Douma Location 4 Frost and "Magical Color Change" Dispute

November 27, 2025

The Michael Kobs Take (So Far)

This rather tedious post will address, if not settle, an issue of contention regarding the alleged chemical attack in Douma on April 7, 2018. A colleague, open-source researcher Michael Kobs, recently published a new report (PDF link), at his slick new website, regarding the gas cylinder on the bed at "Location 4." 

Most of this deals excellently with mathematical and computer modeling to assess the alleged impact of the cylinder and bounce onto the bed. But in a later section, Michael addresses the strange color change of the duvet (or comforter, or bedspread), about which opinions have differed. He compares the 2 views below. At left, we see yellow-green (or chlor) color he reads as chlorine gas floating above the duvet, which appears white. At right, the duvet is seen later in what he takes to be its natural brown color, aside from some dust at the top.

Although brown is not a very popular color for a duvet, it looks plausible enough. He shows a close-up of the blanket pattern with 3 shades of brown. There is foam material and dust from the ceiling scattered at the base of this slope and dust seemingly gathered at its top, like snow on a mountain. 

But the initial white appearance seems a mystery. Noting how no debris is moved to allow some blanket-dyeing operation, it seems a natural process at the scene was responsible. Aside from digital alteration of the image, which didn't seem especially likely, Michael notes "frost would be the only plausible explanation for the magical color change of the duvet." That could be. We know how frost can form on a gas cylinder's underside as its compressed gas contents - effectively a liquid - absorb ambient heat trying to expand back into a gas (auto-refrigeration). Any liquefied gas that dripped out would also absorb some heat in its rapid expansion to gas, although that would happen more quickly in the open air. 

(Note: that's dust either settled or sifted onto the cylinder's top, not frost - it wipes off and is later washed off - no frost on the underside is seen in any published images.)

In a post on X, Michael shows a field test involving a tons-level mass release of compressed chlorine gas, claiming the ground there is frozen afterwards. I'll assume that was the case when such a volume of liquefied gas expanding onsite in a span of minutes. And considering that, it seems possible this same exterior effect would be adequate to frost the bed. But I have my doubts. The fabric is not conductive like the steel cylinder, and the release seems small scale (maybe half the contents drained, steadily or intermittently, over 3 weeks). I further doubt it would stay that way after any frost that had formed on the cylinder is already melted away.

Let's include a view of the white duvet, cropped on the same spot. If the original color is brown, it's quite muted here, leaving the dark spots seeming bluish gray. I don't think frost is causing this.

The report noted no signs of runoff from melted frost on the cylinder's sides, suggesting it never released gas long enough to form much frost. Michael wasn't sure of the timespan of the gas release, but figured it was "short enough that the cylinder does not cool below the dew point and long enough to freeze the bed and valve by evaporating chlorine." It's not clear that such a time span is even possible. 

Besides from the blanket, Michael also sees frost on the valve in the left image below. 

It's not clear why the valve and the blanket would frost, but not the cylinder in contact with the bulk of the liquefied gas. The opening knob at the top is white in the same way although (I think) it's made of plastic, and so would not transmit heat like brass the valve is made of, and it would have no gas evaporating on it, like the blanket would have. But maybe that is metal as well. Either way, the valve cap should be the same bronze material as the valve and in contact, so it would likely frost as well. But even in the left view it lacked the usual whiteness, maybe from warming up during a recent removal so its frost melted away.

My Take (So Far)

I have always seen it differently, with no frost seen anywhere. I will explain that here, starting with a 3-image view I've been using. I'm not sure of the exact dates, but these span from just after the attack (April 8 or so) to the April 21 visit by investigators with the OPCW's Fact-Finding Mission (FFM). As I see it, these show: 1) the white duvet has its lower points and debris there somehow stained a chlor color and/or coated in visibly dense gas - 2) the staining is somehow widened and shifted to brown - 3) then it's all coated again in a wider green-brown stain as the FFM visited. I've been citing that as nearly evident fact, and I still read it this way. However, some explaining and/or questioning is in order.

Frost is not "the only plausible explanation for the magical color change of the duvet." Among the other possible options, I only care to name the one I've always proposed; the blanket might appear white because it was, and the color was added later, somehow, by the escaping gas. 

Consider how there was a similar brown color added pretty widely to the scene in at least the following areas:

- wet spatters around the valve that later forms a brown crust (see images above)

- A pale brown goo (perhaps yellowish? per an image I've seen but don't have handy) was seen on some disposable gloves found outside the bedroom, which were tested and found high in chlorides (17,000 ppm) - likely from released chlorine - and also high in zinc (1,500 ppm) - perhaps from handling the valve made of brass, which includes zinc. (2019 final report)

- The OPCW FFM noted this goo was all over the bedroom. "The FFM team observed a viscous liquid throughout the room, which was not apparent in videos. The same liquid was observed also before the entrance to the apartment and on disposable gloves present at the location" (2019 final report) They don't give a color, but as noted, the goo on the gloves was brown, and that was seemingly chlorine-related.

- Not noted but more interesting: the goo is apparent in videos. There is a reddish-brown liquid trail on the pillow that was once directly beneath the cylinder's valve (see first image above). As the following image shows, the trail is roughly the size of the valve opening (estimated here). This suggests a flow of liquid, requiring a release much faster than evaporation. The FFM and later IIT investigations assumed a slow and steady release, but that would yield single drops that would evaporate in one spot, rather than forming a trail like we see. 

- There several faint brownish circles (at least 7 or 8) on the pillow around the trail. These could be drip marks from different positions, but why would the pillow shift around that much? These could also be where the valve cap was wiped of its dust and goo. The marks seem larger in diameter than the cap, but if it were pushed in, the pillow would also contact the sides, for a wider mark more like we see. A concentric circles appearance may be caused by the cap being rotated as it was pressed in and/or by details of the embroidery. These spots are generally a faint brown color, although one is more vivid and notably reddish in color.  As it happens, the cap appears red with some paint or coating (see images above), some of which may have wiped off with the chlorine residue, or perhaps the usual reddish hue is just more prevalent here. 

In several cases I've seen in Syria, Chlorine gas leaves a brown residue. For example, the March 25, 2017 alleged attack on a White Helmets cave hospital in Al-Latamnah. The flattened end is coated in tacky-looking brown fluid, and there's a puddle of it on the ground underneath the rupture, where liquefied gas would have poured out uncontrolled (this was allegedly and apparently the cylinder's position at the time).

Searching for a link between chlorine and a brown color, the main thing I found was many sites talking about swimming pools, where chlorinated water sometimes turns brown. I gather this is largely from organic materials, but can also be caused by oxidizing metals, like iron. I didn't find much information on situations outside the pool and involving free gas. One easily-available source explains "Chlorine gas can also react with iron to produce iron (II) chloride and iron (III) chloride depending on the conditions. ...  Iron (III) chloride is characterized by its red-brown color and its ability to dissolve in water, forming a yellow-orange solution due to the presence of iron (III) ions." I suspect this is what we see: Iron (III) chloride. It's not from regular oxidization of chlorine, as I had thought, but only with prior contact with the iron in the steel container. As such, it would be regular for any case with a steel vessel, as with all cases I've seen of this brown residue.

That once-white duvet winds up appearing about this same shade of brown, either because it was coincidentally that color to start (+ mysteries) or because this ubiquitous brown color was added to the original white (+ the mystery of how). 

That would make total sense, if chlorine color-staining of fabrics (chlor or brown) were a thing. And really, it's not a thing I have ever heard of. Sources are not clear about this happening or not happening - bleaching seems the main thing that happens regarding color. Michael took that as the fact, but that may be a premature decision. I had found one source, long ago, suggesting that staining also occurs sometimes. A British doctor in World War I noted how, with intense release and a prolonged stay, a dead soldier's face sometimes "assumes a pale greenish yellow colour." (ch. 10) 

I don't recall seeing anything so stained in any Syria chlorine attack, aside from, perhaps, this case. But I don't think I've seen any others on long-exposed skin or on absorbent fabric. I looked for case examples of fabrics involved in chlorine spills, with no luck yet. But last-minute, I noticed an upholstered couch in the White Helmets photo from Al-Latamnah used above. Let's see that again, color and brightness enhanced. If chlorine does not stain fabrics, then this couch was upholstered with a mottled mix of yellow-green and reddish-brown, pretty much the same colors we're wondering about, possibly sitting a mostly yellow original color. It's an almost nauseating color mix. Can we get some frost on this thing? 

The gas cloud would be deep enough to apparently envelop the entire couch. The area in question is a small, mostly enclosed entrance and waiting area outside the hospital. Breeze depending, it seems likely a deep cloud would form here if the contents had poured out of the giant tear we see. 

So yeah... I think chlorine gas can leave such colors on fabrics like the Douma duvet.

An image from the FFM final report shows deep final staining in yellow-orange and brown under the valve, with intense corrosion - water was apparently splashed all over at least once, which would react with chlorine to produce hydrochloric acid and recall "Iron (III) chloride is characterized by its red-brown color and its ability to dissolve in water, forming a yellow-orange solution." The wood here also shows faint chlor staining, while the fabric shows it intensely. This might help us to understand why the duvet, and that couch, could be re-colored even though we don't usually see this effect when other materials are involved.  

Finally, the frosted valve that Michael had also mentioned (see images above). He probably decided on that as the dry concrete dust piled on the top would not coat the valve on all sides like this - unless the surface was sticky, perhaps with the brown liquid evident nearby. I propose this is dust covering the valve on all sides because it is sticky with the brown goo. The cap was red instead of white in the earlier view not because frost had melted, but because the dust was wiped off (on that pillow?)

While something frosting over would go to suggest a rapid and presumably controlled release, the lack of frost that I suspect allows for the same inference, just using different clues.

In my 3 images above, I see in images 2 and 3 a rapid gas release where the gas piled up deeper than it had before, to coat all but the upper pillows. The third image suggests another piling of gas shortly before the FFM visit, with new yellow-green staining added and the brown shift just beginning. (the cloud appears deeper over the pillows, but really the bed has just collapsed and sunk beneath the cylinder, which remains lightly propped on the headboard). 

(it's not clear if the opening knob ever moves - it may be left open, or closed with internal damage allowing a leak, and only the cap holds that in, until the cap is removed. A small deformity and gap in the cap might allow the light spray and spatters seen, even when it's on)

A gas release rapid enough to pile up like that would empty the cylinder quickly if it continued for long. But this never happened, suggesting that these rapid releases were somehow stopped, When they visited 18 days after the alleged attack, the OPCW's Fact-Finding Mission noted that "there did not appear to be any leakage" at the time and "no chlorine gas was detected in the room," but they neglect to mention how full the cylinder still was. On an even later visit he headed, on May 1, whistleblower "Inspector A" saw a cylinder that was still “noticeably heavy, and appeared to be mostly full of liquid.” This means probably less than half had emptied in the last 3 weeks. The valve appeared “intact” to Inspector A, although possibly leaking; he notes that "we smelled a faint trace of chlorine," perhaps from a leak in the cylinder or valve, although none has been noted. (Inspector A, The Syria Scam, p.166, 212) 

But maybe the smell was residual from the latest in a string of controlled releases. The limited amount being released in a rapid manner suggests, as does the color sequence above, and the various patterns of brown discoloration, that the release was intermittent and presumably controlled via the valve. 

Further Developments

I'll leave this space for anything further I discover to shed light on this subject.