e published a paper on skimmer performance in the January 2009 issue of Advanced Aquarist magazine that detailed, for the first time, an experimental methodology to provide meaningful metrics for both the rate at which skimmers removed organics and the extent of the removal of these organics from aquarium water (Feldman, 2009). Highlights of these earlier studies included: 1. Development of a mathematical representation of the skimming process based upon a "continuously stirred reactor" model for both the skimmer and the reservoir (tank) with a given water flow rate between them. 2. Application of that mathematical formalism to both (a) a model system featuring removal of Bovine Serum Albumin (BSA) as a test case protein in freshly prepared saltwater, and (b) authentic TOC (Total Organic Carbon) removal in reef tank water. Key experimental parameters extracted from this mathematical modeling included the rate constant, k, for organic (BSA or TOC) removal, which is a singular metric reflecting the intrinsic capacity of a given skimmer to remove the organics in question, and the total % of the available organics (BSA or TOC) that were removed before the skimmer "flatlined". 3. Analysis of these data for four representative skimmers; a EuroReef CS80 needlewheel skimmer, a Precision Marine ES100 venturi skimmer, a Precision Marine AP624 airstone skimmer, and an ETSS Evolution 500 downdraft skimmer. 4. Conclusions about relative skimmer performance based upon these measurements: ◦ All four skimmers removed both BSA and TOC with similar rate constants; in short, "bubbles is bubbles", and there was no significant difference between these four skimmers in their intrinsic abilities to strip organics from saltwater. ◦ Only about 20 - 30% of the measurable TOC in reef tank water was removed by skimming, whereas almost all of the BSA was removed from saltwater by skimming. Over the intervening year, we have continued and expanded these studies of skimmer performance in several directions. In this article, we report the results of these efforts. Specifically, we have: 1. Modified our mathematical model to take into account the observation that there is a (significant) component of TOC that is not skimmable. We have applied this new model to the old skimmer data as well as to new data with new skimmers. 2. Examined the performance of three new skimmers, all of which have bubble plates; the Bubble King Mini 160 needlewheel skimmer, the Royal Exclusiv 170 Cone needlewheel skimmer, and the Reef Octopus 150 recirculating pinwheel skimmer. The Modified Mathematical Model The mathematical model derived in the January 2009 Advanced Aquarist article was based upon four assumptions: 1. The water reservoir can be treated as a continuously stirred reactor. 2. The skimmer mixing volume can be treated as a continuously stirred reactor. 3. The reservoir volume is much larger than the skimmer mixing volume. 4. The rate of TOC removal in the skimmer is proportion to the amount of TOC present in the water. Page 1 of 24
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
e published a paper on skimmer performance in the January 2009 issue of Advanced
Aquarist magazine that detailed, for the first time, an experimental methodology to provide
meaningful metrics for both the rate at which skimmers removed organics and the extent of the
removal of these organics from aquarium water (Feldman, 2009). Highlights of these earlier
studies included:
1. Development of a mathematical representation of the skimming process based upon a "continuously stirred reactor" model for both the skimmer and the reservoir (tank) with a given water flow rate between them.
2. Application of that mathematical formalism to both (a) a model system featuring removal of Bovine Serum Albumin (BSA) as a test case protein in freshly prepared saltwater, and (b) authentic TOC (Total Organic Carbon) removal in reef tank water. Key experimental parameters extracted from this mathematical modeling included the rate constant, k, for organic (BSA or TOC) removal, which is a singular metric reflecting the intrinsic capacity of a given skimmer to remove the organics in question, and the total % of the available organics (BSA or TOC) that were removed before the skimmer "flatlined".
3. Analysis of these data for four representative skimmers; a EuroReef CS80 needlewheel skimmer, a Precision Marine ES100 venturi skimmer, a Precision Marine AP624 airstone skimmer, and an ETSS Evolution 500 downdraft skimmer.
4. Conclusions about relative skimmer performance based upon these measurements: ◦ All four skimmers removed both BSA and TOC with similar rate constants; in short,
"bubbles is bubbles", and there was no significant difference between these four skimmers in their intrinsic abilities to strip organics from saltwater.
◦ Only about 20 - 30% of the measurable TOC in reef tank water was removed by skimming, whereas almost all of the BSA was removed from saltwater by skimming.
Over the intervening year, we have continued and expanded these studies of skimmer
performance in several directions. In this article, we report the results of these efforts. Specifically,
we have:
1. Modified our mathematical model to take into account the observation that there is a (significant) component of TOC that is not skimmable. We have applied this new model to the old skimmer data as well as to new data with new skimmers.
2. Examined the performance of three new skimmers, all of which have bubble plates; the Bubble King Mini 160 needlewheel skimmer, the Royal Exclusiv 170 Cone needlewheel skimmer, and the Reef Octopus 150 recirculating pinwheel skimmer.
The Modified Mathematical Model
The mathematical model derived in the January 2009 Advanced Aquarist article was based upon
four assumptions:
1. The water reservoir can be treated as a continuously stirred reactor.2. The skimmer mixing volume can be treated as a continuously stirred reactor.3. The reservoir volume is much larger than the skimmer mixing volume.4. The rate of TOC removal in the skimmer is proportion to the amount of TOC present in the
water.
Page 1 of 24
There is no reason to doubt the validity of assumptions 1 - 3. However, assumption #4 does not
take into account the experimental observation that only some of the TOC in reef tank water is
susceptible to removal by skimming. Thus, a more appropriate and general starting point would
involve explicitly breaking down TOC into two functionally distinct components:
1. [TOCl], or labile TOC that the skimmer will remove
2. [TOCr], or refractory TOC that the skimmer won't remove
With this distinction made, the prior math can be adapted to arrive at a slightly modified expression
that once again allows extraction of the two quantities of interest from the raw data: the mass
transfer rate constant, k (units = per minute), for TOCl removal, and the total amount of TOC
remaining, TOCr, when the skimmer is not pulling out any more (labile) organics from the reservoir
water. The recirculating reservoir/skimmer system (Fig. 1) maps quite closely onto a fundamental
textbook problem in mass transfer/fluid flow encountered in introductory chemical engineering
courses called the "continuously stirred" or "well-stirred" reactor problem. (Felder, 2005) In this
instance, both the skimmer and also the reservoir can be treated as "well-stirred reactors" with a
given flow Q between them. A component of the water is depleted in the skimmer by bubble-
mediated removal.
Figure 1. A generic skimmer with all mathematical
quantities defined.
Where:
Page 2 of 24
1. [TOCT]r = the total concentration of TOC in the reservoir, in ppm. Note that:
◦ [TOCT]r = [TOCl]r + [TOCr]r, where [TOCl]r = labile TOC that the skimmer will remove,
and [TOCr]r = refractory TOC, which the skimmer will not remove
2. [TOCT]s = the total concentration of TOC in the skimmer, in ppm. Note that:
◦ [TOCT]s = [TOCl]s + [TOCr]s, where [TOCl]s = labile TOC that the skimmer will remove,
and [TOCr]s = refractory TOC, which the skimmer will not remove
3. Q = the volumetric water flow rate, in gal/min4. Vr = the volume of the reservoir, in gal (30 gal)
5. Vs = the mixing volume of the skimmer, in gal
It is important to note that not all of the TOC present in reef tank water is susceptible to skimmer-
mediated removal - some types of TOC are not picked up by bubbles. To account for this
observation, we have divided the TOC into a labile, or skimmer-removable component [TOCl], and
a refractory, or skimmer-inert component [TOCr]. Therefore, the total TOC that we measure
experimentally is the sum of these two types of TOC: [TOCT] = [TOCl] + [TOCr]. However, only the
labile TOC's concentration [TOCl] is changing upon skimmer action; the refractory TOC's
concentration, [TOCr] does not vary. Thus, we can confine our mathematical derivation to this
labile TOC, TOCl, and at the end take into account the fact that experimentally, we can only
measure the total TOC, TOCT.
Since both "reactors" are interconnected, the level of the labile TOC component will drop in the
reservoir as well, and our task will be to develop a mathematical model that relates the removal in
the skimmer with the measured depletion in the reservoir. In our experimental setup, a liquid
volume Vr (reef tank water in a Rubbermaid tub) has an input stream and an output stream, and
TOC in the water becomes depleted over time via bubble-mediated removal in a skimmer with
mixing water volume Vs (see Fig. 1). Inspection of the skimmers in action permits measurement of
this "active" skimmer mixing volume, which is the value that we will use for Vs. For the purposes of
this analysis, we will assume that all of the active volume is water; that is, we will ignore the void
volume of the bubbles, as we cannot independently assess the relative contributions of bubbles
and water. This assumption will introduce an error into the calculations, but that error should be
systematic for all skimmers, and since we are interested in relative and not absolute skimmer
performance, this error should not affect the conclusions. Knowledge of the precise mechanism by
which the skimmer's bubbles removes the water component(s) is not required; all that we need to
know is that the concentration of the measured water component (TOCT in this case) is diminishing
with time in the reservoir.
It is essential for solving this problem that both the reservoir and the skimmer water volumes are
well mixed to avoid concentration gradients. The reservoir water mixing in the experiments
described below is provided by the skimmer return flow and by two powerheads in the reservoir.
We independently tested the "well mixed" assumption in the reservoir by sampling TOC levels at a
given time point at different locations (i.e., top, bottom, left side, right side) during a skimmer run.
Page 3 of 24
We observed that the site-to-site variation in TOC levels at different locations was no greater than
the sample-to-sample variation at one location (both ~ 10%), suggesting that there is no reason to
suspect that the "well mixed" assumption is not applicable. The mixing in the skimmer reaction
chamber is provided by both rapid water movement and the agitation caused by the motion of the
bubble stream. We have no independent experimental measurement/confirmation of mixing
behavior in the skimmer.
The application of this mathematical approach to the protein skimmer problem leads ultimately to
two important equations, labeled 22 and 24 below. The complete mathematical derivation (i.e.,
Eqs. (1) - (21)) can be found in the next section. This derivation that is largely taken from the 2009
Advanced Aquarist article. This new version of the derivation takes into account the presence of
• Vr = the total volume of the reservoir water, in gal
• [TOCl]r = the concentration of labile TOC in the reservoir at any time t and also the
concentration of labile TOC in the stream leaving the reservoir and entering the skimmer• Q = the volumetric flow through the system, in gpm• [TOCl]s = the concentration of labile TOC in the stream leaving the skimmer and entering the
reservoir
Eq. (8) says that the change in the amount of TOC in the reservoir (the left hand side) is equal to
the difference between the reservoir input and output TOC concentrations ([TOCl]s - [TOCl]r) times
the flow rate (the right hand side). Note that this expression includes information about TOC
concentrations in both the reservoir and the skimmer.
A similar expression can be developed for the fate of the TOC concentration just in the skimmer.
However, in this case, "removal" does not equal 0, as the bubbles in the skimmer actively remove
etc., etc.), there does not appear to be any compelling reason to favor one type of skimmer
design/bubble generation mechanism over any other amongst the seven skimmers examined
(Reef Octopus 150 excepted). That is, the inclusion or omission of a bubble plate does not seem to
have any decisive effect on the rate constant for TOC removal, nor does a change in skimmer
geometry from cylindrical to cone-shaped. Likewise, all methods of bubble generation examined
appear adequate.
One question that impacts on the mathematical model derivation, especially the simplification
discussed with Eq. (25), involves the relationship between the flow rate Q and the intrinsic rate
constant k. There is no provision in the math that links these two quantities; however, if k does
depend upon Q, then the simplification of Eq. (25) will not hold. This point was examined
experimentally with the Reef Octopus 150 skimmer, as this skimmer has a recirculating pump for
bubble introduction and hence the water flow rate Q should not influence the rate of bubble
generation. Flow rates from 2.69 gal/min to 6.50 gal/min (= 162 gal/hr to 390 gal/hr) were
examined. Within this flow regime, there was no significant variation in the derived rate constant k.
So, at least for the Reef Octopus 150 skimmer, and by extension to all of the skimmers, we
Page 20 of 24
proceed with the analysis as if k is not dependent on water flow rate Q. Note that in the original
January 2009 Advanced Aquarist publication, we attempted to examine the same point using the
airstone skimmer (Precision Marine AP624). In those trials using BSA removal as the experimental
parameter, we did observe a non-linear response between Q and k; Q = 156 gal/hr, k = 3.1 min-1;
Q = 318 gal/hr, k = 7.6 min-1; Q = 540 gal/hr, k = 2.5 min-1. The basis for the discrepancy between
the BSA/airstone skimmer results and the TOC/Reef Octopus results is not clear.
The overall rate of TOC removal, as modeled by the exponential term of Eq. (22), is a function of
the rate constant k, the skimmer mixing volume Vs, and possibly the flow rate Q (see Eq. (25) and
the accompanying discussion). Therefore it is perhaps a more relevant metric for answering overall
questions about skimmer performance, since it takes into account the distinct operational
parameters (flow, size) of each skimmer. For this metric, large error bars on the order of 10-40% of
the average value once again suggest caution in (over) interpreting the data. There are t-test-
based statistically significant differences in the values for this rate measurement amongst many
the skimmers. Once again, the Reef Octopus 150 displays a rate that is significantly less than all of
the other skimmers tested. The Bubble King and Royal Exclusiv skimmers do not display
statistically significant differences in their rates of TOC removal, but both of these skimmers do
operate at an appreciably slower rate of TOC removal than the Precision Marine ES100, the
Precision Marine AP624, and the ETSS Evolution 500.
The skimmers all have different mixing volumes Vs, ranging from a maximum of 1.3 gallons for the
Reef Octopus 150 down to a minimum of 0.69 gallons for the Bubble King and 0.62 gallons for the
EuroReef. These differences in skimmer sizes become influential in determining the overall rate of
TOC removal, whereas the flow rate Q has much less significance (see the discussion along with
Eq. (25)). For example, the smallish Bubble King mini 160 skimmer has a relatively small overall
rate of TOC removal even though its intrinsic rate constant k is in the middle range; that is, its
smaller size really limits its ability to remove TOC rapidly. The similarly sized EuroReef CS80 has a
much larger rate of TOC removal since its intrinsic rate constant k is 4x the Bubble King's k value.
Conversely, the large volume of the Reef Octopus 150 does not overcome an intrinsic rate
constant k at the lower edge of the calculated values and so it exhibits the smallest rate of TOC
removal amongst all of the skimmers tested. Of course, all of these skimmers arrive with different
price tags, and so a relevant question might focus on the price/performance trade-off. These data
are illustrated in Fig. 7. The prices listed are the standard retail price that was paid at the time that
these skimmers were purchased. It should not escape notice that the least expensive skimmer (the
Precision Marine ES100) offers the greatest rate of TOC removal.
Page 21 of 24
Figure 7. Price/performance comparison for the skimmers examined in this
study.
In a practical sense, it is important to resist over interpreting these rate-of-TOC removal data. Most
aquarists run their skimmer 24/7, and under that husbandry regime, the major impact of differing
rates of TOC removal will be on how often the skimmer cup needs to be cleaned. If the skimmer
removes TOC with a faster rate, then the cup will be filled faster and hence have to be cleaned
sooner.
One of the more surprising and important observations to emerge from the earlier skimmer studies
was that the four original skimmers tested removed only 20 - 30% of the measurable TOC in the
reef tank water examined; the remaining 70 - 80% of the TOC was not removed by skimming.
Extension of these measurements to the three new skimmers tested in this study did not add much
to the argument. The Reef Octopus' removal amount fell within this range, whereas the Bubble
King and Royal Exclusiv skimmers appeared to remove incrementally more of the extant TOC,
perhaps up to the mid-30% range. An explanation for this observation was offered in the January
2009 Advanced Aquarist article; in summary, skimmers can only remove what bubbles trap, and
bubbles only trap molecules and/or particles (i.e., bacteria, diatoms, etc.) with some compelling
thermodynamic reason to adhere to the bubble's surface. On the molecular level, this surface
association is typically driven by the molecule/particle having a hydrophobic (= water hating) patch
that can be buried in the bubble surface/interior. This arrangement avoids the energetically
penalizing juxtaposition of hydrophobic surfaces with (hydrophilic) water, and so overall the system
energy is lowered (a favorable occurrence). Some of the molecules/particles in aquarium water will
meet this hydrophobic region criterion, and some will not. The ones that do not have a sufficiently
large hydrophobic patch will not interact with bubbles, and hence will not be removed by skimming.
From, the results of the experiments described here, it appears that only 20 - 35 % of the
measurable TOC meets this hydrophobicity criterion (= [TOCl] defined earlier) whereas the
remaining 65 - 80 % does not (= [TOCr] defined earlier). In essence, bubbles are a rather poor
Page 22 of 24
media for removal of organic nutrients from aquarium water compared to, for example, GAC.
However, they do have the distinct benefit of being cheap.
Conclusions
Many factors contribute to the "value" of a skimmer to an aquarist, including quality of construction,
size, footprint, noise level, ease of cleaning, energy efficiency of the pump, and of course, the
ability to remove organic waste from aquarium water. Our data show that there are not compelling
or remarkably large differences in measurable skimmer TOC removal metrics among the seven
skimmers tested, although the Reef Octopus 150 consistently underperformed compared to the
other skimmers. However, in the larger picture, it is equally apparent that if an aquarist runs a
skimmer continuously (24/7), then any of the skimmers tested would perform adequately in terms
of rate of TOC removal; the only practical differences might involve the frequency of skimmer cup
cleaning. A perhaps more interesting observation to emerge from these skimmer studies involves
not the rate of TOC removal, but rather the amount of TOC removed. None of the skimmers tested
removed more than 35% of the extant TOC, leading to the conclusion that bubbles are really not a
very effective medium for organic nutrient removal. If fact, the presence of refractory, or
unskimmable, TOC, coupled with the likelihood that endogenous TOC consumers (bacteria,
among others) also do not remove all of the TOC present (cf. Fig. 4), suggest that in an operational
sense, TOC can be categorized as follows:
1. TOC that a skimmer removes2. TOC that a skimmer does not remove3. TOC that is consumed by microbes4. TOC that is not consumed by microbes5. TOC that is (indirectly or directly) harmful to tank livestock6. TOC that is not harmful to tank livestock
The last two categories must be included as a result of recent work of Forest Rohwer (See the
January 2009 Advanced Aquarist article for a discussion), and they really highlight why an aquarist
might be concerned with rising tank TOC levels. Of course, there will be much overlap between
these categories. Ultimately, the crucial question for sustaining aquarium livestock health over the
long term is, "How much of the harmful TOC (#5) is removed by either biological consumption or
by skimming?" That question remains unanswered at present.
The results to date on protein skimming as a means of aquarium water remediation form a
consistent picture that is at odds with some of the cherished dogma in the marine husbandry area.
According to the data presented in this and the earlier paper (Advanced Aquarist, January 2009),
protein skimmers appear to have a much larger variation in their prices than they do in their ability
to remove TOC from aquarium water. Recent design innovations like bubble plates, conical sides,
or pinwheel impellers do not seem to impact significantly on either rate of TOC removal or amount
Page 23 of 24
of TOC removed, at least for the skimmers tested. Thus, skimmer manufacturer claims about
enhanced organic removal capabilities should be met with skepticism in the absence of compelling
and quantitative TOC removal data.
Acknowledgments
We thank the Eberly College of Science at the Pennsylvania State University and E. I DuPont de
Nemours and Co. for financial support, Dr. Bruce Logan and Mr. David Jones of the Pennsylvania
State University Department of Civil and Environmental Engineering for use of the Shimadzu 5000
TOC Analyzer, Dr. James Vrentas of the Pennsylvania State University Department of Chemical
Engineering for assistance in developing the mathematical model described in this article, and Dr.
Sanjay Joshi for the use of his reef tank and for many helpful discussions.
References
1. Allers, E.; Niesner, C.; Wild, C.; Pernthaler, J. 2008. "Microbes Enriched in Seawater after Addition of Coral Mucus." Appl. Environ. Microbiol., 74, 3274-3278.
2. Covert, J. S.; Moran, M. A. 2001. "Molecular Characterization of Estuarine Bacterial Communities that Use High- and Low-Molecular Weight Fractions of Dissolved Organic Carbon." Aquat. Microb. Ecol., 25, 127-139.
3. de Goeij, J. M.; van Duyl, F. C. 2007. "Coral Cavities are Sinks of Dissolved Organic Carbon." Limnol. Oceanogr., 56, 2608-2617.
4. Felder, R. M.; Rousseau, R. W. 2005. Elementary Principles of Chemical Processes, 3rd Ed., John Wiley and Sons, New York.
5. Feldman, K. S.; Maers, K. M.; Vernese, L. F.; Huber, E. A.; Test, M. R. 2009. "The Development of a Method for the Quantitative Evaluation of Protein Skimmer Performance." Advanced Aquarist http://www.advancedaquarist.com/2009/1/aafeature2/
6. Kirchman, D. L. 1990. "Limitation of Bacterial Growth by Dissolved Organic Matter in the Subartic Pacific." Mar. Ecol. Prog. Ser., 62, 47-54.
7. Mopper, K.; Stubbins, A.; Ritchie, J. D.; Bialk, H. M.; Hatcher, P. G. "Advanced Instrumental Approaches for Characterization of Marine Dissolved Organic Matter: Extraction Techniques, Mass Spectrometry, and Nuclear Magnetic Resonance Spectroscopy." Chem. Rev., 107, 419-442.
8. Sharon, G.; Rosenberg, E. 2008. "Bacterial Growth on Coral Mucus." Curr. Microbiol., 56, 481-488.
9. Wild, C.; Rasheed, M.; Werner, U.; Franke, U.; Johnstone, R.; Huettel, M. 2004. "Degradation and Mineralization of Coral Mucus in Reef Environments." Mar. Ecol. Prog. Ser., 267, 159-171.
10. Zweifel, U. L.; Norrman, B.; Hagström, A. 1993. "Consumption of Dissolved Organic Carbon by Marine Bacteria and Demand for Inorganic Nutrients." Mar. Ecol. Prog. Ser., 101, 23-32.