Carbohydrazide conversion to hydrazine above 250-280 °F changes how you dose and test

You are moving a boiler program off hydrazine, and carbohydrazide is the replacement. US Patent 4,269,717, the 1981 patent that claimed carbohydrazide for boiler water, describes its hydrolysis to hydrazine under typical boiler conditions as “very minor”. Its own test table says that holds only at the cool end. At 343 °F, 60.5% of the carbohydrazide fed came out as hydrazine in 14.5 minutes. Above roughly 250-280 °F part of the feed runs as hydrazine, and hotter still most of it does: dose, test and documents must follow.

Carbohydrazide vs hydrazine: how much of the feed becomes hydrazine at boiler temperature?

In US Patent 4,269,717 (1981), 100 ppm of carbohydrazide passed through a once-through test coil for 14.5 minutes converted 7.0% to hydrazine at 282 °F, 60.5% at 343 °F and 80.2% at 400 °F. Above roughly 250-280 °F, carbohydrazide in the loop behaves as partly or mostly hydrazine.

Those figures come from Example 9, Table VI, run at feedwater pH 8.0 and 600 psi with carbohydrazide fed at 100 ppm. One run at 400 °F held the water 15.0 minutes and reached 83.0%. The patent took the hydrazine formed, less 2 ppm of free hydrazine already in the feed, as a share of 71.1 ppm, the yield of complete hydrolysis.

Each carbohydrazide molecule can hydrolyze to two molecules of hydrazine, which sets the 71.1 ppm ceiling. The test coil delivered 45 ppm of hydrazine at 343 °F, and 59 and 61 ppm at 400 °F.

Where does the patent’s “very minor” hydrolysis stop being true?

The patent gives 190-350 °F as the typical range of boiler water temperatures over which carbohydrazide scavenges oxygen and passivates metal. Its low-hydrolysis claim holds near the bottom of that range and fails near the top: 343 °F, inside the stated range, produced the 60.5% result. The authors knew hydrolysis mattered. Their Example 4 ran comparison tests at 135 °F and pH 9.0, the patent says, to prevent hydrolysis of the carbohydrazide.

Two later patents, US 4,681,737 (1987) and US 4,895,703 (1990), say in their background sections that carbohydrazide decomposes to hydrazine above 360 °F. On the measured data that marks rapid, near-complete decomposition, not the onset. If a supplier sheet quotes 360 °F as the start of conversion, it disagrees with Table VI.

The independent check is a 2006 kinetics study by Fujiwara and co-workers in the Japanese corrosion journal Zairyo-to-Kankyo. In a flow-through autoclave with 15 minutes at temperature, 0.1 and 0.3 ppm of carbohydrazide at pH 9.3 changed little at 373 K. It decomposed to hydrazine at 393 K (248 °F) and above, reaching about 30-50% at 423 K (302 °F) and about 100% at 453 K (356 °F). Between those two concentrations, the decomposition ratio did not depend on the starting concentration.

The two data sets measure different things: the patent counted hydrazine formed, the 2006 study carbohydrazide lost. A 2024 Corrosion Science paper writes, with no test conditions attached, that above 200 °C carbohydrazide decomposes with water to ammonia, nitrogen, hydrogen and carbon dioxide. The patent’s 400 °F runs sit just above that temperature, so their hydrazine yield may understate the loss.

Extrapolated from the 2006 study’s pH 9.3, 0.1-0.3 ppm tests to the patent’s pH 8.0, 100 ppm feed, 343 °F and residence time, its kinetics predict about 90% decomposition, which would put the 60.5% on the low side. The authors describe carbohydrazide and its hydrazine coexisting between 398 and 428 K; their 1997 conference abstract had already reported carbohydrazide stable below 373 K.

How thin are the conversion data?

Thin. The conversion curve a buyer needs rests on one patent table with four temperatures, one pH and one feed concentration, run in a once-through test coil in 1981 with reagent-grade carbohydrazide that carried about 2% hydrazine. The independent kinetic data, from one research group, add a second pH and a far lower concentration.

Time is the other gap. The 2006 rate constants came from holding times of 15 to 50 minutes on a first-order fit, so conversion keeps rising with time at temperature, and boiler water is not a once-through coil. Where water spends 15 minutes or more at temperature, plan for conversion at or above the published figures; short-transit points such as a feedwater heater can sit below them, and treat 250-280 °F as a band, not a sharp line.

What changes in dosing when carbohydrazide replaces hydrazine?

Start from the oxygen reaction: one mole of carbohydrazide consumes two moles of oxygen, giving nitrogen, water and carbon dioxide. On that equation the 2006 authors calculate that 3 ppm of carbohydrazide and 2 ppm of hydrazine consume nearly the same oxygen if both react completely. By mass, then, carbohydrazide needs more than the hydrazine it replaces, not less. The patent’s claim 1 sets at least 0.5 mole of carbohydrazide per mole of dissolved oxygen in alkaline boiler water.

Be careful with the half-dose rule. The patent concluded from its Example 3 that the oxygen removal reached with a given hydrazine concentration was generally reached with half as much carbohydrazide. That came from one once-through test at about 227-239 °F and pH about 10.5, and its own table shows half the dose removing 57.4% of the oxygen against 70.5% for hydrazine. Read as half the old hydrazine pounds, the rule delivers roughly a third of the old feed’s oxygen-removal capacity, so it only works where the old hydrazine feed carried a wide margin over the oxygen load.

Speed depends on conditions. At 373 K the 2006 study found hydrazine the faster scavenger, while a 2021 study summarized in a 2024 Corrosion Science paper ranked carbohydrazide’s initial oxygen reaction ahead of hydrazine’s at eight-fold excess in simulated steam generator conditions. Your measured oxygen, not either ranking, sets the dose.

Compare published feed rates on active content. One water-treatment handbook gives about 5.6 mg/L of carbohydrazide per mg/L of oxygen. By its own reaction equation the stoichiometric figure is 1.41 mg/L, so that is about 4 times stoichiometric, not the “2.5 times” printed beside it. A formulator’s bulletin for a 6.5% solution doses 23.0 ppm of product per ppm of dissolved oxygen, plus the required product residual.

Ask every quote to state the active content and whether the dose is written as product or as carbohydrazide. Two manufacturer bulletins send the feed to the deaerator storage section (one also allows the feedwater tank), undiluted and not mixed with other treatment chemicals.

Downstream, the hydrazine that forms keeps scavenging. A 2023 field study in Water Resources and Industry says carbohydrazide has the same limits to its effectiveness as hydrazine, because it decomposes to hydrazine in the boiler. What hydrazine never brought is carbon dioxide.

The 2024 loop tests showed the direction, at a dose far above field feed rates. In simulated pressurized water reactor secondary water dosed with ammonia to pH 9.8, 85 ppm of carbohydrazide against 7.5 ppm of inlet oxygen, at 212-228 °C and 288 °C, lowered pH by up to 1 unit near temperature transitions and by roughly 0.2 at temperature, read on cooled outlet water. Acid conductivity rose in the first hours.

Read the pH drop as an upper bound from a stress test, not a forecast for a boiler fed at field rates. Trend condensate pH and acid conductivity after the swap, and expect to revisit the neutralizing amine feed.

UFC 3-230-13, the US Department of Defense industrial water treatment manual (1 March 2023), lists carbohydrazide among the specialty volatile oxygen scavengers in boiler water treatment for boilers over 900 psig, normally held in the parts-per-billion range in boiler water.

If you are still choosing a scavenger, see where sodium sulfite’s pressure ceiling sits and what replaces it, and for spray water, why attemperation water, not pressure alone, decides where sulfite can go.

How should residual testing change when the loop holds two scavengers?

If your old hydrazine test is the p-dimethylaminobenzaldehyde method, it will not see the new chemical. One maker’s technical data sheet for that method, referenced to ASTM D1385-07, states that carbohydrazide up to 100 ppm does not interfere. On a carbohydrazide program, that test reads the hydrazine already formed, plus any that arrived with the product.

The iron-reduction method sees every scavenger at once. One instrument maker’s method sheet says it reacts with all oxygen scavengers without telling them apart, over 5 to 600 µg/L as carbohydrazide. The sample is read at 25 ± 3 °C straight after sampling, after 10 minutes in the dark, with ferrous iron measured and subtracted. An older procedure reads as diethylhydroxylamine (DEHA) and multiplies by 1.3 for carbohydrazide; another maker uses 1.31.

The sample point decides which test you need. Where the water has stayed below roughly 250-280 °F, the scavenger is still largely carbohydrazide, and only the iron-reduction test reads it. Past the hot sections, part or most of it is hydrazine, which that hydrazine test reports and the iron-reduction test lumps in with the rest. Run both at each point after the swap and log two numbers, not one converted figure.

The 2023 field study shows why a target must name its molecule. In a 67 barg boiler, carbohydrazide fed at the deaerator outlet at 0.5 ppm held a residual of 20-40 ppb. Residual hydrazine, specified at around 20 ppb in feedwater, read inconsistently and most often zero within detection limits. On the patent data, a hydrazine target copied to a cool feedwater point asks for a number the chemistry there may not produce.

One formulator’s bulletin allows the hydrazine test where the sample point runs above 260 °C before cooling, multiplying the result by 23.0 to give ppm of its 6.5% product. On the 2006 data, carbohydrazide is almost fully decomposed by 453 K, well below that temperature. The shortcut does not transfer to cooler points.

Which documents change when carbohydrazide replaces hydrazine?

1. Program specification. Write each residual target with its molecule, its basis (as carbohydrazide, as product or as hydrazine), the test method, any multiplier and the sample point.

2. Certificate of Analysis (CoA). Ask for the free-hydrazine content of each lot. The patent’s reagent-grade carbohydrazide carried about 2% hydrazine, and European patent application EP 0 567 275 A2 states that carbohydrazide solutions hydrolyze to hydrazine on standing (its tests held 6.5% solutions at room temperature and at 125 °F for about ten days).

3. Safety Data Sheet (SDS). Take hazard statements from the current SDS for the carbohydrazide grade you buy; the SDS governs. Keep the hydrazine documentation that covers sample points past the hot sections.

4. Food-contact steam. Carbohydrazide does not appear in the current text of 21 CFR 173.310, the boiler water additives section (eCFR, current to 24 September 2026). If your steam touches food, settle that with your regulatory lead before the swap.

5. Cycle chemistry basis. The IAPWS air in-leakage guidance (TGD9-18, 2018) names carbohydrazide as an acceptable substitute for hydrazine as the reducing agent in AVT(R), added to the condensate or feedwater.

On listings, check the molecule in the loop as well as the one in the shipping drum you receive. ECHA CHEM returned no legal-obligation entries for carbohydrazide, such as Candidate List, authorization or restriction lists, when queried on 27 September 2026. That record covers the substance you buy; the hazard classification of hydrazine applies to the hydrazine your boiler makes from it; which listing obligations follow depends on the regime and your legal role, so take that question to your regulatory lead.

The swap takes hydrazine off the purchase order, not out of the loop: conversion starts around 250-280 °F and, in the patent’s own test, passed half by 343 °F. Dose carbohydrazide on measured oxygen and active content, test for both molecules by sample point, and document what the boiler makes as well as what you buy.

Method: conversion figures are from US Patent 4,269,717 (1981), Example 9, Table VI, and from Fujiwara et al., Zairyo-to-Kankyo 55(6), 2006; dosing and testing details are from the patent, UFC 3-230-13 (2023), instrument makers’ method sheets, manufacturer bulletins and water-treatment handbooks; identity from PubChem CID 73948. Patent figures are published test data under the stated conditions, not a recommendation to operate at them; dosing and testing notes summarize published sources, not a program design, so set changes with your water-treatment specialist.