Is copper safe for Neocaridina shrimp?

Evidence review: August 12, 2026

"Contains copper" is not a dose, and "copper-free" is not a complete safety test. You need the chemical form, route, amount, water concentration, exposure time and water chemistry before an ingredient name becomes an exposure estimate.

Five different questions hide inside "copper"

The U.S. EPA describes copper as an essential nutrient at low concentrations and a toxicant at higher concentrations.¹ Its freshwater Biotic Ligand Model uses multiple water inputs because a total-copper number alone does not predict bioavailability.² That environmental framework is useful chemistry. It is not a home-aquarium shrimp dosing calculator.

There is no defensible universal copper ppm here

This review found no controlled experiment that establishes a safe home-aquarium copper concentration for N. davidi. Direct 2015 and 2018 papers exposed shrimp under historical Neocaridina labels to nominal copper series, but neither independently identified the animals as N. davidi or measured a safe concentration.^21,23 An LC50 is the concentration associated with median mortality in a particular experiment. It is not a no-effect level, a chronic target or permission to dose below it.

Values from another crustacean, an EPA mixed-community criterion, a fish medication or a water company report cannot silently become a cherry-shrimp limit. Species, life stage, chemical form, exposure duration and water chemistry all have to travel with a toxicity number.

What direct Neocaridina toxicant experiments show

The silver result is especially useful for one narrow lesson: the same element in different forms did not produce one interchangeable toxicity value. Do not use any of those acute numbers as aquarium targets.

The acetaminophen and ibuprofen study adds a different lesson: a lower mixture LC50 is not automatically proof of synergy. The paper tested 10- to 14-day-old, approximately 1.5 mm juveniles for 96 hours. It reported LC50 values of 6.07 mg/L for acetaminophen, 6.60 mg/L for ibuprofen and 6.23 mg/L for its 1:1 series.²⁴ The equal-ratio result does not clearly state whether that concentration is the amount of each drug or their sum.

In the unequal series, the reported 4.78 mg/L value is the changing acetaminophen concentration with another 0.5 mg/L ibuprofen present. The reported 6.78 mg/L value is the changing ibuprofen concentration with another 0.5 mg/L acetaminophen present. Comparing those labels with single-drug LC50 values without carrying the fixed co-exposure changes the quantity being compared. Their confidence intervals also overlap substantially.

The paper calls the unequal patterns synergistic and antagonistic, but it does not report a concentration-addition, independent-action, toxicity-unit, model-deviation or isobologram analysis. Concentrations were nominal, DMSO stocks lacked a stated vehicle-matched control and the paper does not reconcile four stated repeats, duplicate treatments and Table 2's n of four. The result is a set of ratio-specific acute mortality curves, not a general rule for mixing medications or predicting household wastewater.

The 2015 copper and chlorpyrifos paper used three parallel 2-litre beakers per nominal concentration, 20 fasted shrimp per beaker and daily solution renewal. Three shrimp per group were destructively sampled each day, but the report does not explain how those animals were distributed among or pooled across beakers. The 96-hour LC50 values came from earlier work used to choose sublethal treatments; they were not results of this five-day series. The predecessor is now identified in the research queue as a 2014 acute paper, but its numerical tables still await direct rendered-page inspection. Nominal exposure, uncertain taxon identity and muscle biomarkers cannot establish a safe home concentration.²¹

The 2021 imidacloprid paper used a commercial 28.8 percent formulation and apparently one exposure tank per treatment. Measuring several animals or videos from that tank does not create independent treatment-tank replication. The shrimp were fasted, no formulation-only control or measured water concentration was reported, and acetylcholine itself changed the measured endpoints. The later docking model used a snail acetylcholine-binding protein rather than a shrimp receptor. These results do not validate acetylcholine as an antidote or let a keeper identify imidacloprid from slowed movement.²²

The 2018 copper transcript paper used a separate non-exposed ten-animal library to discover candidate sequences, then pooled whole-body RNA from three shrimp per copper treatment. Its figure labels a 24-hour nominal dose series from 0 to 100 micrograms per litre, but the paper does not state the concentration used for the time series, identify the exposure vessel or report independent vessel replication. It also reports 28 PSU salinity despite describing freshwater lake animals. The changed mRNA targets are candidates for further validation, not a field test for copper exposure or evidence that a particular shrimp was harmed.²³

The two nickel publications are also one research sequence. They are not two independent dose replications. The 2024 paper selected nominal 3.63 mg Ni/L after about half the animals died by 14 days during preliminary concentration work. Its main experiment followed surviving adults through one- or two-week exposures and one or two later clean-water weeks. The 2025 continuation used that experimental scheme and added cell-cycle, proliferation, energy, mitochondrial and protein measures, with five specimens per method group.^5,16 Some measures moved toward control values while others did not. That is an organ- and endpoint-specific clean-water response, not proof that the animals, colony or aquarium recovered.

The nickel concentration was nominal rather than analytically verified, the control animals remained in the continuous culture instead of a separately described time-matched cup group, and the later clean-water groups necessarily contained survivors. Neither publication measured behaviour, reproduction or a household-safe concentration. Do not turn the selected high treatment, its preliminary mortality observation or a cellular marker into an LC50, a clearance schedule or an aquarium diagnosis.

The two linked CC BY data deposits add useful denominators without making the experiment safer or more independent. The 2024 deposit has three numbered TXRF samples per group and reports whole-shrimp nickel, but it does not measure nickel in the exposure water. Its reactive-oxygen tables have uneven row counts and repeated analysis dates without identifying every row as a separate animal. The 2025 deposit has five values per group and organ for mitochondrial potential, ATP and ADP/ATP, but no specimen identifiers linking one endpoint to another.^27,28 Raw rows improve auditability. They do not create extra biological replication.

The microplastics paper described slower growth and reproduction, but it counted animals in broad size classes rather than following individual lengths, and it did not measure fecundity, eggs or gonads. Births, growth, mortality and missed animals can all change a size-class count. The direct result is therefore a treatment difference in mortality and population structure under exposure to smooth 38 to 45 micrometre spheres, not a measured reproductive rate or a household-plastic diagnosis.

The shorter polystyrene study adds behaviour and oxygen-consumption evidence, but it also needs a statistical caution. Exposure was assigned by aquarium, while the published mixed model listed individual and day effects without listing aquarium as a random effect.¹⁰ Treat the treatment difference as bounded evidence from this design, not 80 independent water replicates or a household-plastic safety line.

A separate 2021 experiment used purchased White Pearl N. palmata, not N. davidi. In one run, individual shrimp occupied eight 500 mL vessels per treatment for 24 hours. Particle counts in rinsed whole-shrimp lysates rose across the tested concentration series, and particles were recovered from excretions after a separate four-hour transfer period.²⁵ Those are ingestion and passage observations, not evidence of injury or safety.

The immediate and four-hour groups were different animals and their particle counts varied widely. The supplement provides group summaries rather than animal-level records, does not identify the post-hoc outliers and gives only one example for nominal-versus-actual particle delivery. Beads and fragments also differed in polymer, size distribution, preparation and detection, so a non-significant shape comparison does not isolate shape. Whole-shrimp lysate does not prove tissue transfer, three deaths across different treatments do not demonstrate toxicity, and the experiment does not define a clearance time or household-plastic threshold.

A companion study exposed purchased White Pearl N. palmata to concentrated, particle-filtered extracts from one recycled LDPE material and one starch-blend foil.²⁶ Duplicate 24-hour leachates were combined into one extract per material and weathering condition before eight one-shrimp vessels per concentration were observed for 14 days. Those vessels replicate shrimp responses under a shared extract. They do not independently replicate chemical release from the plastic.

The extract concentrations were reported as source-plastic equivalents, not as identified chemical concentrations or particle densities. The extracts strongly inhibited bacterial luminescence, but moved distance and frozen events in shrimp were highly variable and showed no consistent concentration-response. The study did not identify the causal compounds, test N. davidi, or measure growth, reproduction, histology, chronic aquarium exposure or natural weathering. A bacterial assay effect and two mostly negative movement endpoints cannot be turned into either a danger verdict or a safety certificate for household plastics.

The diclofenac study used 22 replicate 1.5-litre aquaria per treatment and checked concentrations at 0 and 72 hours in two replicates per treatment. The overall means were 0.0075, 0.1320 and 1.0730 mg/L: both control checks contained diclofenac, at 0.006 and 0.009 mg/L. The first eggs had partly developed before exposure, and later brood results depended on females surviving and spawning. Hatchling counts did not differ significantly among treatments, while the significant abnormality result came from the small second-spawn sample. Those details make this useful direct evidence, but not a safe-concentration study for accidental trace exposure. The SSRN and Research Square postings are preprint versions of this experiment, not independent replications.¹¹

The chlordane and lindane evidence is a three-paper research sequence, not one experiment published three times and not three independent replications. The first 2004 paper used juveniles in an acute mortality series and a separate 28-day hormone-assay experiment. The second used identified males for hormone-associated assays and measurements of the masculine appendage. The 2006 paper used female hormone groups and separate mixed-sex reproductive groups.^19,20,12

The two 2004 papers describe triplicate exposure beakers, but neither maps its animal-level assay counts clearly to those beakers. Both used repeated paired t tests without a reported correction for the many treatment-by-time comparisons. Their alkali-labile phosphate assay was an acknowledged nonspecific proxy for vitellogenin-like protein, and the estradiol positive control used alcohol without a described alcohol-matched control. The juvenile paper additionally conflicts with its own acute-treatment count and prints one confidence interval in descending order. The male appendage results varied by time and concentration and did not test reproductive competence.

The 2006 reproductive comparison is weaker at the treatment-unit level. Each treatment placed 20 males and 20 females in one shared 10-litre beaker, after which only the females that became ovigerous supplied egg and hatch outcomes. Some groups produced only one or two ovigerous females and one produced none. The sequence supports treatment-associated responses under its protocols, but not a household threshold, a precise effect size, a visible diagnostic sign or a proven endocrine mechanism.

The nonylphenol paper demonstrates another common evidence trap: a changed transcript is not automatically a sick shrimp. The one-day experiment used nominal concentrations, pooled whole-animal RNA and semi-quantitative RT-PCR. The accessible methods do not report independently replicated exposure aquaria, and the abstract says six transcripts responded at 0.01 mg/L while the results text says five. A figure also calls the comparator PBS-treated even though the methods describe a DMSO solvent control.¹⁷ The result supports a short-term transcriptional response under that protocol. It does not establish protein abundance, impaired function, chronic harm, recovery, a visible diagnosis, a consumer-product rule or a home safe level.

The dipropyl phthalate paper reinforces the endpoint lesson and adds several design cautions. Ten shrimp were pooled for each RNA sample, while the statistics section instead says each replicate used at least five shrimp. The paper does not report the number of independent exposure vessels. DPrP stocks were prepared in acetone, but the control used DPrP-free pond water rather than a described acetone-matched control.¹⁸

Reporting also conflicts. The English abstract says all six responsive targets at 50 mg/L were upregulated. The Chinese abstract, main results, figures and table show them as downregulated, and the table footnote mistakenly names nonylphenol. The source can show that selected pooled transcripts changed under its protocol. It cannot establish protein abundance, immune performance, survival, disease susceptibility, chronic harm, a consumer plastic cause, a home threshold or a ready biomarker panel.

The EE2 study is a useful warning against reading growth alone as benefit. Females in the two lowest nonzero nominal treatments were heavier and longer, yet they produced fewer eggs, had lower hatching outcomes and had lower adult survival than the control. Exposure was not analytically verified, the clean-water control was not matched for ethanol, the written stock-volume calculation conflicts with the reported maximum ethanol percentage and control survival was only 57.3 percent. Males and offspring were not directly exposed, and eggs were artificially incubated. These limits prevent a household safe level or beneficial-dose claim.¹³

The two BPA papers report different cellular endpoints from one continuing acute exposure program, not two independent replications. Adults were held one per 0.5-litre container in nominal 1, 5 or 10 mg/L BPA for 24, 48 or 72 hours. The first paper measured midgut structure, autophagic signals, reserves and cell-cycle measures; the second added reactive-oxygen, cell-death, mitochondrial, glutathione and DNA-damage endpoints.^14,15

The changing timelines matter. Reactive-oxygen signals were highest after 24 hours and later declined, while other damage measures changed at later times. A falling marker therefore cannot be called recovery. Concentrations were not analytically verified, controls were not described as separate time-matched groups, and five animals or cell preparations per method do not reveal the total independent animal count across all assays. Neither paper measured chronic exposure, whole-animal survival, behaviour, reproduction or recovery. Do not use these milligram-per-litre treatments to diagnose an aquarium, condemn an unnamed plastic item or publish a household safe level.

How to evaluate common product questions

Food with copper on the label

Copper is a trace nutrient. Its presence on a complete-food label does not tell you the amount entering the water, the fraction eaten, the chemical form or the final concentration. That label alone establishes neither safety nor poisoning. Feed a measured amount for a recorded purpose and remove recoverable excess using the feeding protocol.

Plant fertilizer

"For planted aquariums" and "shrimp safe" are marketing categories, not exposure measurements. Keep the exact product, guaranteed analysis, dose, water volume and lot. A correctly used trace-nutrient product is not equivalent to a copper medication, but no brand label can validate every line, water chemistry, overdose or product version. Do not dose a production colony merely to chase a plant chart.

Medication and algaecide

A fish dose is not a shrimp dose. Read the active ingredient, concentration, target species, contraindications and current directions. If the label does not explicitly address the invertebrates present, it does not establish their safety. Use a commissioned hospital system and qualified diagnostic guidance rather than testing a treatment in an irreplaceable colony. The word "natural" changes none of these requirements.

New plants, botanicals and hardscape

Ask for source and treatment history. Rinsing can remove loose material; it cannot certify removal of a systemic pesticide, absorbed compound, every surface residue, egg or epibiont. Tissue culture reduces some wet-transfer routes, but it is not a chemical-analysis certificate. Unknown wet additions belong in a separate, commissioned plant-observation system. No fixed dip or soak time proves safety.

Household aerosols and topical pet treatments

Our prevention rule is to keep insect spray, cleaner, fragrance, paint, nicotine liquid, lotion and topical flea or tick products away from open water, wet equipment and air-pump intakes. We do not spray in the aquarium room. Record nearby pest-control, renovation, smoke and spill events even when direct contact was not seen.

Read a test result literally

• Record the analyte: total copper, dissolved copper or another stated fraction.
• Record the unit, range, reporting limit, expiry, calibration and sample timing.
• "Not detected" means below that method's reporting boundary, not zero.
• A general TDS or conductivity meter cannot identify copper or pesticide.
• A copper non-detect cannot exclude another toxicant.

If a laboratory result matters to a diagnosis, preserve the sample collection, filtration, digestion and chain-of-custody details rather than copying only the number.

Suspected chemical exposure: act without inventing a diagnosis

1. Stop the continuing source. End the dose, isolate a leaking item and prevent more aerosol, runoff or treated water from entering.
2. Protect respiration. Restore a flow or aeration failure with verified clean equipment and air. Do not pull contaminated room air through the tank.
3. Close movement. Do not sell, transfer or share plants, media, water or wet tools from the system.
4. Preserve the evidence. Keep labels, lots and receipts. Record the timeline, photographs, deaths, water changes and every wet or airborne addition.
5. Measure immediate competing hazards. Verify temperature, total ammonia nitrogen, nitrite, pH and dissolved oxygen when available. Check the source-water and chlorine or chloramine treatment.⁸
6. Prepare known replacement water. A water change can dilute a conservative dissolved substance; it cannot calculate adsorption, release, metabolism or a safe endpoint. Use the water-change guide.
7. Compare transfer risks. An uncommissioned container can add nitrogen, oxygen, temperature and handling hazards. Use a commissioned hospital system when the source tank cannot be controlled.

Do not add a second chemical merely to neutralize a guessed first chemical. Carbon, resins, adsorbents and conditioners have compound-specific capacities and limitations. Ongoing mortality, loss of balance or an uncertain exposure warrants an aquatic veterinarian or qualified diagnostic laboratory. The American Association of Fish Veterinarians provides a North American locator.⁹ Ask whether the clinician accepts aquatic invertebrates; a directory listing does not establish that scope.

What the evidence does not prove

• That any detectable copper automatically caused a death
• That a copper-free label excludes pesticides, disinfectants or other metals
• That every fertilizer containing a trace nutrient is unsafe
• That a water conditioner neutralizes an unidentified exposure
• That rinsing or a fixed plant quarantine removes every residue
• That clear water, surviving animals or one negative kit clears the system
• That one water-change percentage or clean-water duration fits every compound

Sources and evidence boundaries

1. U.S. Environmental Protection Agency. Aquatic Life Criteria - Copper. Essentiality, toxicity and the national ambient-water framework.
2. U.S. Environmental Protection Agency. Copper Biotic Ligand Model. Water-chemistry inputs governing bioavailability.
3. Hu et al., 2019. Effects of two juvenile hormone analogue insecticides, fenoxycarb and methoprene, on Neocaridina davidi. Environmental Pollution 253: 89-99.
4. Ostróżka et al., 2022. Can insecticide-free clean water regenerate the midgut epithelium of the freshwater shrimp after dimethoate treatment? Micron 155: 103162.
5. Ostróżka et al., 2024. Toxic effects of nickel on tolerance and regeneration in the freshwater shrimp Neocaridina davidi. The European Zoological Journal 91(1): 180-205.
6. Hugger, 2025. Differences in solubility behaviour and acute toxicity of ionic silver and silver nanoparticles towards Neocaridina davidi. Master's thesis, University of Vienna.
7. McIvor, Baring, Qin and Mitchell, 2026. Microplastics Affecting Population Growth, Mortality and Life Stages of Cherry Shrimp, Neocaridina davidi. Aquaculture Research 2026(1): 6618042.
8. U.S. Centers for Disease Control and Prevention. About Water Disinfection with Chlorine and Chloramine. Drinking-water disinfectant distinctions.
9. American Association of Fish Veterinarians. Find a Fish Vet.
10. Kucera et al., 2022. Polystyrene Microparticles and the Functional Traits of Invertebrates: A Case Study on Freshwater Shrimp Neocardina heteropoda. Fishes 7(6): 323.
11. Zanitti, Medesani, Rodríguez and López Greco, 2023. Long-Term Exposure of the Red Cherry Shrimp Neocaridina davidi to Diclofenac: Impact on Survival, Growth, and Reproductive Potential. Archives of Environmental Contamination and Toxicology 85(2): 181-190.
12. Huang, Chen, Wu and Wang, 2006. Reproduction obstacles for the female green neon shrimp Neocaridina denticulata after exposure to chlordane and lindane. Chemosphere 64(1): 11-16.
13. Razekenari et al., 2023. Impacts of sublethal concentrations of 17 alpha-ethinylestradiol on growth, reproductive performance, and survival in red cherry shrimp Neocaridina davidi during consecutive spawnings. Aquatic Toxicology 259: 106519.
14. Gorol et al., 2025. How does bisphenol A damage cells? Studies on the freshwater shrimp Neocaridina davidi. The European Zoological Journal 92(1): 390-409.
15. Gorol et al., 2026. Bisphenol A disrupts epithelial homeostasis in the midgut of the freshwater shrimp Neocaridina davidi. The European Zoological Journal 93(1): 563-582.
16. Ostróżka et al., 2025. Relationship between different toxic effects of nickel on tolerance and recovery in the freshwater shrimp Neocaridina davidi. The European Zoological Journal 92(1): 876-895.
17. Liu and Sung, 2011. Genes are differentially expressed at transcriptional level of Neocaridina denticulata following short-term exposure to nonylphenol. Bulletin of Environmental Contamination and Toxicology 87(3): 220-225.
18. Tsai and Sung, 2013. Development of Ecotoxicogenomic Biomarkers on the Freshwater Shrimp (Neocaridina denticulate) Following Short-Term Exposure to Dipropyl Phthalate. International Journal of Ecology 2(4): 38-49.
19. Huang and Chen, 2004. Effects of chlordane and lindane on testosterone and vitellogenin levels in green neon shrimp Neocaridina denticulata. International Journal of Toxicology 23(2): 91-95.
20. Huang, Wang and Chen, 2004. Effects of the endocrine disrupter chemicals chlordane and lindane on the male green neon shrimp Neocaridina denticulata. Chemosphere 57(11): 1621-1627.
21. Li et al., 2015. Physiological Response of Neocaridina denticulate to the Toxicity of Cu2+ and Chlorpyrifos. Environmental Science 36(2): 727-735.
22. Siregar et al., 2021. Exploiting the Freshwater Shrimp Neocaridina denticulata as Aquatic Invertebrate Model to Evaluate Nontargeted Pesticide Induced Toxicity by Investigating Physiologic and Biochemical Parameters. Antioxidants 10(3): 391.
23. Kim et al., 2018. Gene expression profiling and expression analysis of freshwater shrimp (Neocaridina denticulata denticulata) using expressed sequence tags and short-term exposure to copper. Journal of Environmental Biology 39(1): 51-57.
24. Sung et al., 2014. Acute toxicity of mixture of acetaminophen and ibuprofen to Green Neon Shrimp, Neocaridina denticulate. Environmental Toxicology and Pharmacology 38(1): 8-13.
25. Klein, Hess, Nungess, Schulte-Oehlmann and Oehlmann, 2021. Particle shape does not affect ingestion and egestion of microplastics by the freshwater shrimp Neocaridina palmata. Environmental Science and Pollution Research 28: 62246-62254.
26. Klein, Hess, Schulte-Oehlmann and Oehlmann, 2021. Locomotor behavior of Neocaridina palmata: a study with leachates from UV-weathered microplastics. PeerJ 9: e12442.
27. Ostróżka, 2025. The effect of nickel on the midgut of the freshwater shrimp Neocaridina davidi: oxidative stress, cell ultrastructure and nickel concentration. RepOD, version 1, CC BY 4.0. Data supporting the 2024 article.
28. Ostróżka, 2025. The effect of nickel on the midgut of the freshwater shrimp Neocaridina davidi: mitochondria, HSP70 and MnSOD. RepOD, version 1, CC BY 4.0. Prepublication data corresponding to the 2025 article by title, authors, grant, design and endpoints.

Use the health and mortality triage guide.

Read water measurements without turning them into diagnoses.