The Islamic State appears to have manufactured rudimentary chemical warfare shells and attacked Kurdish positions in Iraq and Syria with them as many as three times in recent weeks, according to field investigators, Kurdish officials and a Western ordnance disposal technician who examined the incidents and recovered one of the shells.
The development, which the investigators said involved toxic industrial or agricultural chemicals repurposed as weapons, signaled a potential escalation of the group’s capabilities, though it was not entirely without precedent.
Beginning more than a decade ago, Sunni militants in Iraq have occasionally used chlorine or old chemical warfare shells in makeshift bombs against American and Iraqi government forces. And Kurdish forces have claimed that militants affiliated with the Islamic State, also known as ISIS or ISIL, used a chlorine-based chemical in at least one suicide truck bomb in Iraq this year.
Firing chemical mortar shells across distances, however, as opposed to dispersing toxic chemicals via truck bombs or stationary devices, would be a new tactic for the group, and would require its munitions makers to overcome a significantly more difficult technical challenge.
Chemical weapons, internationally condemned and banned in most of the world, are often less lethal than conventional munitions, including when used in improvised fashion. But they are indiscriminate by nature and difficult to defend against without specialized equipment — traits that lend them potent psychological and political effects.
In the clearest recent incident, a 120-millimeter chemical mortar shell struck sandbag fortifications at a Kurdish military position near Mosul Dam on June 21 or 22, the investigators said, and caused several Kurdish fighters near where it landed to become ill.
The shell did not explode and was recovered nearly intact on June 29 by Gregory Robin, a former French military ordnance disposal technician who now works for Sahan Research, a think tank partnered with Conflict Armament Research, a private organization that has been documenting and tracing weapons used in the conflict. Both research groups are registered in Britain.
The tail of the shell had been broken, Mr. Robin said by telephone on Friday, and was leaking a liquid that emanated a powerful odor of chlorine and caused irritation to the airways and eyes.
It was the first time, according to Mr. Robin and James Bevan, the director of Conflict Armament Research, that such a shell had been found in the conflict.
In an internal report to the Kurdish government in Iraq, the research groups noted that the mortar shell appeared to have been manufactured in an “ISIS workshop by casting iron into mold method. The mortar contains a warhead filled with a chemical agent, most probably chlorine.”
Conflict Armament Research and Sahan Research often work with the Kurdistan Region Security Council. Mr. Robin and Mr. Bevan said the council had contracted a laboratory to analyze residue samples removed from the weapon.
“Soon we should have an exact composition of the chemical in this projectile, but I am certain it is chlorine,” Mr. Robin said.
He added, “What I don’t know is what kind of burster charge it had,” referring to the small explosive charge intended to break open the shell and distribute its liquid contents. The shell had not exploded, he said, because, inexplicably, it did not contain a fuse.
Whether any finding from tests underwritten by Kurdish authorities would be internationally recognized is uncertain, as the Kurdish forces are party to the conflict.
The week after Mr. Robin collected the shell, on July 6, another investigator found evidence that the research groups said indicated two separate attacks with chemical projectiles in Kurdish territory in the northeastern corner of Syria.
Those attacks, at Tel Brak and Hasakah, occurred in late June and appeared to involve shells or small rockets containing an industrial chemical sometimes used as a pesticide, the investigators said.
In the incidents in Syria, Mr. Bevan said, multiple shells struck in agricultural fields near three buildings used by Kurdish militia forces known as the Y.P.G., or Peoples Protection Units, in Tel Brak. More shells, he said, landed in civilian areas in Hasakah; at least one struck a civilian home.
Late on Friday, the Y.P.G. released a statement denouncing what it called “criminal actions” and said that in the last four weeks its forces had captured gas masks from Islamic State fighters.
The attacks at Tel Brak sickened 12 Y.P.G. fighters, who suffered many symptoms, including headaches, breathing difficulties, nausea, vomiting, eye irritation, disorientation, temporary paralysis and, in some cases, loss of consciousness, said a Western investigator for Conflict Armament Research who asked that his name be withheld for security reasons.
The investigator said he examined two impact craters at Tel Brak and also a hole in a reinforced concrete wall at Hasakah where the munitions had landed. The odor, he said, was strong, unfamiliar and soon became a painful irritant.
“It smelled like a spicy onion smell,” he said. “It was strange; it wasn’t something I could put my finger on immediately.”
He added, “We were there for perhaps 30 seconds when it started burning the nose; more than 90 seconds and the throat started to burn.”
Based on laboratory results provided by Kurdish medical officials in Qamishli, where the afflicted fighters were treated and tested, the research groups said they tentatively concluded that the shells contained phosphine, a chemical sometimes used to fumigate stored grains.
A document from the medical authorities, translated by The New York Times, referred to the laboratory tests but did not describe their methodology or show specific results.
Mr. Bevan also noted that tests so far were not conclusive.
No direct samples of the substance in the shells had been independently gathered, he said, in part because the field investigator, who did not have chemical protective equipment, experienced the onset of symptoms while working near the impact craters and had to leave the area.
Some of the shells’ characteristics from the incidents in Syria, based on photographs of the fragments, did not appear consistent with chemical weapons, including that the shell walls appeared to be thick; chemical weapons often have thinner metal skins than weapons designed to fragment.
But both Mr. Robin and the field investigator said it was possible that the attacks were tests of new weapons from the Islamic State’s makeshift munitions production lines.
The field investigator noted that at Tel Brak, the Kurdish fighters pulled back from the front-line positions after the attack, and that their former squad-sized outpost was now an observation post with fewer fighters.
“My guess is that this is going to happen again,” he said, “because it was effective.”
In an internal report provided to Kurdish officials in Syria, the two research groups recommended that Kurdish forces improve their readiness for chemical warfare incidents.
Kurdish forces “require immediate training and equipment to identify and counter chemical IED threats,” the report noted, using the acronym for improvised explosive device. It continued: “This capacity will be fundamental in maintaining civilian confidence, given the enhanced psychological impact of chemical weapon use.”
By C. J. CHIVERS