Priority-setting regarding offshore substances and preparations

4. Substitution of offshore chemicals

4.1 Criteria for substitution

For preparations that required to be phased out (P1 preparations), two possibilities are available:

1. reformulation of the preparation to eliminate the undesirable substances or properties giving rise to the P1 grouping, or
2. substitution by an existing alternative preparation.

Alternative preparations discussed in this chapter comprise both of these possibilities. For the evaluation of possibilities of substitution, three sets of criteria are proposed.

4.1.1 Substitution criteria # 1: P3 for P1

The highest priority is given to possibilities of replacing P1 preparations by P3 preparations which are regarded to be of low risks to the environment and to human health.

4.1.2 Substitution criteria # 2: P2 for P1

If substitution criteria # 1 cannot be fulfilled, the replacement of P1 preparations by P2 preparations should be considered. In most cases, P2 preparations will be better than P1. However, if the use of higher volumes is necessary, the improvement must be evaluated further. For preparations placed in P2(E) (due to environmental hazards), the evaluation should be confirmed by e.g. the CHARM model (when possible), because the model also includes the volumes used. For preparations placed in P2 (due to occupational health hazards), the use and handling should be evaluated further.

4.1.3 Substitution criteria # 3: Different P1 preparations

If neither P3 nor P2 preparations with the same technical properties can be found, it might for a limited period of time be acceptable to choose between different P1 preparations. Preparations are placed in group P1 due to a content of S1 substances (environmental properties)(resulting in the placement in P1(E)) and/or a classification regarding health hazards (resulting in the placement in P1(H)). It is not possible to deem one of the two hazards (to environment or health) as worse than the other. This will under any circumstances depend on a concrete assessment of the actual hazards. However, within each of the two types of hazards a ranking may be possible. The following criteria for these two considerations are proposed:

Environmental properties

The S1 substances in each of the P1(E) preparations are evaluated and the potential new preparation may be preferred if the following criteria for improvement are fulfilled:

The environmental performance is improved with respect to the following properties:

	acute toxicity: The sum of the toxic units1 in the alternative preparation is > 10 times less than in the original preparation
	bioaccumulation: The sum of the bioaccumulative units2 in the alternative preparation is 1 unit lower than in the original preparation
	biodegradation: The amount of readily biodegradable substances is greater in the alternative preparation than in the original preparation

CHARM evaluation

When it is possible to conduct a CHARM evaluation: The risk quotient of the alternative preparation is > 10 times lower than the risk quotient of the original preparation. When a CHARM evaluation is not possible, the percentage of the S1 substances in the preparation(s) must not increase.

To discriminate between different preparations, at least one of the above parameters should be fulfilled and none should be worse.

Occupational health (P1)

For comparing P1 preparations according to occupational health hazards, the following criteria for evaluation are proposed:

	Long-term effects (carcinogenecity, reproduction toxicity, mutagenicity) are worse than short-term toxicity (very toxic, toxic, acutely toxic) and imminent hazards (E, Fx, O)
	Carcinogenicity is worse than mutagenicity
	Toxicity: Tx is worse than T

At least one of the above parameters should be fulfilled and none of them should be worse.

Evaluation of preparations

As a minimum for deeming one preparation better than another, this preparation must be improved with regard to one of the environmental hazard properties or to one of the occupational health hazard properties. This assessment include the possibility of introducing technical measures to avoid human exposure to the preparation.

4.2 Identification of possibilities of substitution

In order to identify possibilities of substitution, the function of each of the preparations was identified. The function groups are those defined by OSPAR for reporting on offshore chemicals discharges, and it is important to bear in mind that preparations in the same function group may not necessarily be interchangeable, since a preparation may have been developed for specific use in one particular situation, or for use together with a suite of other preparations from the same supplier. The function groups are rather ambiguous and also overlap.

For function groups, in which relatively many different preparations from different preparation hazard categories are present, the number of preparations from each of the categories are shown in Table 4.1. Preparations from other function groups, in which no direct possibilities for substitution considerations are seen, are not included.

The above grouping of preparations describes the main technical functions, which may be further subdivided depending on the specific technical function of each of the preparations. It is not within the frame of the present project to make a detailed assessment of actual possibilities for substitution. This will require a detailed technical knowledge that is only available among operators and suppliers. In stead, the basic principle of substitution considerations are presented.

It is, however, obvious from the above table that for some of the function groups, preparations of different hazard (i.e. from different categories) are used. Especially for the function groups Nos. 5, 19, 27, 33 and 35, it seems obvious that the technical possibilities of substituting P1 and/or P2 preparations with P3 preparations should be further evaluated.

4.3 Evaluation of environmental hazards of preparations by CHARM

In order to demonstrate the principles for comparing preparations from the different categories, a generic environmental risk assessment was performed using the hazard assessment module in the CHARM model (Karman et al. 1996, Karman & Vik 1996). As examples, the environmental hazard of offshore chemicals with the same main technical function, but placed in different preparation categories (P1(E), P2(E) or P3(E)), is compared (cf. Section 4.4). Also the environmental hazard of alternative offshore chemicals that have already replaced chemicals considered more hazardous, was compared with the environmental hazard of the chemicals originally used (cf. Section 4.5).

Although the CHARM model is proposed to be used for evaluation of both offshore preparations and offshore substances, the basic principles in the model for estimating emission to the aquatic environment and the subsequent distribution between environmental compartments (i.e. the PEC estimate) and for estimating the PNEC are based on the international experience gained by evaluating single substances. This means that the use of preparation data on toxicity in the CHARM model is in principle not scientifically valid as for instance toxicity data related to preparations cannot be compared to exposure data for substances. As a trial, the CHARM model was, however, run also with preparation data.

In general, no information on the typical amounts of the chemicals used was available and consequently, it is not possible to estimate the PECs for the chemicals. Instead, a so-called "critical dosage" was estimated for each chemical. The "critical dosage" is defined as the dosage of the preparation resulting in the PEC of the most critical substance equalling the PNEC of the substance for the most critical of the environmental compartments included in the CHARM model, i.e. water phase, sediment and biota. Thus, if a dosage greater than the critical dosage is used, the estimated PEC will - at least in the most critical compartment - be higher than the PNEC. The critical dosage is estimated for each of the preparations based on preparation data or on the basis of data on the most critical substance, respectively.

4.4 Comparison of preparations with the same technical function

The comparison of preparations with the same main technical function was prepared by use of the CHARM model although they in practice may not be directly exchangeable. Each of the function groups 19 (gelling agents) and 35 (well stimulation fluids) contains a relatively high number of different preparations and among them both P1(E), P2(E) and P3(E) preparations. Obviously, the technical possibilities of replacing P1(E) and P2(E) preparations with P3(E) preparations should therefore be evaluated. Only the environmental hazards of the preparations are compared here and the actual technical possibilities for substitution are not considered.

In general, no comparison based on the preparations data could be performed as sufficient data are only available for 2 out of 21 gelling agents and 3 out of 28 well stimulation fluids. Therefore, only comparisons based on the most critical substances in the preparations are discussed and evaluated. More detailed results of the comparisons are presented in Appendix 4. An overview of the results based on the most critical substances is given below.

3 of the gelling agents are placed in group P1(E). One because it contains an EU list I substance, and two because they contain a substance on the Esbjerg Declaration list. The critical dosages of the preparations based on the most critical substances are very different ranging from 0.0001 mg/l for two of the preparations to 226 mg/l (cf. Appendix 3). However, neither the list I substance nor the substances on the Esbjerg Declaration list are the most critical substances in the preparation. This is probably due to a combination of the content of the substances in the preparations, the estimated emission with the produced water, the fate in the aquatic environment and the intrinsic ecotoxicological properties. It should moreover be mentioned that the EU list I substance is not very toxic and the present inclusion in the EU list I might be reconsidered by the European Commission.

3 gelling agents are placed in group P2(E). The critical dosages range from 0.25 mg/l to 25 mg/l. Finally, the 15 preparations, which are grouped in P3, have critical dosages ranging from 1.4 to 1000 mg/l.

Of the well stimulation fluids, 3 are placed in group P1(E) and the estimated critical dosages range from 0.08 mg/l for two preparations to 18000 mg/l for the last preparation. The reason for this great difference is that the latter contains a substance on the Esbjerg Declaration list. As no toxicity data are available on this substance, it is not taken into account in the calculation of the critical dosage.

5 preparations are placed in group P2(E) with estimated levels of critical dosages ranging from 0.00003 mg/l (for two preparations) to 30 mg/l. 20 preparations are placed in group P3(E) and the estimated critical dosages for these preparations range from 7.7 to 10000 mg/l.

The comparison is hampered by lack of data on the environmental properties of both the preparations and the substances contained in the preparations. When data on the most hazardous of the substances in a preparation are missing, the critical dosage is estimated for one of the less hazardous substances, which will most likely result in an overestimation of the critical dosage and thus an underestimation of the hazard. It should, moreover, be noted that the substance with the most hazardous intrinsic ecotoxicological properties does not necessarily need to be the most critical substance. If the most hazardous substance is used in the preparation in a low concentration or if it is retained in the oil phase due to its lipophilic properties and therefore not discharged to the environment with the produced water, another substance that is used in a higher concentration in the preparation or is more water soluble may be the most critical substance.

4.5 Comparison of original and alternative preparations

7 pairs of original and alternative preparations were evaluated covering corrosion inhibitors, gelling agents, defoamers and well stimulation fluids. The comparison was made applying both the PROSPECT environmental pre-screening method, the total content of toxic and bioaccumulative substances in the preparations (toxic units and bioaccumulative units, respectively) and the modified CHARM generic risk assessment model (calculation of critical dosage). The evaluation by use of the CHARM model is described in more details in Rasmussen & Pedersen (1997).

An overview of the results of the evaluation exercise is presented in Table 4.2. For some of the substances in the preparations, basic data are missing and consequently, the estimated critical dosage of the substance is not necessarily referring to the most critical substance. The grouping of the substances for which data are missing, is shown in the table as well.

PROSPECT pre-screening

Only for two of the preparation pairs, it was possible to compare the PROSPECT environmental pre-screening grouping of the original preparation with the grouping of the alternative preparation. For these two preparation pairs, it is seen that none of the alternatives can be evaluated as less hazardous than the original preparation as one P1(E) preparation is used in stead of another P1(E) preparation. For the remaining substitutions, all of the alternative preparations are placed on the waiting list due to lack of data of the inherent properties of the constituents, and the possible improvement of the environmental performance can therefore not be evaluated.

Comparison based on TU

The content of toxic substances, defined as the sum of the toxic units (S TU), were estimated for all substances contained in the preparations with an estimated PNEC < 0.1 mg/l. As seen in Table 4.2, only for one of the alternative preparations the S TU is lower than for the original preparation. For 3 of the remaining 6 preparations, the S TU of the alternative preparation is equal to the original, and for the other 3 preparations the S TU of the alternative is greater than of the original. Thus, in general the content of toxic substances in the alternative preparations is greater than or equal to the content in the original preparations and the alternatives can therefore not be regarded to have a better environmental performance than the original preparations. However, for most of the preparations data are lacking for some of the substances contained therein which hampers the comparison.

Comparison based on BU

Similarly, the content of bioaccumulative substances, defined as the sum of the bioaccumulative units (S BU), were estimated for all substances with a log P_ow > 3. Only for two of the alternative preparations, the S BU was reduced compared to the original preparations. For 4 of the preparations, the S BU was increased. As for the content of toxic substances, the alternative preparations do not seem to have a better environmental performance than the original.

Results of comparison of offshore preparations (p) and alternative preparations (a)

n.a. Not available due to lack of data.

The comparison by use of the CHARM generic risk assessment model based on preparation data was only possible for 3 pairs of preparations. For the remaining preparations, the estimates of the critical dose for one or both of the preparations could not be performed due to lack of data. 2 of the comparisons showed that the critical dosage for the alternative preparation was considerably higher than for the original preparation and therefore, the alternative preparation may be considered less hazardous for the environment than the original preparation. However, none of these two alternatives can be evaluated thoroughly as data for some of the substances are missing. Therefore, the critical dosage might have been lower. However, this is also the case for the original preparations. For one pair of preparations, the critical dosage for the alternative was lower than for the original and the alternative preparation is therefore considered more hazardous than the original preparation. Both of these preparations are grouped as P1 preparations and for both, data are missing for some of the substances.

CHARM - substance data

When the comparison by use of the CHARM model was based on data on the most hazardous of the substances in the preparations that could be evaluated and assigned to one of the groups (S1, S2, S3), the comparison could be made for 5 of the pairs of preparations. One of the comparisons showed that the alternative was considerably less hazardous than the original, 2 showed that the alternatives were slightly less hazardous (but probably not significantly), 1 showed that the alternative was slightly more hazardous, and 1 showed no difference between the 2 preparations.

For only 3 pairs of preparations, the comparison using the CHARM model could be prepared on both preparation and substance data. 2 of the comparisons showed the same result, namely that the alternative preparation was less hazardous than the original preparation, while the remaining comparison showed no difference when it was based on the critical component, but when the preparation data was used the alternative was evaluated as more hazardous than the original preparation.

For all preparations but Optiflo AP, data are missing for some of the substances. If one of these substances is the critical substance of the preparation, the estimated critical dosage will be too high. For three of the preparations, substance data are missing for S1-substances, which may be the critical substances of the preparation.

Lack of data

It should, however, be noted that only for one of the preparations, sufficient data on all of the substances were available so they could be grouped correctly into one of the 3 substance groups. This has almost certainly influenced the results of the comparisons of the preparations and the most hazardous substances.

Based on the above comparison it must be concluded that, despite the thorough search for data on the ingredients in the offshore chemicals, the comparison is hampered by a lack of data on the ecotoxicological properties of the substances contained in the preparations as well as of the preparations themselves.

4.6 Comparison of evaluation methods

In the above Section 4.5, 7 pairs of original and alternative preparations were compared by use of 5 different evaluation methods (PROSPECT environmental pre-screening, sum of toxic units, sum of bioaccumulative units, CHARM generic risk assessment of preparations and of substances). In the present section, the results obtained by the various methods are compared. An overview of the results is given in Table 4.3.

n.a. Comparison result not available, - alternative more hazardous than original, + alternative less hazardous than original, ~ no difference between original and alternative.

In general, only for two preparation pairs the different methods resulted in the same conclusions regarding the comparison of original and alternative preparations. This was the case for the two HAI preparations and the two J preparations for which the substitutions were evaluated to result in an improvement of the environmental performance and no significant change, respectively. For the remaining preparation pairs, the various evaluation methods gave different results.

As mentioned above the lack of data seriously hampers the comparison both of the individual pairs of preparations and also of the various evaluation methods. However, it should also be considered that the PROSPECT evaluation methods (environmental pre-screening and toxic and bioaccumulative units) are based upon the inherent properties of the individual substances contained in the preparations, while the CHARM method is based on a generic risk assessment where both the potential emission to the aquatic environment and the resulting effects are taken into account. Especially, the log P_ow of the substances is an important parameter as substances with a high log P_ow result in a high S BU, but on the same time, the CHARM model estimates these to be more or less retained in the oil phase and therefore, only a limited amount of the substances is estimated to be emitted with the produced water.

Thus, as the PROSPECT and CHARM methods are applying different approaches, different results of the evaluations may be expected. On the one hand it is tempting to use the CHARM generic risk assessment method, despite its conceptual limitations as previously discussed, as it intents to simulate the actual use and emission pattern - at least for production chemicals. On the other hand there is no evaluation of the potential hazards of the part of the chemicals that remains in the oil phase and might be emitted later in the life-cycle of the oil. Moreover, in general the CHARM model does not take into account the biodegradability of the substances in the preparations (only in estimating concentrations in the sediment compartment). Thus, the CHARM model can not be used directly for prioritising the substances with the highest intrinsic hazard.