among the highest MIB removal efficiency (lowest C/Co) and had the lowest unit cost. Correspondingly, brand “B” also had the lowest Index Value. Therefore, we would recommend brand “B” for the City of XXX PAC supplier.

However, other PAC brands are capable of performing equally well in removing MIB and Geosmin. Other high performing PAC brands would include: H and E, followed by A and G. The combination of MIB removal efficiency and bid price just does not make them as good a choice as brand “B”. The City may also want to consider the following issues in final selection of a PAC supplier:

• 
  Availability of product
  
• 
  Product handling issues
  
• 
  Size and settling characteristics of the PAC
  

Figure 1 – Summary of Fraction MIB or Geosmin Remaining



Figure 2 – Summary of PAC Index Values


Appendix A
Experimental Results and Computations

Experimental Results and Computations for 15 ppm PAC Doses


Experimental Results and Computations for 25 ppm PAC Doses