Summary
A complaint was made in September 2003 about a new type of Billy bookcase. It involved a possible health risk from formaldehyde.
Even now (6th February 2004), despite several prompts from myself, IKEA have provided no reassurance that the product is safe.
Each prompt from myself leads to a polite letter or phone call from IKEA promising to do safety tests on the product and to let me know the result, but every time the promises remain just that, promises; the tests never start, and as far as I know have still not started. The last such promise was on 21st January, but, as usual, I have heard nothing from them since.
IKEA have twice before had problems with formaldehyde being emitted from their products. On one of those occasions there were newspaper reports about it following which IKEA did take action.
Formaldehyde can cause serious health problems, particularly in those too young to complain, e.g. babies who might have such furniture in their bedrooms and yet sleep there night after night, being unable to communicate that they feel ill.
The details of the way that IKEA have dealt with my complaint are documented below.
I would be interested in hearing from anyone who has had a similar problems with IKEA products.
Email: paperclip444 at peterre dot demon dot co dot uk
(the address is deliberately obscured to deter address harvesting by spammers)
If you have a website dealing with this this sort of situation, you are welcome to link to this one.
Update
IKEA have at last provided the test results and their assessment of the same. Unfortunately, the author of the assessment appears to have misunderstood what it is that the test measures. The test results, IKEA's assessment and some discussion of the whole matter are included in the details that are given below.
Further update regarding another customer with an experience similar to mine - 5th April 2005
Following a report of another customer with an experience similar to mine, IKEA have performed a further test. The result is given below.
Further update - April 2005 - but not uploaded to this site until June because of a lack of a working scanner
IKEA have provided test results for VOCs other than formaldehyde. The results and some discussion of the same are given below.
Dear Sir,
I believe that you may have again a problem with formaldehyde emission from one of your products.
I know that IKEA has a reputation as a responsible manufacturer with regard to environmental and health issues, and that your policy is not to use formaldehyde in surface finishes, and for particleboard and MDF to keep to the German E-1 standards.
Nevertheless, with your using a large number of suppliers in several different countries, mistakes can occur.
On 13th September, 2003, I purchased an 80cm x 202cm Billy bookcase in the Oak finish from your Wednesbury (UK) store.
I assembled the bookcase that same day and placed it in position by the side of the bed in which I slept that night.
In the morning, I had uncomfortable symptoms consistent with my understanding of the effects of breathing formaldehyde (including effects on eyes, nose, throat and chest/breathing). I can supply further details if you wish. Even now (17th September), despite not having slept in that room since Saturday night, some of them persist slightly.
One of my concerns is that babies and children could be being placed to sleep near such bookcases. My symptoms were unpleasant and very noticeable, so I imagine that if they were indeed caused by the bookcase, then the effect on (say) a baby sleeping by one of the bookcases every night could be severe.
I am also concerned that I have purchased goods that seem on present evidence to be unsuitable for my (completely reasonable) purposes. I shall, of course, be asking for a refund, or a replacement by a bookcase made of safe materials, in the event that it turns out that my current bookcase is emitting formaldehyde or other dangerous substances.
I did telephone Customer Services to report the matter; the reference number is 4097234. However, I was annoyed to find (when another department phoned back) that my complaint had been documented as a complaint about a "strong smell" from the bookcase. There is indeed a smell, but it is not all that strong and it was not that that I had complained about. I had complained that the bookcase appeared to have made me ill.
The bookcase is still in position and is available for testing by yourselves and the Trading Standards Office, and tests commissioned by myself should that prove necessary. I imagine though that you can more easily perform an initial check by using the stock that you still have, and by looking at records of the materials are used in the construction of the bookcases.
I look forward to your response.
Yours sincerely,
(name withheld in this web site copy of the letter)
Dear Sir,
I refer to your letter of 23 September 2003.
Your letter passes my complaint back to the local store, which was the location that I originally contacted about it. I have arrived back there after being passed through several departments including yourselves.
The last that I heard was a message left on my answering machine on 25th September to the effect that the matter had been further passed on to "Dave Baker", a Heath and Safety Officer, who would be calling me. I have heard nothing from him.
I have begun to take further action. I have had two interviews with the Trading Standards Office in Stafford.
The possible criminal aspect has been passed to their Walsall office, since the store is in their area.
The aspect of my having been sold goods not of suitable quality is being handled by the Stafford office; they will be contacting you shortly, or may indeed have done so already.
I am surprised that you care not enough about your reputation simply to pass a complaint on and on from department to department in this way, and eventually, apparently, drop it without even the courtesy of a final answer on the matter. This is particularly surprising since there is a possibility of a health risk to your customers.
In the absence of a satisfactory reply within one week I shall take further action, and in the case that it turns out that after the one week the possibility of a health risk to the public still exists, arrange for considerable publicity.
I look forward to your response.
Yours sincerely,
(name withheld in this web site copy of the letter)
At last Dave Baker has phoned, this evening.
He apologised for the delay, saying that he had been awaiting information from Sweden. They have asked him to find out the details of any numbers, bar codes, etc. on the bookcase in order that they can investigate the manufacturing history of that particular example.
The cynical might suspect that my letter of 21st October had set in motion a train of events that caused him to phone.
However, what he said seemed very satisfactory. He offered a full refund, and I have arranged to look for the details that he requires and pass them to him tomorrow.
IKEA took back the bookcase shortly after the above events. I was informed by Dave Baker that samples of the materials would be sent for testing (presumably for formaldehyde and perhaps other toxic substances). That, of course, is a correct and responsible action for them to take.
I have received today a letter dated 18th November from Dave Baker (see below). It is a response to a letter that I sent on 22nd October to IKEA Customer services in Rotherham to inform them of the existence of this web site.
Dave Baker's letter assures me that the bookcase matter is under investigation and that I will be advised of the outcome.
I will report the outcome here. Dave Baker told me in October that the testing would probably take "several weeks". Several weeks have now passed so presumably the outcome will soon be known.
I hope that the bookcase arrived back at IKEA in a complete state and suitably identified. The arrangement had been for it to be collected on a Monday morning, but the IKEA van arrived unexpectedly on the previous Friday evening, after I had dismantled it so that it would go down the stairs but before I had packed the components ready for transport.
Dear Sir,
This letter concerns a possible serious public health danger from an IKEA product.
The details are documented on a web page at http://www.peterre.com/ikea/formaldehyde/index.html so I will not repeat them here.
I have received none of the information that was promised, about the result of the tests that I was told were being carried out (please see the web site for details).
Please let me have that information in writing. If necessary I will cause considerable publicity in order to help eliminate this possible health risk. All that it takes from you is the result, in writing, of your tests and for them to show that there is no health risk from that product.
Be aware that already a search on Google with keywords IKEA and formaldehyde returns a list with the above page at the top of the list.
Yours faithfully,
(name withheld in this web site copy of the letter)
This evening I had a phone call from IKEA in the UK. They had been contacted by IKEA headquarters in Sweden as a result of my letter of 13th January (above). It seems that the letter did not reach them until today because I had, in error, addressed it to IKEA's store in Älmhult instead of their Head Office in that same town.
I was told that because of "an oversight", the samples for testing were never sent to Sweden after all.
I was told that samples would now be sent to Sweden, and I was told the name of the laboratory in Germany where the tests would take place.
I was assured that I would be telephoned when the samples had arrived at the laboratory and testing was underway.
But IKEA promised test results before, yet the tests were not even started. I await developments with interest and will report them here.
There had been no communication from IKEA since 21st January.
IKEA's letter of 6th February provides a good outline of the standard test procedure.
However, their letter of 23rd February demonstrates that the author misunderstood what it is that the test measures. It is clear that the author believes that the test measures the "long term emission value". It does not. It measures the emission value at the time that the test is done. Even their final conclusion about the test (that it is unlikely that "a different long term emission value would have been measured had the sample been sent for analysis earlier") is couched in terms of that misunderstanding.
I give below some discussion of the test procedure, IKEA's letters, and of what conclusions may be drawn from the result of the test.
If a furniture item that emits formaldehyde is placed in a room, the level of formaldehyde in the air will vary with the temperature, humidity and amount of ventilation. The test procedure puts a representative sample from the furniture item into what is in effect a room (the test chamber), where the temperature, humidity and amount of ventilation are maintained at standard values. The level of formaldehyde in the air will then change as, for example, the moisture content of the sample approaches equilibrium with the humidity of the air in the chamber.
The result is taken when the steady state is reached, this being defined as the state when the rate of change of formaldehyde level is below some predefined value, sufficiently small that we can assume that the sample is very close to being in equilibrium with the environment within the test chamber. No doubt the specification of the standard test gives a precise definition of what rate of change of formaldehyde level is to be regarded as indicating that the steady state has been reached.
That seems to be a very satisfactory method of testing.
In order better to understand the matter, it is important to realise that the steady state as defined above is not an absolute steady state, since the formaldehyde emission will still be changing as a result of ageing of the sample. The changes due to ageing will typically be much slower than the changes that take place as a result of the sample coming into equilibrium with the environment within the chamber.
We can better understand ageing if we realise that the formaldehyde emission is typically due to two processes, with both of those processes declining as the sample ages. Initially, much of the formaldehyde emission is due to the release of trapped formaldehyde - that emission falls as the trapped formaldehyde is used up. When the trapped formaldehyde has gone, emission of formaldehyde that is produced by chemical reactions continues - that process declines more slowly. The net effect is that as the product ages, the formaldehyde emission falls initially comparatively rapidly, then later more slowly, but never quite reaches zero. The timescale involved is typically a matter of months or years.
That the formaldehyde emission never quite reaches zero is not a significant problem since formaldehyde is present at very low levels in the environment anyway.
IKEA's letter of 23rd February states that the purpose of the test chamber method is "to determine the steady state emission behaviour". That is correct, provided that we mean the steady state defined as above, that indicates that the sample is very close to being in equilibrium with the environment within the test chamber.
But the letter then goes on to say that that steady state emission behaviour reflects "the emission that a product will have over a period of time, often years". That is incorrect. It reflects the emission at the time that the test is done.
Although it does not "reflect" the emission that a product will have over a period of time, it does provide an upper bound on emission behaviour subsequent to the test, since formaldehyde emission falls as the product ages. But it provides no upper bound on emission behaviour prior to the test. For example, the test result obtained with my bookcase does show that the measured emission behaviour would be at or below 0.05 ppm if the test were repeated at any time later. But it provides no such guarantee for the behaviour at any time prior to the test.
It could be that the author was confusing the steady state as defined above with another type of steady state.
One could certainly talk about a steady state in the context of a product ageing, perhaps being defined as the (almost) steady state reached after the initial comparatively rapid decrease in formaldehyde emission. One could then relate that to "the emission that a product will have over a period of time, often years". But we would be talking about a different steady state from the one that is relevant to deciding when to take the result of the test procedure.
Unfortunately, IKEA's letter of 23rd February changes horses in mid-sentence, making a correct statement about the first type of steady state (that the purpose of the test is to determine the steady state emission behaviour), then continuing the sentence as though we were talking about the second type of steady state, saying that it "will reflect the emission that a product will have over a long period of time, often years".
The test result does not reflect "the emission that a product will have over a period of time, often years". It reflects the emission at the time that the test is done. It seems that the author of the letter did not understand that.
The nature of the author's misunderstanding is displayed even more clearly in their final conclusion about the test, which is that it is "unlikely that a different long term emission value would have been measured had the sample been sent for analysis earlier." It is clear there that the author believes that the test measures the "long term emission value". It does not. It measures the emission value at the time that the test is done.
As to whether that conclusion is true, all that we can say is that the conclusion doesn't make sense, since the test doesn't measure the long term emission value anyway.
Had the author had not misunderstood what it is that the test measures, perhaps they would have omitted the words "long term". In that case the conclusion would at least have made sense, but would almost certainly be incorrect for the following reasons.
Their conclusion would then mean that the decrease in emission (with ageing), over the period of the several months by which the test was delayed, was zero.
Other literature on formaldehyde emission describes the fall in formaldehyde emission with ageing as being initially rapid, then slower. One source describes the level as falling to 50% of its initial value after 6 months, but I do not know how accurate that result is, and it refers to uncoated board whereas the IKEA bookcases have most of the surface coated. And anyway, it will vary with the type of board. Another manufacturer's website states that they test their products within 2 to 3 weeks of manufacture. If anyone wants the references and contacts me, I will seek them out again.
Therefore, it seems very unlikely that the decrease over several months would be zero. That view is also supported by an understanding of the processes that lead to the emission of the formaldehyde. However, neither from the other literature, nor from any information that IKEA have supplied, can we know how how rapid the decrease would be.
Of course, as stated earlier, the decrease with ageing will be slow compared with the changes that take place as the test sample comes into equilibrium with the environment within the test chamber. But now we are talking about changes on a timescale of months or years. What can fairly be described as slow on one timescale can also fairly be described as fast on another.
The author of IKEA's assessment does state that the decrease will be "quite slow over time", but provides no valid justification for that. Perhaps the author believes that their sentence prior to that provides some justification, but that is the sentence that shows confusion about what it is that the test measures and it is difficult to know what was in the mind of someone who was confused in that way.
For the reasons given above, almost certainly the emission value would have been higher if the sample had been sent in earlier. But we have no way of deducing how much higher the earlier emission value would have been, e.g. if the sample had been tested during the first few weeks after purchase. Therefore we cannot be reassured by the result about the safety of the product during that time.
On the other hand, neither can we deduce that the product would have been shown to be unsafe. Because of the delay, the test result is inconclusive about the safety of the product during the first few weeks after purchase.
However, we can conclude with reasonable confidence that if someone bought a bookcase similar to mine, and kept it in a similar environment to that in which IKEA stored my bookcase while they delayed doing the test, then four months later it would be safe. The readers must decide for themselves about the value of this limited reassurance.
In case you wonder, IKEA's reference to my letter of 7th February is to one where I enquired about the identity of the sample that had been sent, whether it was from the actual bookcase that I purchased and returned to them, or another sample that had been mentioned to me by telephone, or one from current production. From the wording of their letter, it seems clear that it was my actual bookcase that was tested.
The cynical might note that IKEA did not inform me, as they had promised, when the sample arrived at the test laboratory and testing was underway. They did not inform me that the test was underway until they had received the interim report that indicated that the final result would be below the 0.10 ppm limit.
On 9 Feb 2005 I received the following email from a gentleman in the Netherlands.
We have bought Ikea furniture Malm 300.539.21 manufactured on 0451-4 23:06 (Birch effect).
Date of purchase 3rd feb 2005. We also assembled it right away. We left it in a well ventilated room for two days and then tried to sleep in the same room. My wife left the room within an hour because of the smell and the effects on breathing. In the morning I had a headache and a dry throat. We have ventilated the room during daytime with the drawers open, but in the evening after closing the windows the smell and irritations started again. I had first placed the drawers in our washing room (they seemed to be the worse), but today I have also placed the chest in the washing room in our attic.
I have asked my local store for a comment on this smell and the health implications. I have bought several furniture at home an in my office from IKEA but none of them presented this severe discomfort.
As soon as I hear more from IKEA I will keep you informed.
On 17th Feb 2005 I received from him the following email.
I have received a phone call from IKEA Heerlen. She had received my E-mail. Although I did not mention the word Formaldehyde she told me that the cause of my problems was caused by a high concentration of Formaldehyde. I had to return the chest to the shop (assembled). And we received our money back. We have not been able to buy a new one yet.
The above emails are reproduced here with the permission of their author.
I wrote to IKEA (UK) enquiring whether IKEA are now admitting that there is a problem with formaldehyde emission from their products and received the following reply.
The Malm test result provides some reassurance about IKEA products, but only of a limited nature.
The test result assures us that if we purchased a product similar to the Malm item, and stored it for the same length of time for which IKEA stored the Malm item before they did the test, and in a similar environment (particularly temperature and humidity), it would then meet the 0.1 ppm criterion by a 50% margin and therefore be safe for normal household use.
Judging by the date of purchase and the date of IKEA's letter, then a period of a few weeks would seem to be sufficient.
However, the test result provides no assurance that the level of formaldehyde emission from the Malm item was safe shortly after purchase. That is the case since formaldehyde emission falls with time (provided that there are no significant changes in other factors such as temperature), so the level of emission may have been significantly higher shortly after purchase than at the time that the test was done. But we do not know if that was the case.
It would be good if we could be sure that IKEA furniture had a safe level of formaldehyde emission even when newly purchased. If the level is not safe, customers might expose themselves to a health risk, e.g. by doing the same that I and the gentleman in the Netherlands did, i.e. sleeping in the same room as a newly purchased item. Even worse, they might put babies or children to sleep in the same room as such furniture.
In the case of my own bookcase, I returned it to IKEA so that they could perform tests, but instead of accepting a refund I took a replacement bookcase since I knew that formaldehyde emission falls with time.
The second bookcase had the same smell as the first one.
I stored it, unwrapped but unassembled, in the roof space of my detached garage throughout the summer of 2004, bringing it into the house in November.
The smell had gone, and I have had no problems with that second bookcase.
It is good that IKEA have ordered a further test to investigate the emission of other volatile organic compounds from the product. Other volatile compounds could well pose a health risk equal to or greater than that posed by formaldehyde.
If IKEA send me the result then I will be happy to post it to this site. Indeed, I am happy to post any information or comments that they send to me.
The whole matter of emission of volatile organic compounds such as formaldehyde is important since they can not only produce unpleasant symptoms but can be very injurious to health.
I find IKEA's letter of 14th April to be worrying, but not because of the figures that it gives for VOC emission.
Regarding the VOC figures, then the 1200 micrograms/m³ and 600 micrograms/m³ criteria, and the way that IKEA use them, may well be entirely reasonable. I have no information on this matter.
However, IKEA's description of of their test procedure for VOCs raises a concern that their test procedure for formaldehyde uses a similar basis, in particular that they look at the figure for formaldehyde emission after 48 hours, then if it is not below the 0.1 ppm they leave the sample in the test chamber for 28 days, then regard the test as satisfactory if the 0.1 criterion is satisfied then.
If the 0.1 ppm is their criterion for the 48 hour point, then that seems satisfactory. But if they have a higher figure for the 48 hour point, and the 0.1 ppm limit is for the 28 day point, then that would mean that freshly purchased furniture could well have a level of formaldehyde perhaps far in excess of 0.1 ppm. We cannot know if this is the case without further information from IKEA.
If the latter possibility is the case, then if a customer buys an item and ages it at normal room temperature and humidity for 28 days they should have no problems. But there is no reassurance about a product when it is 48 hours old, or indeed of any age less than 28 days. The customer may, or may not, have problems depending on how long the item has been in store since manufacture.
If IKEA do use the above procedure, then that is consistent with the experience of myself, the man in Holland and his wife. We experienced problems with freshly purchased furniture but had no problems with furniture that had been left to age for a bit. The man in Holland has in fact emailed me to inform me that a second Malm chest that he bought and left to stand for a fortnight has caused no problems.
It is also consistent with their letter of 23rd February 2004 where they say, "When testing according to the chamber method, the purpose is not to catch any initial emission (e.g. reflecting emission behaviour immediately after unwrapping a product), but to determine the steady state emission behaviour, which will reflect the emission that a product will have over a long period of time, often years."
If that is intended as a description of the purpose of a single chamber test then it is incorrect, as explained previously.
But if it is intended as a description of the purpose of their entire test procedure, including allowing the sample to age up to 28 days if necessary to meet the 0.1 ppm criterion, then the statement makes more sense. However, in that case it is consistent with the worrying possibility that IKEA's test procedure provides no guarantee as to the safety of their furniture when freshly purchased. In fact it is more than consistent with that worrying possibility - it would seem to indicate it in a fairly positive way.
Without further information from IKEA, we still cannot be sure of the safety of their products when freshly purchased. On the other hand, we cannot be sure that they are unsafe.
My logs show me that this page is viewed every few days by one or more users from the domain ikea.com, so IKEA may well write to me again with further information. I will post here any further information that they provide.